Copyright is owned by the Author of the thesis.  Permission is given for 
a copy to be downloaded by an individual for the purpose of research and 
private study only.  The thesis may not be reproduced elsewhere without 
the permission of the Author. 
 
The effectiveness of on-farm control 
programmes against wildlife-derived 
bovine tuberculosis in New Zealand 
A thesis presented in partial fulfilment 
of the requirements for the degree of 
Doctor of Philosophy at Massey University 
Carola Sauter-Louis 
2001 
Printed by Massey University Printery, 2001 
To obtain a copy of this thesis write to: 
EpiCentre 
Wool Building 
Massey University 
Palmerston North 
New Zealand 
Ph: +64 (0)6 350 5270 
Fx: +64 (0)6 350 5716 
E-mail: carola@sauter-louis.de 
Abstract 
In New Zealand the Australian brushtail possum (Trichosurus vulpecula), introduced in the middle 
of the 19th century, is the main wildlife reservoir for Mycobacterium bovis infection for farmed 
livestock and other wildlife species. Thus, control of tuberculosis (TB) has to involve both 
livestock and vector animals. Areas with endemic wildlife infection constitute 23% of New 
Zealand's land area. Vector control is mainly performed by large scale poisoning operations, by 
both aerial and on-ground baiting, conducted by official agencies, such as Regional Councils. The 
costs of vector control rose from NZ$18 million in 1995 to NZ$28 million in 1998/99, and finances 
are not available to cover all areas with endemic wildlife infection. There is a need for farmers to 
be involved and participate in TB control to complement the official control efforts. This thesis 
comprises a number of studies that looked in detail at on-farm control measures that could be 
applied at farm level, their efficiency and cost-effectiveness, in order to determine if and how 
farmers could take on-farm measures which would complement the official TB control programme. 
In an initial survey of 27 Wairarapa herd managers, whose cattle herds were TB infected, 
'grounded theory' was used to identify factors related to farm management and TB infection in 
cattle. Most farmers had knowledge or suspicion about potential high risk areas on their farm, 
where cattle were more likely to become infected with TB. Farms that grazed cattle in paddocks 
with TB hot-spot areas had a greater herd TB incidence than farms that excluded cattle from such 
areas, and used adjacent paddocks. Grazing management was found to be flexible, more so on beef 
farms than on dairy farms. These results formed the basis for designing on-farm control measures. 
A subsequent intervention study used 67 Wairarapa farms. On-farm control measures were 
implemented for three years on 34 randomly selected 'focused control' farms. On-farm control 
measures included targeted vector control in spring and autumn, and adoption of grazing 
management in summer and winter that excluded cattle from TB hot-spots during these times. 
These measures were implemented by the research team during the first two years and farmers 
continued the control work in the third year. At the end of three years the effect of the interventions 
was evaluated. Focused control farms achieved more effective TB control than standard control 
farms. They were significantly less likely to have multiple TB animals per year, a higher proportion 
of focused control farms came off Movement Control, and the two-year cumulative TB incidence 
was reduced more on focused control farms than on standard control farms. 
Part of the project was also to compare the Wairarapa project with a contemporary intervention 
study. The study was conducted on a national scale in four separate areas of New Zealand by a 
11 
national organisation, using 35 focused control and 70 standard control cattle/deer farms. Farmers 
were advised by a multi-disciplinary team on possible management changes and vector control for 
two years. The implementation of these measures was the responsibility of the individual farmers. 
Three and a half year after the start of the project the effectiveness was evaluated as part of this 
thesis. Focused control fanns reduced the two-year cumulative TB incidence more than standard 
control fanns. Comparison with the Wairarapa project indicated that the hands-on operational 
approach of the Wairarapa project had advantages over the 'advice only' approach in the national 
project. 
All fanners involved in the two intervention studies were surveyed at the end of the intervention 
studies using a questionnaires, asking about fann management and TB related issues. Only the 
Wairarapa focused control farmers were interviewed during the project period. Only slight 
differences existed in these variables between focused and standard control farms in each of the 
projects, indicating that the allocation of farms to the two fann groups was adequate. Questions 
were also asked about attitudes towards TB and its control. Overall fanners rated the importance of 
TB eradication as very high. However, the majority of fanners were not in favour of stricter 
Movement Control regulations, removal of compensation or having to pay TB testing costs 
directly. Many farmers saw organisations, such as Government and Regional Council, as being 
responsible for eradicating TB and did not see any need to conduct control programmes 
themselves. 
An economic analysis of the adoption of on-farm control measures was conducted using 
deterministic, stochastic and decision analysis. Under the current compensation level of 65% for 
TB test positive animals, the adoption of on-farm control measures generally was beneficial to 
dairy farms, but for beef farms only if they achieved TB free herd status. Reducing the 
compensation level to zero did not alter the situation significantly. The net gain in dairy farms 
increased, the situation in the beef breeding farms changed minimally and on beef finishing farms 
the adoption of control programmes became beneficial if the number of TB animals was reduced at 
least by two, without achieving TB free status. 
The final stage of the project described in this thesis was the development and use of 
FarmORACLE, a whole-fann simulation model, that allows the user to combine knowledge about 
TB and its occurrence on farms with farm-specific grazing strategies. The model was used to 
compare traditional grazing strategies with alternative strategies, that excluded cattle and deer from 
grazing TB hot-spot paddocks during high-risk times. Four farms were described in detail. In all 
four farms an alternative grazing strategy was found that resulted in higher production or greater 
economic returns, while protecting the herd against exposure to tuberculous possums. 
111 
Acknowledgements 
I came to New Zealand for the first time in 1993 and thanks to the encouragement and support I 
received during that time I decided to start a PhD. It was due to the vision of Roger Morris, my 
chief supervisor, that the project was developed and continued. Often I would think the problems 
are too big, the benefits too low, but Roger re-assured me and his enthusiasm would lift my spirits. 
I am especially grateful to Roger for providing the opportunity for me to do a PhD and for his 
personal and professional support. To Peter Wilson, my second supervisor, I am particularly 
grateful for the time and guidance I received during all these years. Especially in the last year, 
trying to teach me how to write not only correct English, but academic English. Thanks also to 
Dirk Pfeiffer, my third supervisor. For his guidance in many analytical matters and for his and 
Susanne' s  friendship I am especially grateful. The combination of these three supervisors was the 
best I could have wished for, both on a personal and on a professional level. 
I would like to express a special thanks to the late Ron Goile and to his wife, Donna. Ron 
conducted most of the fieldwork for me; he and Donna autopsied many possums and enabled me to 
go back and see my family in Germany. Without their help and input this project would never have 
been completed. Ron and Donna were like a second family to me here in New Zealand. Through 
their help and even more through their friendship I was able to grow with the project and not only 
learn about project related issues, but also about life and the importance of other things. I am 
extremely sad that Ron is not able to see this thesis finished, but I know that he will be with me in 
all the years and all the projects to come. Ron had a wonderful personality and I will never forget 
the talks we had about life, work, and personal things, sitting on the motorbike, having lunch under 
a tree, or when doing autopsies of smelly possums on the back of his truck. 
The project involved three years of intensive fieldwork on 35 farms in the Wairarapa. Thank you to 
all these farmers and their families who made us feel so welcome and part of their families. 
Without their commitment this project would never have been completed and without their 
friendship it would not have been as enjoyable. Thank you also to all the farmers who participated 
in the various questionnaires. 
This PhD would never have been completed without the support and friendship from all the people 
at the EpiCentre. Thanks to Ron Jackson, who encouraged me many times, to Fiona, whose help in 
writing English and publishing was often required; to Deb, who had to spend many hours of 
'counselling' when things didn' t  go the way I expected them to go; thanks also to Leigh and Laurie 
for their support and encouragement. A special thanks to Joanna, without her encouragement, re­
assurance and proof-reading, I probably would not have finished this PhD. Thanks to NigeI and 
Mark for their support. Thanks to Jarg, Sonja, Klim and Kathrein for the 'German' evenings and 
iv 
the great company. And a special thanks to my office colleagues over the last few years - Joanna, 
Naomi, Rene and Nina, who supplied encouragement when times were difficult and a constant flow 
of chocolate at the end of the writing-up stage to keep my (and probably their own) sanity. 
I would also like to acknowledge the tremendous help and friendship I have received from 
members of AgriQuality NZ based in Masterton, who provided me with data and field information, 
in particular Gillian Atkinson, Garth Pannett, and Alan Comelius and all the livestock officers. I 
received also considerable help from staff in the Wellington Regional Council, based in Masterton. 
Thanks also to Chris Carter of AgriQuality and Tony Rhodes of Agriculture New Zealand for the 
help in obtaining the data from the national project, so that I could conduct the analysis. 
I also would like to acknowledge the support of the AHB who funded this project and were helpful 
and patient many times during this PhD. 
Finally, the greatest and deepest thanks goes to Tommi, my husband. We met one year into my 
PhD, got married, and lived for most of the time half a world apart, he in Germany, me in New 
Zealand. It was hard for both of us at times, having a marriage by e-mail, but he was supportive all 
the years, although he really wished nothing more than that I would be at home with him. He built 
databases, entered data totally unfamiliar to him ('shoot them down' instead of 'down the chute' a 
colloquial term for condemned carcasses), created queries and analytical outputs. He would 
encourage, support, give me confidence, and edge me on many times over the phone. I am unable 
to adequately express my gratitude to him and I am looking forward to enjoy 'real' married life 
with him. Thanks also to his family for supporting him while I was away. 
I am particularly grateful to my parents, my sister and my brother. Without their encouragement, 
help and support, I would never have come to New Zealand in the first place and I would not have 
stayed that long. My father once was asked what I would be, once I have finished my PhD in New 
Zealand. He only smiled and said 'Forty'. I managed to finish it before then, but sometimes it made 
me feel even older than that. Their support and love, even when separated by 20,000 km helped me 
to continue and conquer this challenge. 
Thanks to all these people, and to all the students and staff from the EpiCentre for helping me to 
grow in myself, grow in personality, grow in character and grow professionally. 
Carola Sauter-Louis 
EpiCentre, Institute of Veterinary, Animal and Biomedical Sciences, 
Massey University, Palmerston North, New Zealand 
February 2001 
v 
Table of Contents 
ABSTRACT ....................................................................................................................................................... 1 
ACKNOWLEDGEMENTS ............................................................................................................................. III 
TABLE OF CONTENTS .................................................................................................................................. v 
LIST OF FIGURES ........................................................................................................................................ XI 
LIST OF TABLES ......................................................................................................................................... XV 
INTRODUCTION A GUIDE TO THE METHODOLOGIES AND RATIONALE FOR RESEARCH 
PRESENTED IN TillS THESIS .................................................................................................................... 1 
BIBLIOGRAPHY .............................................................................................................................................. 6 
CHAPTER 1 CHANGING BEHAVIOUR: A LITERATURE REVIEW ................................................ 9 
INTRODUCTION ........................................................................................................................................... 11 
BEHAVIOUR CHANGE ................................................................................................................................. 11 
COGNITIVE PROCESS ..................................................................................................................................... 11 
lNFORMA TION CHANNELS AND SOURCES ...................................................................................................... 13 
BRIEFOVERVlEWOF ADULT EDUCATION ...................................................................................................... 14 
INNOVATION DIFFUSION MODEL ........................................................................................................... 15 
INNOVATION PROCESS .................................................................................................................................. 15 
CHARACTERISTICS AND ADOPTION OF INNOVATIONS ................................................................................... 16 
CHANGE AGENT ............................................................................................................................................ 18 
CONSEQUENCES OF INNOVATIONS ................................................................................................................ 19 
DECENTRALISED DIFFUSION SYSTEMS .......................................................................................................... 19 
HUMAN BERA VIOUR CHANGE: THE SMOKING EXAMPLE ................................................................ 20 
EXAMPLE FROM AGRICULTURAL EXTENSION ................................................................................... 23 
DEFINITION OF EXTENSION ........................................................................................................................... 24 
ADOPTION PROCESS AND EFFICIENCY OF TECHNOLOGY TRANSFER ....... ............... ......................................... 25 
RECENT TECHNIQUES IN AGRICULTURAL EXTENSION ............................................. ...................................... 27 
'Fanner first' model ................................................................................................................................ 27 
Communication ....................................................................................................................................... 28 
Fanners' and consumers' goals .............................................................................................................. 29 
Motivation ......................................................................................................... ...................................... 30 
Computer aided programs ....................................................................................................................... 30 
AGRICULTURAL EXTENSION IN NEW ZEALAND ............................................................................................ 31 
History of extension ................................................................................................................................. 31 
Communication channels ........................................................................................................................ 34 
Factors influencing change in fann practices ......................................................................................... 36 
Commercialisation of advisory services .................................................................................................. 38 
RELEVANCE TO THE HYPOTHESIS RESEARCHED IN THIS THESIS ................................................. 39 
CONCLUSION ................................................................................................................................................ 39 
BIBLIOGRAPHY ............................................................................................................................................ 40 
vi 
CHAPTER 2 ANALYSIS OF WAIRARAPA FARMER PERCEPTIONS OF TUBERCULOSIS AND 
MANAGEMENT OPTIONS FOR CONTROL .......................................................................................... 57 
ABSTRACT .................................................................................................................................................... 59 
INTRODUCTION ........................................................................................................................................... 59 
MATERIALS AND METHODS ..................................................................................................................... 60 
GROUNDED THEORY ..................................................................................................................................... 60 
ANALYSIS OF QUALITATIVE DATA ................................................................................................................ 61 
SELECTION OF FARMS ................................................................................................................................... 62 
INTERVIEW PROCESS .•.................................................................................................................................. 63 
CONTENT ANALYSIS ......................................•.....••....................................................................................... 63 
RESULTS •.........•...•......................................................................................................................................... 65 
DESCR1P11VE ANALYSIS OF STUDY FARMS .................................................................................................... 65 
General farm characteristics .................................................................................................................. 65 
TB history from existing records kept by AgriQuality ............................................................................ 66 
BUILDING THEORIES USING THE INTERVIEWS ............................................................................................... 68 
Tuberculosis related observations by farmers ........................................................................................ 68 
General farm management ...................................................................................................................... 75 
Grazing management .............................................................................................................................. 76 
DISCUSSION .................................................................................................................................................. 79 
DESCR1P11VE ANALYSIS OF STUDY FARMS .................................................................................................... 79 
METHODOLOGY ........................................................................................................................................... 80 
INTERVIEW CONTENT ANALYSIS ................................................................................................................... 81 
BIBLIOGRAPHY .....•.............•....................................•................................................................................... 87 
CHAPTER 3 EFFECTIVENESS OF ON-FARM TUBERCULOSIS CONTROL PROGRAMMES: 
FARMS LOCATED IN THE W AlRARAPA ............................................................................................. 93 
ABSTRACT .................................................................................................................................................... 95 
INTRODUCTION ...........•............................................................................................................................... 95 
MATERIALS AND METHODS ..................................................................................................................... 97 
STUDY AREA ................................................................................................................................................ 97 
POSSUM CONTROL EFFORTS BY mE REGIONAL COUNCIL ............................................................................ 98 
FARM SELECTION PROCESS ........................................................................................................................... 99 
CATTLE TB DATA ...............................................................................................................•...................... 101 
CONFIRMA nON OF TB STATUS .................................................................................................................. 101 
FARM VISITS .............................................................................................................................................. 102 
TB CONTROL MEASURES EMPLOYED IN TIllS STUDy ................................................................................... 103 
Basis of the hypothesis .......................................................................................................................... 103 
Targeted localised possum control ....................................................................................................... 105 
Livestock grazing management practices ............................................................................................. 107 
ANALYSIS OF DATA .................................................................................................................................... 107 
RESULTS .......•........•.........•........................................................................................................................... 108 
POWER ANALYSIS ....................................................................................................................................... 108 
VECTOR CONTROL CONDUCfED ON FOCUSED CONTROL FARMS BY THE RESEARCH TEAM .......................... 110 
VECTOR CONTROL BY REGIONAL COUNCILS .............................................................................................. 113 
ANALYSIS OF TUBERCULOSIS TESTING RECORDS ........................................................................................ 114 
Time spent on Movement Control and Herd TB status at the end of the project .................................. 114 
Number of TB cattle .............................................................................................................................. 116 
Cumulative incidence and its reduction over three years ..................................................................... 117 
DISCUSSION •...............................................................••...•................•..............•..••..................•................... 121 
vii 
CONCLUSIONS ...................................................................................•....•................................................... 127 
BmLIOGRAPHY .......................................................................................................................................... 127 
CHAPTER 4 EFFECTIVENESS OF ON-FARM TUBERCULOSIS CONTROL PROGRAMMES: 
COMPARISON OF WAIRARAPA STUDY WITH A CONTEMPORARY NATIONAL STUDY .... 133 
ABSTRACT ................................................................................................................................................... 135 
INTRODUCTION ......................................................................................................................................... 135 
MATERIALS AND METHODS ................................................................................................................... 136 
STIJDY AREAS ............................................................................................................................................. 136 
FARM SELECTION PROCESS ......................................................................................................................... 137 
METHODS EMPLOYED ................................................................................................................................. 138 
CATILE AND DEER TB DATA AND CONFIRMATION OF TB STATUS .............................................................. 139 
ANALYSIS OF DATA .................................................................................................................................... 139 
RESULTS ...................................................................................................................................................... 140 
VECTOR CONTROL CONDUCTED ON THE PROJECT FARMS BY THE REGIONAL COUNCIL .............................. 140 
ANALYSIS OF TUBERCULOSIS TESTING RECORDS ........................................................................................ 141 
Time spent on Movement Control and herd TB status at the end of the project .................................... 141 
Number of TB animals ........................................................................................................................... 142 
Cumulative TB incidence and its reduction over three years ................................................................ 143 
COMPARISON OF THE NATIONAL PROJECT wrrn THE W AIRARAPA PROJECT ................................................ 147 
Comparing only cattle fanns ................................................................................................................. 147 
Comparing all farms ............................................................................................................................. 149 
DISCUSSION .......................................................................................................... ...................................... 151 
EVALUATION OF THE NATIONAL ONE-ON-ONE PROJECT .............................................................................. 151 
COMPARISON BETWEEN NATIONAL AND W AIRARAP A PROJECT .................................................................. 154 
BIBLIOGRAPHY .......................................................................................................................................... 154 
CHAPTER 5 ATTITUDES OF FARMERS TO BOVINE TUBERCULOSIS CONTROL IN NEW 
ZEALAND .................................................................................................................................................... 159 
ABSTRACT ................................................................................................................................................... 161 
INTRODUCTION ......................................................................................................................................... 161 
MATERIALS AND METHODS ................................................................................................................... 162 
SURVEYS .................................................................................................................................................... 162 
Wairarapa project farms ....................................................................................................................... 162 
National project farms ........................................................................................................................... 163 
DEFINITION OF TERMS USED ....................................................................................................................... 165 
ANALYSIS OF DATA .................................................................................................................................... 165 
RESULTS ...................................................................................................................................................... 166 
W AIRARAP A FARMS .................................................................................................................................... 166 
General farm characteristics ................................................................................................................. 166 
Herd manager ....................................................................................................................................... 171 
Stock management ................................................................................................................................. 172 
TB risk assessment ................................................................................................................................. 173 
Vectors and vector control on farms ..................................................................................................... 175 
Attitudes towards TB and its control ..................................................................................................... 179 
Perceived cost of TB and its control by farmers .................................................................................... 18 6 
Multivariate analysis between Wairarapa focused and standard control farms ................................... 187 
NATIONAL STIJDY FARMS ........................................................................................................................... 187 
General farm characteristics ................................................................................................................. 187 
viii 
Herd manager ....................................................................................................................................... 190 
Stock management ................................................................................................................................ 192 
TB risk assessment ................................................................................................................................ 193 
Vectors and vector control on farms ..................................................................................................... 195 
Attitudes towards TB and its control ..................................................................................................... 197 
Perceived cost of TB and its control by farmers ................................................................................... 202 
Multivariate analysis between national focused and standard control farms ....................................... 202 
DISCUSSION ............................................................................................................................................... 203 
COMPARING FOCUSED CONTROL FARMS WITII STANDARD CONTROL FARMS .............................................. 203 
A TITIlJDES TOWARDS TB AND ITS CONTROL ....................................................................................•........ 204 
CONCLUSIONS ...........................................................................................................................................•. 206 
BmLIOGRAPHY ........................................................................................................................................ 207 
CHAPTER 6 ECONOMIC EV ALUA TION OF TB CONTROL PROGRAMMES AND 
POTENTIAL BENEFITS OF USING INCENTIVES OR AN INSURANCE SCHEME FOR 
DIFFERENT FARM TYPES ..................................................................................................................... 209 
ABSTRACT .•................•....••..•...•..................•.....................•........•.....•....•.••.••••.••...•.•••.•..•.•...•..•..••......•........ 211 
DEFINITIONS ............................................................................................................................................. 211 
INTRODUCTION ....................................................................................................................................... 212 
NEED FOR EcONOMIC ANALYSIS ......................................................................................................•........ 213 
TECHNIQUES AVAILABLE FOR ANALYSIS ....................................................................................•.......•....... 214 
AIM OF TIllS STUDY .................................................................................................................................... 214 
INSURANCE ................................................................................................................................................ 214 
MATERIALS AND METHODS ................................................................................................................ 216 
REpRESENTATIVE FARMS INCLUDED IN THESTUDY .................................................................................... 216 
TuBERCULOUS ANWALS ............................................................................................................................ 217 
DETAILS OF COSTS AND REVENUES USED IN lEE PARTIAL BUDGETING ....................................................... 218 
Additional returns resulting from the implementation of control measures ......................................... 218 
Reduced costs resulting from the implementation of control measures ................................................ 219 
Additional costs resulting from the implementation of control measures ............................................. 219 
Returns foregone as a result of the introduction of control measures .................................................. 221 
PARAMETERS USED IN lEE ECONOMIC ANALYSIS ....................................................................................... 221 
Deterministic model .............................................................................................................................. 221 
Stochastic model ................................................................................................................................... 222 
Decision analysis .................................................................................................................................. 225 
RESUL TS ..................................................................................................................................................... 226 
DETERMINISTIC MODEL ..............................................................................................................•............... 226 
Current situation (65% compensation) ................................................................................................. 226 
Alternative situations with reduced or no compensation for reactor animals ...................................... 228 
Alternative situation with subsidies on control costs ............................................................................ 230 
Break-even points .................................................................................................................................. 232 
STOCHASTIC MODEL (@RISK) ..............................................•...•............................................................... 233 
Stochastic model on dairy farms ........................................................................................................... 233 
Stochastic model on beef breeding and beef finishing farms ................................................................ 236 
Comparing 65% compensation level for reactors with zero compensation .......................................... 239 
DECISION ANALySIS ...................................................................................................................•......•........ 241 
Decision analysis on dairy farms .......................................................................................................... 241 
Decision analysis on beef breeding farms ............................................................................................. 242 
Decision analysis on beef finishing farms ............................................................................................. 246 
PROVISION OF VOUCHER FOR CONTROL AND OFF-MC-PA YMENT FOR TIlE W AIRARAP A REGION .•....•......... 249 
DISCUSSION ............................................................................................................................................... 251 
FACTORS CONSIDERED AND OMITTED IN TIlE PARTIAL BUDGET ................................................................. 251 
ix 
FARMS USED IN THE STIJDY AND TB REACTOR SCENARIOS ANALYSED ....................................................... 253 
RESULTS FROM THE DETERMINISTIC AND STOCHASTIC MODELS ................................................................. 254 
Farm types ............................................................................................................................................. 255 
Compensation level and subsidies ......................................................................................................... 255 
RESULTS FROM THE DECISION ANALYSIS .................................................................................................... 256 
INSURANCE AS AN ADDmONAL METHOD .................................................................................................... 257 
EFFECTS FOR FUTURE CONTROL OF TB ....................................................................................................... 258 
TB and its effect in the wider economy .................................................................................................. 262 
CONCLUSIONS ............................................................................................................................................ 263 
BIBLIOGRAPHY .......................................................................................................................................... 264 
CHAPTER 7 FARMORACLE, A FARM SIMULATION MODEL .................................................... 271 
ABSTRACT ................................................................................................................................................... 273 
INTRODUCTION ......................................................................................................................................... 273 
OVERVIEW OF FARMORACLE ................................................................................................................ 275 
SETTING UP PADDOCKS AND ANIMAL CLASSES/MOBS ................................................................................. 277 
GRAZING PLANS IN FARMORACLE ........................................................................................................... 280 
OUTPUT OF FARMORACLE ....................................................................................................................... 282 
USE OF FARMORACLE ON SOME OF THE WAlRARAPA FOCUSED CONTROL FARMS .............. 285 
COMPARING GRAZING PLANS ON DAIRY FARMS .......................................................................................... 285 
Dairy farm A .......................................................................................................................................... 285 
Dairy farm B .......................................................................................................................................... 288 
COMPARING GRAZING PLANS ON BEEF FARMS ............................................................................................ 291 
Beef breeding farm ................................................................................................................................ 291 
Beef finishing farm ................................................................................................................................ 296 
DISCUSSION ................................................................................................................................................ 301 
POTENTIAL FURTHER DEVELOPMENT OF FARMORACLE ........................................................................... 305 
CONCLUSIONS ............................................................................................................................................ 305 
BIBLIOGRAPHY .......................................................................................................................................... 306 
CHAPTER 8 GENERAL DISCUSSION ................................................................................................. 313 
INTRODUCTION ......................................................................................................................................... 315 
ARE THERE ANY CONTROL MEASURES AVAILABLE FOR FARMERS? ......................................... 316 
ARE THESE ON-FARM METHODS PRACTICAL AND EASILY IMPLEMENTED? ............................ 316 
ARE THESE ON-FARM CONTROL MEASURES EFFECTIVE? ............................................................. 317 
ARE THESE ON-FARM CONTROL MEASURES FINANCIALLY WORTHWIDLE? ........................... 317 
ATTITUDES OF FARMERS TOWARDS TB CONTROL.. ........................................................................ 318 
WHAT UNDERSTANDING OF BEHA VIOUR IS IMPORTANT FOR ACHIEVING CHANGE? .......... 319 
OPPORTUNITIES FOR ENHANCEMENT OF ON-FARM PROGRAMMES ........................................... 320 
FUTURE RESEARCH PROPOSALS EMANATING FROM THIS STUDY .............................................. 320 
ADOPTION OF CONTROL METHODS STIJDIED ............................................................................................... 320 
F ARMORACLE .......................................................................................................................................... 321 
IMPLICATIONS OF THE METHODS STUDIED FOR TB CONTROL .................................................... 321 
x 
CONCLUSION ............................................................................................................................................. 322 
BmLIOGRAPHY .......................................................................................................................................... 322 
APPENDIX ................................................................................................................................................... 325 
APPENDIX I: QUESTIONNAIRE USED FOR THE INTERVIEWS (CHAPTER 2) ................................ 327 
APPENDIX 11: LIST OF CATEGORIES AND CODES USED IN WINMAX98 ....................................... 331 
APPENDIX Ill: REGULATIONS REGARDING TB CONTROL AND TESTING ................................... 334 
APPENDIX IV: QUESTIONNAIRE USED ON W AIRARAPA AND NATIONAL PROJECT FARMS .. 339 
APPENDIX V: SPREADSHEETS USED FOR ECONOMIC ANALYSES ................................................ 351 
xi 
List of Figures 
Figure 1: Text coding windows for TB history of one sample farm using the qualitative 
software programme WinMAX98 ............................................................................................ 64 
Figure 2: Example of a Mind Map with categories and events used in describing the TB 
situation and TB perception of one farmer in the sample ......................................................... 65 
Figure 3. Box-plot of five-year cumulative TB incidence rates (lesioned animals 1 990-1 994) 
for the three main farm types (excluding one extreme outlier in the beef breeding group 
with an incidence of 0.87) . ....................................................................................................... 67 
Figure 4. Location of study farms in the Wairarapa in the North Island of New Zealand . .............. 98 
Figure 5. Average monthly point prevalence of TB in possums (data obtained from the 
longitudinal study in Castlepoint) . ......................................................................................... 104 
Figure 6. Proportion of possums dying from tuberculosis per month of population at risk . .......... 105 
Figure 7. Relationship between power and sample size at four different proportions of focused 
control farms and 30% of standard control farms remaining on Movement Control.. ........... 109 
Figure 8. Relationship between sample size and power to detect a difference in cumulative 
incidence of TB of 0.01 and 0.02 between focused and standard control farms with a 
common standard deviation of 0.045 . .................................................................................... 1 10 
Figure 9. Percentage of farms under RC vector control, assuming that a control operation lasts 
for four years . ......................................................................................................................... 1 13 
Figure 10. Percent of focused and standard control farms that had equal or more than one, 
two, or three reactors in any one of the two years .................................................................. 1 1 7 
Figure 1 1. Frequency histogram of pre-study cumulative incidence of TB animals in 1995/96 
for focused and standard control farms . .......................................... H ••••••••••••••••••••••••••••••••••••• 1 18 
Figure 12. Frequency histogram of cumulative incidence of TB animals 1998/99 for focused 
and standard control farms . .................................................................................................... 1 19 
Figure 13. Distribution of 2yr cumulative incidence (skin test reactors plus lesioned culls) for 
1 995/96, stratified by farm group and herd type . ................................................................... 120 
Figure 14. Distribution of 2yr cumulative incidence (skin test reactors plus lesioned culls) for 
1 998/99, stratified by farm group and herd type . ................................................................... 120 
Figure 15. Study farm locations within four areas of New Zealand ............................................... 137 
Figure 16. Percent of farms under RC vector control, assuming that a control operation lasts 
for four years . ......................................................................................................................... 140 
Figure 17. Time (in months) spent on Movement control by focused and standard control 
farms between 1 996 and 1 998 . ............................................................................................... 14 1 
Figure 18. Proportion of farms with one, two, three or more TB animals in the years 1993/94; 
95/96 and 98/99 for focused control and standard control farms . .......................................... 143 
Figure 19. Reduction in cumulative TB incidence between 1 993/94 and 1 997/98 for focused 
and standard control farms . .................................................................................................... 144 
Figure 20. Cumulative TB incidence in different herd types of focused and standard control 
farms for 1993/94 and 1997/98 . ............................................................................................. 145 
Figure 2 1. Comparison of reduction in cumulative TB incidence in cattle farms of Wairarapa 
and national focused and standard control farms . .................................................................. 149 
Figure 22. Effective farm size distribution of all farms included in the Wairarapa study .............. 167 
Figure 23. Violin plots for effective farmed area of Wairarapa focused control, standard 
control and non-TB farms . ..................................................................................................... 168 
Figure 24. Cattle herd size distribution of farms included in the study in livestock units . ............ 169 
Figure 25. Violin plots for the cattle proportion of total livestock units for focused control, 
standard control, and non-TB farms ....................................................................................... 170 
Figure 26. Violin plots for cattle density on focused control, standard control, and non-TB 
farms . ...................................................................................................................................... 170 
XlI 
Figure 27. Distribution of age groups of herd managers on focused control, standard control, 
and non-TB farms . ... . . . .... . ....... ........ . . . ..... .. .. . . ........... ............................................................... 17 1  
Figure 28. Personality trait means for focused control, standard control, and non-TB farmers 
(adapted from Seabrook, 1984) ............................. . . . . . . . . . ...... ... .... . . .... . ....... ......... .... ..... . .. .. ...... 172 
Figure 29. Time spent per year on vector control by focused control, standard control, and 
non-TB farmers . ..................... ...... . . . . . . . . .. . . . . . . . . . . .. . ......... . . ......... ..... .............. ..... ... . ................... 178 
Figure 30. Box and Whisker plot of the proportion of farm area controlled by focused control, 
standard control, and non-TB farmers . ........................................... .... .................... ................ 179 
Figure 3 1 .  Importance of TB eradication as considered by Wairarapa focused control, standard 
control and non-TB farmers . .... ........................ ......................................................... ............. 1 80 
Figure 32. Attitudes of farmers towards Movement Control restrictions, if they should be 
stricter or less strict, stratified by Wairarapa focused control, standard control, and non-
TB farmers . .. ................................ ... . ....... ...... .... . ...... ... ............ ..................... ........................... 1 8 1  
Figure 33. Farmers' belief regarding the effect of removing compensation for TB reactor 
cattle, stratified by Wairarapa focused control, standard control, and non-TB farmers . ........ 1 82 
Figure 34. Organisations/institutions perceived as being responsible for TB eradication . ............ 1 84 
Figure 35. Effective farm size distribution of all farms included in the national study . ................ 1 88 
Figure 36. Violin plots for effective farmed area of national focused and standard control 
farms . .......... . . .. .. . ...... . ................ . . . . . . .... . .. .................... . . . ........ .. ..... . ................. ... . ..... ................ 189 
Figure 37. Distribution of age groups for herd managers of national focused and standard 
control farms . .......................................................................... . . . . . . . .. . . . . .. .. . . . ............................ 190 
Figure 38. Personality trait means for national focused and standard control farms (adapted 
from Seabrook, 1984) . .. . ........ . . . . ................ . .. . .. ... .......... ............... ...... .... ....................... ..... .... . 1 92 
Figure 39. Days spent per year on vector control by national focused control and standard 
control farmers .................... .... ....... .............. .. ............................ . . ..... ...................... .............. . .  196 
Figure 40. Importance of TB eradication as considered by national focused and standard 
control farmers ........................................................................................................................ 1 97 
Figure 41 .  Percent of national focused and standard control farmers who believed Movement 
Control restrictions should be more or less strict. .................................................................. 198 
Figure 42. Organisations/institutions perceived as being responsible for TB eradication . ............ 200 
Figure 43. Expected net returns on dairy farms for different reductions in reactor numbers 
using different compensation levels for reactor animals . ...... . . . . ........ . . ..... . . . . . .. . . . . . . ...... . . . . . . ..... 228 
Figure 44. Expected net returns on beef breeding farms for different reductions in reactor 
numbers using different compensation levels for reactor animals . .................. ....... ... .. .......... 229 
Figure 45. Expected net returns on beef finishing farms for different reductions in reactor 
numbers using different compensation levels for reactor animals . ...................................... . .  229 
Figure 46. Additional revenues per animal reduction in reactor numbers and if the herd came 
off Movement Control. . .... ............. . . ....... . . . . . ..................................................... ...................... 233 
Figure 47. Economic outcome distributions from @Risk stochastic partial budgeting model 
for returns minus costs for six scenarios of reductions in reactor numbers for a dairy 
farm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  234 
Figure 48. Tornado graph with results of sensitivity analysis showing the importance of 
influence of different input variable for the ituation where a dairy farm had 5 reactors 
and reduced it to 2/yr due to the implementation of on-farm control programmes . ............... 235 
Figure 49. Comparison of the range of expected returns minus costs from the stochastic 
@Risk partial budgeting model for the six scenarios of reducing reactor numbers in a 
dairy farm . ..... . . . ... ..................... ...... ........................................................................................ 236 
Figure 50. Economic outcome distributions from @Risk stochastic partial budgeting model 
for returns minus costs for six scenarios of reductions in reactor numbers for a beef 
finishing farm (ordered from worst case scenario to best) . .................................................... 237 
Figure 5 1. Tornado graph with results of sensitivity analysis showing the importance of 
influence of different input variables for the situation where a beef finishing farm had 2 
reactors and reduced it to one every 2nd year due to the implementation of on-farm 
control programmes . . . .......... .. .. ................ ........ . . . ..... ...... ............... ..... . . .. . ..... ........ . . .. . .. .. .... . ... .. 238 
Figure 52. Tornado graph with results of sensitivity analysis showing the importance of 
influence of different input variables for the situation where a beef breeding farm had 5 
reactors per year and reduced it to zero due to the implementation of on-farm control 
xiii 
programmes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  238 
Figure 53. Expected revenues minus costs from stochastic partial budgeting (@RlSK) for all 
three farm types and all six scenarios of reducing reactor numbers with 65% and zero 
compensation . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  240 
Figure 54. Decision tree for expected fmancial outcomes for adoption or non-adoption of on-
farm control measures on a dairy farm with five reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  242 
Figure 55. Decision tree for expected financial outcomes for adoption or non-adoption of on-
farm control measures on a beef breeding farm with five reactors . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . 243 
Figure 56. Decision tree for beef breeding farm starting with two reactors and changed 
probabilities of reducing reactor numbers after implementing on-farm control 
programmes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  245 
Figure 57 . Subsidies and off-MC-payments for beef breeding farms in order to make adoption 
of on-farm control programmes the preferred option in the decision analysis for the five-, 
two, and one-reactor starting situation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  246 
Figure 58. Subsidies and off-MC-payments for beef finishing farms in order to make adoption 
of on-farm control programmes the preferred option in the decision analysis for the five-, 
two, and one-reactor starting situation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  248 
Figure 59. Graphical display of FarmORACLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  277 
Figure 60. Creating a paddock map in FarmTracker. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  278 
Figure 6 1 .  Setting up grazing plans in FarmORACLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  28 1 
Figure 62. Location of animal groups on the farm map in FarmORACLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  282 
Figure 63. FarmORACLE output of animal intake and farm cover for dairy cows . . . . . . . . . . . . . . . . . . . . . . .  283 
Figure 64. FarmORACLE output of animal liveweight plus supply and demand for dairy 
cows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  284 
Figure 65. Paddock map of dairy farm A, with TB hot-spot area in grey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  286 
Figure 66. Grazing routines during summer and winter on dairy farm A, using traditional 
grazing plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  287 
Figure 67. Alternative grazing routines during summer and winter on dairy farm A, excluding 
TB hot-spots from grazing at these times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  287 
Figure 68. Paddock map and TB hot-spot (in grey) on dairy farm B. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  289 
Figure 69. Aerial photograph of beef breeding farm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  292 
Figure 70. Aerial photograph of beef breeding farm with paddock layout. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  293 
Figure 7 1 .  Paddock map on beef breeding farm (TB hot-spot in grey, pine plantation in dark 
green) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  294 
Figure 72. Grazing locations of cattle and sheep, using traditional grazing plans, for August . 
1 998 on the beef breeding farm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  295 
Figure 73. Paddock map and TB hot-spot (in grey) of the beef finishing farm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  297 
Figure 74. August grazing plans under the traditional grazing scheme for cattle and sheep on 
the beef finishing farm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298 
Figure 75. December grazing plans under the traditional grazing scheme for cattle and sheep 
on the beef finishing farm . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  298 
Figure 76. Alternative grazing plans for cattle during high risk times (winter and summer) . . . . . . . .  299 
xiv 
xv 
List of Tables 
Table 1. Mean [and range] of some characteristics of study farms in the Wairarapa . . . .. . . . . . . . . . . . . . . . . .  66 
Table 2. Distribution of focused and standard control farms by herd type and Regional 
Council vector control area (December 1996) . ...................................................... ................ 100 
Table 3. Details of possum trapping and poisoning on the focused control farms, together with 
their effective farmed area . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . .. . . . . . . . . .. . . . .. .. . . . . .  1 12 
Table 4. Number of focused and standard control farms by TB status at the end of the 
intervention programme (December 1999) . .............................. . .. ... . ...................................... 1 15 
Table 5. Transitions of TB status of focused and standard control farms during the project 
period 1997-1999. (1= infected, C= clear status, D= disbanded) . . . . . . . . . . . . . . . . . . . . . ... . . . .. . . . . . . .. .. . . . .  1 16 
Table 6. Number of focused and standard control farms with one, two or more reactors in any 
one of the two years . . . . .  '" ........................................................................................................ 1 17 
Table 7. Average two-year cumulative incidence (cum inc.) of TB animals in focused and 
standard control farms and the reduction versus the 1995/96 cumulative incidence . . .. . . . . . . . . .  1 1 8  
Table 8. Within group average of two-year cumulative incidence of TB animals in focused 
and standard control farms, stratified for herd type; and reduction in cumulative 
incidence achieved. (The number of farms in each category is shown in brackets. )  . . . . . . . . . . . . . .  1 19 
Table 9. Two-year cumulative TB incidence for 1 995/96 and 1998/99 for focused and 
standard control farms, stratified by Regional Council control received prior to June 
1 996 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  121  
Table 10. Number of focused control and standard control farms by TB status at the end of 
1 998 - in brackets the percentage of total farms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  142 
Table 1 1 . comparing TB status at end of 1998 within herd type, in brackets percentage of 
farms in that herd type group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  142 
Table 12. Number of focused and standard control farms with one, two or more TB animals in 
any one of the two years . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .  143 
Table 1 3. Average two-year cumulative TB incidence (Cum inc.) for national focused and 
standard control farms . . . . . . . . . . . . . . . . . . . . . . .  , .................................................................................... 144 
Table 14. Average two-year cumulative incidence (Cum inc.) of TB in focused control and 
standard control farms, stratified for herd type and reduction in cumulative incidence 
achieved . . . . . . . . . . . . . . . . . ... . . . . . . . ... . . .. . .. . . . .. . . . ... ... . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  145 
Table 15. Comparison of two-year cumulative TB incidence (Cum inc.) and reduction 
between focused and standard control farms stratified by regions (the number of farms is 
shown in brackets) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  146 
Table 16. Two-year cumulative TB incidence (Cum inc.) on focused and standard control 
farms stratified on whether they had received vector control prior to the start of the 
project mid 1995 (the number of farms is given in brackets) . . . . . . . . . . .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .  146 
Table 17. TB status ofWairarapa and national cattle study farms at the end of the projects . . . . . . . . .  147 
Table 1 8. Number of Wairarapa and national study cattle farms with one, two, three or more 
TB animals in the final study year . .. . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  147 
Table 19. Comparison of two-year cumulative TB incidences in the years prior to the 
commencement of the intervention studies and the last two years of the projects of 
Wairarapa and national study farms (cattle farms only) ......................................................... 148 
Table 20. TB status of Wairarapa and national study farms at the end of the projects . ................. 1 50 
Table 21 .  Number of Wairarapa and national study farms with at least one, two or three TB 
animals in the final study year . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . .. . ... . . . . . . . . . . . . . . .... . . . . . . . . . . . . . .  150 
Table 22. Comparison of two-year cumulative TB incidences (Cum inc.) in the years prior to 
the commencement of the intervention studies and the last two years of the projects of 
Wairarapa and national study farms (all farms) . .. . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . .. . . ... . .. . . . . . .. . . . . . . .  15 1 
Table 23. Number of farmers responding to the questionnaire in the Wairarapa and National 
project. ... . . . . . . .. .... . . . . .. . . . . .. .. . . . .. . . . . . . . . . . . .. . . . . . . . .. . . . . . . . . . .. .. . . .. .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . .. . .. . .. .. . .. . . . .  164 
xvi 
Table 24. General characteristics of focused control, standard control, and non-TB farms . . . . . . . . . . .  1 67 
Table 25. Average farmed area stratified by herd type and Wairarapa farm group (with range 
in brackets) . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . .. .. . . . . .. . . . . . . . . . . . . . . .  168  
Table 26. Information on stock movements on and off farms for focused control, standard 
control, and non-TB farms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1 73 
Table 27. Information on the TB situation and perception of focused control, standard control, 
and non-TB farms . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . .  1 75 
Table 28. Assumed likelihood of contact between possums/ferrets and livestock, as indicated 
by focused control, standard control, and non-TB farmers . . . . . . . . . . . . . . . . . . . . . . . . . . . .. .. . . . . . . . .. . . . . . . ... . .. 176 
Table 29. Vector control by Regional Council and farmers as stated by focused control, 
standard control, and non-TB farmers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1 77 
Table 30. Attitudes towards TB control by Wairarapa focused control, standard ontrol, and 
non-TB farmers (percentages of farm group in brackets) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1 8 1  
Table 3 1 .  Importance of farmer conducted vector control as seen by the three farm groups -
focused control, standard control, and non-TB farmers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . .  1 83 
Table 32. Organisations considered by Wairarapa herd managers to be responsible for 
eradicating TB (in brackets the number of farms that ranked these organisations with the 
highest priority). The last column gives the percentages of all farmers that nominated this 
organisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . ... . . . . . . . . . . .  1 84 
Table 33. Organisations considered by Wairarapa herd managers to be responsible for doing 
the actual work to eradicate TB (in brackets the number of farms that ranked these 
organisations with the highest priority). The last column gives the percentages of all 
farmers that nominated this organisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . .. . . . .  1 85 
Table 34. Estimated costs of TB and its control to Wairarapa focused and standard control 
farms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1 86 
Table 35. General farm characteristics of National focused and standard control farms . . . . . . . . . . . . . . .  1 88 
Table 36. Average farmed area stratified by herd type and national farm group (with range in 
brackets) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1 89 
Table 37. Livestock units and stock densities of cattle and deer on national focused and 
standard control farms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .  1 90 
Table 38. Information on herd managers for national focused and standard control farms . . . . . . . . . . .  1 9 1  
Table 39. Information on stock movements on and off farms for national focused and standard 
control farms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . .. . . . . . . . . . . . . . . . . . . . . . . . .  1 93 
Table 40. Information on the TB situation and perception of national focused and standard 
control farms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . .  1 94 
Table 41.  Assumed likelihood of contact between possums/ferrets and livestock, as indicated 
by national focused and standard control farmers . . . . . . . . . . . . . .. . . . . . . .. .... . ... .............. ........ . . . .... . . .... 1 95 
Table 42. Vector control by Regional Council and farmers as stated by national focused and 
standard control farmers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . .  1 96 
Table 43. Attitudes towards TB control by national focused and standard control farmers 
(percentages of farm group in brackets) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . .  1 98 
Table 44. Importance of farmer conducted vector control as seen by focused and standard 
control farmers . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . ..... . . ... . . . . . . . ..... .......... . . . .... . . . . . . . . ...... ....... ... . . .... .... ....... 1 99 
Table 45. Organisations considered by national herd managers to be responsible for 
eradicating TB (in brackets the number of farms that ranked these organisations with the 
highest priority). The last column gives the percentages of all farmers that nominated this 
organisation . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . .. .. . . .. . . . . . . . . . . . . . . .  200 
Table 46. Organisations considered by national herd managers to be responsible for doing the 
actual work to eradicate TB (in brackets the number of farms that ranked these 
organisations with the highest priority). The last column gives the percentages of all 
farmers that nominated this organisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  20 1 
Table 47. Estimated costs of TB and its control to national focused control and standard 
control farms . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . .... . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . . . . . .. ... . . . . . . . .  202 
Table 48. Some characteristics of farms in the Hawke's Bay-Wairarapa District (from 
Ministry of Agriculture and Forestry, 1 998) . ................ ....................... ...... . . . . . . . . ..... . . . ............ 2 1 7  
xvii 
Table 49. Summary of data and assumptions used in partial budgeting of implementing on-
farm control programme on the three farm types in the Wairarapa . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . .  221 
Table 50. The six scenarios analysed in the study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  222 
Table 5 1 .  Distribution parameters for input variables for costs and returns used in the 
stochastic @ RISK partial budgeting model for the three farm types in the Wairarapa . . . . . . . .  224 
Table 52. Assumed probabilities of reducing reactor numbers per farm if conducting on-farm 
TB control or not . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  226 
Table 53. Expected economic outcomes of the partial budgeting for reducing the number of 
reactors in three different farm types (using current compensation of 65%) . . . . . . . . . . . . . . . . . . . . . . . . .  227 
Table 54. Expected economic outcomes in partial budgeting of reducing the number of 
reactors in the three different farm types with zero compensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  230 
Table 55. Expected net returns with subsidy to cover costs of poison and bait stations for beef 
breeding farms and additional subsidies necessary to achieve net gain in all and all 
except one scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  231 
Table 56. Expected net returns with subsidy to cover costs of poison and bait stations for beef 
finishing farms and additional subsidies necessary to achieve net gain in all and all 
except one scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  231 
Table 57. Expected net returns with subsidy to cover costs of poison and bait stations for dairy 
farms and additional subsidies necessary to achieve net gain in all and all except one 
scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  232 
Table 58. Descriptive statistics for the probability distributions of the difference between 
returns and costs resulting from simulation modelling of the six different scenarios for 
dairy farms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  234 
Table 59. Expected financial values of adoption and non-adoption of on-farm control 
measures on dairy farms under different levels of compensation and slaughter levies with 
different starting numbers of reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  24 1 
Table 60. Expected financial values of adoption and non-adoption of on-farm control 
measures on beef breeding farms under different levels of compensation and slaughter 
levies with different starting numbers of reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  244 
Table 6 1 .  Yearly off-MC-payments and subsidies for material and labour necessary to make 
'adoption' of on-farm control programmes the preferred option on beef breeding farms 
under 65% and zero compensation for reactor animals, stratified for reactor starting 
situations with five, two and one reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  246 
Table 62. Expected financial values of adoption and non-adoption of on-farm control 
measures on beef finishing farms under different levels of compensation and slaughter 
levies with different starting numbers of reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  247 
Table 63. Yearly off-MC-payments and subsidies for material and labour necessary to make 
'adoption' of on-farm control programmes the preferred option on beef finishing farms 
under 65% and zero compensation for reactor animals, stratified for reactor starting 
situations with five, two and one reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  248 
Table 64. Subsidies used for the three farm types under two different amounts of off-MC­
payments, stratified for starting reactor numbers under 40% and 65% reactor 
compensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  249 
Table 65. Regional cost for providing vouchers for control work and two different off-MC-
payments under two different compensation levels for reactor animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  250 
Table 66. Expected financial outcome of adopting on-farm control measures for the three herd 
types and the different reactor scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  259 
Table 67. Economic outcomes of deterministic and stochastic partial budgeting on adopting 
on-farm control methods for TB, stratified by farm types, compensation level and 
reduction of TB reactor numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  263 
Table 68. Outcomes of decision analysis whether to adopt on-farm control programmes or not, 
stratified by farm types, reactor compensation level and number of reactors to start with . . . . 264 
Table 69. Liveweight targets for dairy cows and heifers (in kg), used in FarmORACLE . . . . . . . . . . . . .  279 
Table 70. Liveweight targets for beef cattle and sheep (in kg), used in FarmORACLE . . . . . . . . . . . . . . . .  280 
xviii 
Table 71 .  Production outcomes in modelling traditional (base) and alternative grazing plans 
on Dairy fann A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . .  288 
Table 72. Production outcomes in modelling traditional (,Base List' and 'Base High') and 
alternative grazing plans on Dairy fann B. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  290 
Table 73. Economic comparison between grazing plans used on Dairy farm B . . . . . . . . . . . . . . . . . . . . . . . . . . . .  291 
Table 74. Summary of cattle and sheep liveweights for sale using different grazing plans . . . . . . . . . . .  296 
Table 75. End liveweights of cattle and sale ewes for three different grazing plans . . . . . . . . . . . . . . . . . . . . . .  299 
Table 76. Difference in carcass weights and economic outcome comparing different grazing 
plans on the beef finishing farm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  300 
Table 77. Comparison of grazing regimes using 800 versus 1600 kg DMlha residual . . . . . . . . . . . . . . . . . .  301 
Introduction 
A guide to the methodologies and rationale for research 
presented in  this  thesis 
1 
2 
3 
Bovine tuberculosis (TB) is a disease of farmed cattle globally, which has been eradicated from 
many countries within the last century. The eradication is complicated in some countries by the 
existence of wild animal reservoirs for tuberculosis, that are able to transmit the disease back to 
domestic animals. Such reservoirs of TB have been found in badgers in England and Ireland 
(OReilly and Daborn, 1995; Hughes et al. , 1996), in Cape buffalo in South Africa (OReilly and 
Daborn, 1995), in cervids in North America (Schmitt et al. , 1997) and in several wildlife species in 
New Zealand such as farmed and feral deer (Nugent, 1994), sheep (Cordes et al. , 1981), feral pigs 
(Wakelin and Churchman, 1991) ,  feral cats (Lugton, 1997), ferrets (Walker et al. , 1993), possums 
(de Lisle et al. , 1993; Jackson et al. , 1995a; Jackson et al. , 1995b). Amongst the wildlife species, 
the Australian brushtail possum (Trichosurus vulpecula) is considered the main vector for 
tuberculosis in New Zealand (Morris et al. , 1994; Jackson, 1995).  
Tuberculosis control therefore not only has to include the management and control of TB in 
livestock (cattle and deer), but also in the main vector species, possums. Areas where possums are 
known to be infected with tuberculosis are termed Vector Risk Areas (VRA) while those, where no 
tuberculosis has been found in vector species are termed Vector Free Areas (VFA). The Vector 
Risk Areas receive the majority of vector control efforts with the main control method being to 
reduce the possum population. 
The Animal Health Board (AHB), the national organisation for the control and eradication of 
bovine tuberculosis, produced a national pest management strategy (PMS) for the five-year period 
from 1995 in accordance with provisions of the Biosecurity Act 1 993 (Animal Health Board, 
1 995). In the following, some of the issues covered by the PMS are summarised: 
Impacts on human health, on the health and productivity of farmed cattle and deer and the impact 
on New Zealand' s  export trade in beef, dairy and venison products were cited as the actual or 
potential negative impacts of tuberculosis. The impact on human health is considered low due to an 
overall low incidence of TB, sound slaughterhouse inspection and pasteurisation of milk. If the 
disease was not controlled, negative effects on the farm productivity could also occur due to TB 
being a chronic, usually fatal disease. Under current practice the disease generally is detected early 
and the animals removed from the herd. Therefore, the impact on farm productivity under current 
practice is associated principally with control measures - by disrupting farming practices and 
compulsory slaughtering of infected or suspected tuberculous animals. The impact on New 
Zealand' s export trade of beef, dairy and venison products, which is worth approximately $5 billion 
per year, is considered as the most serious aspect of tuberculosis. The incidence of bovine 
tuberculosis in New Zealand is higher than in many other countries, therefore potentially creating a 
disadvantage for New Zealand in the current international trading environment. 
4 
Therefore, the long-tenn goal of the AHB is "to eradicate bovine tuberculosis from New Zealand". 
However, the report stated that "eradication is not a realistic possibility within the tenn of the 
strategy with current technology. Therefore the primary focus of the five year strategy is on the 
reduction, and where technically feasible, elimination of the transmission of M. bovis to and within 
domestic livestock, specifically cattle and farmed deer" (Animal Health Board, 1995). 
The presented thesis addressed a national requirement and aimed to assist the AHB by assembling 
infonnation required to further succeed in the control and eradication of TB. The following key 
points of the AHB strategy provided the framework for the studies presented in this thesis. 
Four objectives were stated for the five-year period: 
to reduce the number of infected herds in TB Vector Free Areas from 0.7 percent to 0.2 percent 
of the total herds in those areas, 
to prevent the establishment of new TB Vector Risk Areas and/or the expansion of existing TB 
Vector Risk Areas into fannland free of TB vectors, 
to decrease the number of infected herds in TB Vector Risk Areas from 17 percent to 1 1  
percent of the total numbers of herds in those areas, 
to encourage individuals to take action against TB on their properties and in their herds. 
Disease control and vector control are proposed as two main areas for achieving these objectives. 
The strategy states that "managing and eliminating disease risk from wild animals through 
increased vector control operations is critical to the success of the strategy" (Animal Health Board, 
1 995).  At the time of publishing the strategy, an increase in vector control expenditure from $ 1 8  
million in 1995/96 to $3 1 million in 1999/2000 was expected. Following priorities were listed for 
vector control (Animal Health Board, 1995): 
Establishment of protection zones to prevent leakage of infected vectors from infected areas. 
Encouragement and assistance to farmers in infected areas to improve their disease status 
through: 
Assistance with development and implementation of regional and locally initiated vector 
control programmes. 
Reclassification of areas as vector risks are reduced. 
Assistance to farmers with high risk herds. 
5 
Self-help programmes. 
One-on-one programmes.  
Commitment to control of infected wild animals on Crown land adjacent to farmland. 
This thesis presented addresses the priority 'encouragement and assistance to farmers to improve 
their TB status' .  The general objective of the thesis was to test whether intensive management 
advice and its adoption could accelerate the control and eradication of TB from individual farms, 
particularly those with persistent TB problems. 
• Chapter one: Literature review of the field of human behaviour change and agricultural 
extension, to examine how change can be achieved and what factors limit the scope for change 
in practice. 
• Chapter two: Interview survey to establish management, practices and attitudes of farmers who 
own or manage TB-infected farms, in order to establish baseline data which would be 
incorporated into the intervention study. 
• Chapter three: Intervention study of using on-farm TB control programmes on farms in the 
Wairarapa region, in order to evaluate the effectiveness of on-farm TB control programmes 
under field conditions. 
• Chapter four: An evaluation of the national one-on-one TB programme, undertaken by 
AgriQuality, the state veterinary service, and Agriculture New Zealand, an agricultural 
advisory company, and comparison with the Wairarapa programme. 
• Chapter five: Questionnaire analysis, comparison of focused control farms (receiving 
interventions) and standard control farms (receiving no interventions) for both projects 
analysed separately; farmers' attitudes towards TB and its control. 
• Chapter six: Economic analysis of TB and on-farm TB control programmes on farms in the 
Wairarapa, in order to assess the cost effectiveness of such programmes and propose potential 
motivational incentives. 
• Chapter seven: FarmORACLE, a computer model that allows modelling of each specific farm 
situation and compares different farm management policies (such as different grazing 
strategies), in order to facilitate and assist decision making in relation to TB control. 
6 
A final discussion of the results of all studies presented in this thesis and a discussion on the 
application of findings in future TB control programmes is presented in the General Discussion. 
This thesis has been written in the form of a series of papers. The papers will be submitted in 
adjusted form following the submission of this thesis. 
Bibl iography 
Animal Health Board ( 1995). National TB Strategy: proposed pest management strategy for bovine 
tuberculosis. Animal Health Board, Wellington, New Zealand. 
Cordes, D.O., Bullians, J.A., Lake, D.E. and Carter, M.E. ( 198 1 ). Observations on tuberculosis 
caused by Mycobacterium bovis in sheep. New Zealand Veterinary Journal, 29: 60-62. 
de Lisle, G.W., Crews, K., de Zwart, J. et al. ( 1993). Mycobacterium bovis infections in wild 
ferrets. New Zealand Veterinary Journal, 41 (3): 148-149. 
Hughes, M.S., Neill, S.D.  and Rogers, M.S. ( 1996). Vaccination of the badger (Meles meles) 
against Mycobacterium bovis. Veterinary Microbiology, 5 1  (3-4): 363-379; 1 14 ref. 
Jackson, R ( 1995). Transmission of tuberculosis (Mycobacterium bovis) by possums. Unpublished 
PhD thesis, Massey University, Palmerston North, New Zealand.  
Jackson, R, Cooke, M.M., Coleman, J.D. and Morris, RS. ( 1995a). Naturally occurring 
tuberculosis caused by Mycobacterium bovis in brushtail possums (Trichosurus vulpecula): 1. 
An epidemiological analysis of lesion distribution. New Zealand Veterinary Journal, 43 (7): 
306-3 14. 
Jackson, R, Cooke, M.M., Coleman, J.D., Morris, RS., deLisle, G.W. and Yates, G.F. ( 1995b). 
Naturally occurring tuberculosis caused by Mycobacterium bovis in brushtail possums 
(Trichosurus vulpecula): Ill. Routes of infection and excretion. New Zealand Veterinary 
Journal, 43 (7): 322-327. 
Lugton, I. W. ( 1997). The Contribution of Wild Animals to the Epidemiology of Tuberculosis in 
New Zealand. Unpublished PhD Thesis, Massey University, Palmerston North. 
7 
Morris, R.S . ,  Pfeiffer, D .U. and Jackson, R. ( 1 994). The epidemiology of Mycobacterium bovi s  
infections.  Veterinary Microbiology, 40 0) :  1 53- 177.  
Nugent, G. ( 1 994) . Deer - Perspectives of the future. In :  National Science Strategy Committee 
Workshop on Ecology and Epidemiology of Bovine Tuberculosis and its Wildlife Vectors .  
Unpublished. 
O'ReiUy, L.M. and Daborn, C.J. ( 1 995). The Epidemiology of Mycobacterium bovis Infections in 
Animals and Man: A Review. Tubercle and Lung Disease, 76 (Supplement 1 ) :  1 -46. 
Schmitt, S .M. ,  Fitzgerald, S .D. ,  Cooley, T.M. et al. ( 1997) .  Bovine Tuberculosis in free-ranging 
White-tailed Deer from Michigan. Journal of Wildlife Diseases, 3 3  (4) : 749-758 .  
Wakelin, C .A .  and Churchman, O.T. ( 199 1 ) . Prevalence of  bovine tuberculosis in feral pigs in  
Central Otago. Surveillance, 18  (5 ) :  19-20. 
Walker, R., Reid, B. and Crews, K. ( 1993). Bovine tuberculosis in predators in the Mackenzie 
Basin. Surveillance, 20 (2) : 1 1 - 14. 
8 
Australian Brushtail Possum 
Farm in the Wairarapa 
9 
Chapter 1 
Changing Behaviour:  a l iterature review 
10  
1 1  
Introduction 
In order to get farmers to take more responsibility in disease control programs, such as programmes 
for tuberculosis (TB) control in New Zealand, two main factors are necessary: First that farmers 
have the required tools (knowledge, understanding and methods) available to them; and second that 
they then adopt and implement these tools and methods. 
Ashby wrote in 1926: 
"If we want to know how or why a farmer acts in a certain way or how to induce him to act 
in a certain way, we have to enquire why men act, and especially why men act as they do 
when they live in the sort of social environment and general circumstances in which 
farmers live." 
Therefore it was necessary to look into the field of behaviour change and methods applied, 
especially in agriculture, to understand this change. This chapter reviews the published literature on 
the process of change, and the behaviour involved. Two examples are reviewed in more detail of 
how these behaviour changes have been applied in real live situations: one example in the human 
health field of smoking cessation, and the second example in agricultural extension. Extension is a 
term commonly used in agriculture for the process of transferring information from 
scientists/industries and others to the end users of the information (Morris et aI., 1995) .  
Behaviour Change 
Any behaviour change involves processes at the cognitive level, which can be stimulated by 
information channels and sources .  Through the diffusion of new ideas or innovations eventually a 
behaviour change can be achieved on a broad basis throughout the populations or social system. 
Cognitive process 
Behaviour is guided by the physical, social and economic environment of the individual. Each 
individual' s personal characteristics (such as beliefs, values ,  opinions, norms, objectives, 
expectations, attitudes, habits, and intellectual and physical capabilities) play a role in that 
individual' s behaviour. These characteristics lead to the way an individual views the world, which 
can be considered to be the individual' s  mode of cognition. Cognition and environment provide 
opportunities and restrictions to activity (Triandis, 197 1). Each individual has a tendency towards 
12 
consistency. If there is inconsistency, cognitive and behavioural processes start with the aim of 
reducing this inconsistency and re-establishing consistency. The reduction of inconsistency can 
either be obtained through manipulating cognition, by suppressing, rejecting or rationalising 
information and not changing the behaviour, or through a change of behaviour. 
The process of behaviour change is a complex cognitive process including four stages: need 
recognition '" search for information '" evaluation '" choice CFairgray, 1979; Rogers, 1 995; 
Steenkamp, 1997). 
Fairgray ( 1979) and Rogers ( 1 995) provide a review on the literature of this process, which in the 
following is elaborated. The initiating factor in behaviour change is a challenge to the existing state 
of satisfaction with current behaviour. This can occur when events or information suggest a 
disadvantage of the present practice or the non-use of an alternative one. Generally, the fIrst 
reaction to such information and the challenge is to ignore it, reject it as being irrelevant or 
inaccurate, or change the information in a way that it will be consistent with current 
behaviour/attitudes. However, if the information is accepted, it creates a dissonance between the 
existing attitude and the new knowledge, an inconsistency between the old belief Cs) and the new 
belief Cs), that the alternative practice should be used. Mostly the acceptance of the information 
comes from close evidence or from a trusted source. 
Once the individuals have recognised the challenge, they will assess alternatives that are available. 
This step involves drawing from the person's own experience and ideas, as well as seeking 
additional information. The objectives and capabilities determine the viable alternatives, whereas 
the importance of the challenged behaviour determines the extent of the search for new 
information. The potential gains and losses of each alternative are compared with those of the 
others and mainly with those of the existing practice. Main determinants in this process of 
evaluation are the perceived risks and uncertainties. The existing practice/behaviour has advantages 
in terms of being established and having less risk of failure than a new practice. Therefore it 
requires more effort to make a decision to change than to decide to keep the existing practice. For 
each situation the costs being saved and/or risks avoided by not changing are compared against the 
anticipated advantages to be gained by changing. The level where the fIrst outweighs the latter one 
is called the threshold level, which depends on factors such as the risks and costs of the alternative 
option involved, the scale of change and the satisfaction with the existing practice (Fairgray, 1979). 
At the end of this stage of persuasion or evaluation the individual makes a tentative decision to 
adopt or reject the new alternative and tries to confIrm this decision. If the decision was made to 
change the individual seeks information that supports this and ignores, if possible, information that 
contradicts the decision. The last stage of the decision-making process towards change is the 
13  
adherence to the decision, where the behaviour i s  confirmed and the challenge resolved or 
suppressed. Therefore decision making can result in a behaviour change, a reinforcement of the 
existing practice, or a situation where the challenge is accommodated and accepted, but no 
behaviour change occurs because no decision is being made (Rogers, 1995). 
For an 'extension agent' , anybody conveying information to others, there are two critical points in 
the decision-making process: the challenge to existing practices and the evaluation stage. The agent 
has to challenge the existing practice by challenging the existing attitudes and providing new 
information in favour of the new practice/innovation. The individual' s objectives are most 
important. Often the individual has to be made dissatisfied with the existing practice by pointing 
out alternatives that better achieve the individual's  goals or that achieve higher or alternative 
objectives. The resulting dissonance has to be strong enough to be too uncomfortable to be ignored 
or suppressed, which occurs when the perceived advantages outweigh the cost of the adoption 
(Fairgray, 1979). 
Information channels and sources 
In order to change behaviour, information on the innovation must be available, understood, 
accepted, retained and acted on. The acceptance of information however depends very much on the 
source of it. 
Three types of information channels are available: personal direct, personal indirect and impersonal 
indirect. The personal direct information channels are the five senses (sight, sound, taste, touch, 
smell), whereby the information is sifted before, during and after the receipt of the message. The 
information is also prone to distortion through selective attention, limited capacity of the nervous 
system to receive information and further alteration after it has been received and understood 
(Fairgray, 1979). 
If the information comes through another person, such as by face-to-face contact or by telephone, it 
is termed personal indirect information channel. Impersonal indirect channels include letters, 
books, journals, newspapers, radio, television and more recently the Internet. These three 
information channels related strongly to the source of information: intrapersonal, interpersonal and 
media (Schrarnm, 1973). 
Six characteristics were identified by Schrarnm (Schrarnm, 1973) as affecting the utility of 
information channels: a message is more likely to be received 
• the more senses it affects, 
14 
• the more control the receiver has over the speed of the delivery, 
• if the message is in permanent form, 
• if it is easily understood, 
• the higher the multiplicative power of the channel (determines the number of people 
reached and the area covered), and 
• the greater the opportunity for immediate feedback between source and receiver. 
Interpersonal communication is generally more effective in obtaining acceptance of a message than 
mediated communication, as interpersonal communication has the additional power of persuasion. 
There are three processes, whereby information can be accepted (Kelman, 1967): through 
compliance, if the change agent has authority over the recipient; through identification, where the 
agent has attractiveness, such as prestige, experience, whereby the content of the information may 
be irrelevant; or through internalisation, if the contents of the message seem logically correct and 
rewarding and therefore warrant acceptance. For the third one, the agent needs credibility, such as 
expertise, objectivity, success and proof. In many extension services the first process, compliance, 
is unimportant. 
Brief overview of adult education 
The rate of learning is a major factor in the adoption of innovations and achieving change. Only if 
an individual's conceptual framework is changed, new ideas can be applied. Thus, in order to 
achieve change, it is not enough to provide purely information and knowledge, but also 
opportunities for learning (Stantiall, 2000). Learning, the continued updating with new information, 
is an essential element of today's  knowledge (Barr, 2000). Information and knowledge are two 
distinct concepts. Information can be shared, stored, communicated, while knowledge is within the 
minds of each individual (Stantiall, 2000). 
The process of adopting innovations depends on learning, and an individual' s  ability to learn is 
enhanced by strategies that conform with the learning style preferred by this individual (Kolb, 
1984). There are two main parts to learning, first to grasp the information, either through 
conceptualising or through experience and the second part is the transformation of the information 
into knowledge, either through reflection or through experimentation (Kolb, 1984). Individuals can 
be grouped into four distinct theoretical learning styles, depending on how they grasp information 
and transform it into knowledge: Accomodator who learns by experience; diverger who learns by 
weighing up different perspectives; converger, who learns by putting theories into practice; and 
1 5  
assimilator, who learns by sorting information into concise logic (Kolb, 1984). In practice, learning 
styles are influenced by past and present experiences. 
Lewin, a psychologist in the first half of the twentieth century did a lot of work on experiential 
learning. It will result in altered cognitive structures ,  modified attitudes and expanded range of 
behavioural skills. Within experimental learning there are several principals: people will believe 
more in knowledge they have discovered themselves than in knowledge presented by others; the 
more supportive the social environment is, the more likely a person is to experiment with new 
behaviours/attitudes; and learning is more effective when it is active than passive. (Jackson and 
Caffarella, 1994; Ballantyne and Packer, 1996; Botha, 2000). In participatory learning the learner is 
involved throughout the learning process, even at the design stage. Participatory learning empowers 
the individual, resulting in 'education for change' rather than 'education for adaptation' (Peet and 
Peet, 1995). 
As individuals may have different conceptions when dealing with a learning task, they will learn 
different things from the same event and apply their understanding differently (Ballantyne and 
Packer, 1996). 
Innovation diffusion model 
In order to achieve behaviour change on a broad basis, the concepts have to be disseminated 
throughout the population. The main model used to describe this process is the innovation diffusion 
model, as described by Rogers ( 1995) .  It is the process whereby an innovation is diffused or 
communicated to populations or social systems over time. Spontaneous diffusion is distinguished 
from directed or managed diffusion (Rogers, 1995) .  If the spread of information is unplanned it is 
called a spontaneous diffusion, while directed or managed diffusion describes the deliberate 
attempt to spread the innovation or new idea. Because of the directed diffusion this theory has 
found wide interest and application. The theory involves the innovation process, innovations 
themselves and their characteristics, the adoption of innovations and the communication channels 
or change agents. 
Innovation process 
The entire innovation process can be divided into two main parts: Firstly, the innovation­
development process comprises all decisions and activities and their impacts that result from 
recognition of a need. Included are research (basic as well as applied), development and 
16 
commercialisation of the innovation through diffusion and adoption by the users and finally its 
consequences. 
The second part is the innovation-decision process an individual goes through: Knowledge � 
Persuasion � Decision � Implementation � Confirmation (Rogers, 1 995). In the beginning the 
individuals gain a first knowledge of the innovation, then they form an attitude towards this 
innovation and decide then to adopt it or to reject it. If they adopt it, they will implement the 
innovation and confirm the decision (Rogers, 1995). 
A similar model of stages was proposed by Prochaska et al. (1992), cancer prevention researchers, 
who investigated how individuals change an addictive behaviour. Their 'transtheoretical' five-stage 
model was first proposed in 1983 and is now widely used in the human public health field to 
explain the adoption of preventive health innovations, such as dietary measures (Ni Mhurchu et aI. , 
1997; Spomy and Contento, 1995), quitting cocaine, smoking, adolescent delinquent behaviour, 
safe sex, contraception, use of sunscreen and others (Prochaska et al. , 1994). The first stage is 
precontemplation, where the individual is aware of the problem and starts thinking about 
overcoming it. The second stage is contemplation, where the individual is seriously thinking about 
overcoming the existing problem, but they have not made a commitment yet. In the third stage, the 
preparation stage, the individual intends to make a commitment in the near future, but has not done 
so. The fourth stage is the action stage, where the individual actually changes the behaviour or the 
circumstances in order to overcome the problem. The fifth stage is maintenance, at which the 
individual consolidates and continues the behaviour change. Individuals generally cycle through 
these five stages several times before overcoming the addiction (Prochaska et aI. , 1992). It is 
important to assess the stage of an individual' s readiness for change and to tailor interventions 
accordingly. If the majority of individuals are not yet in the action stage, then action-oriented 
programmes will only serve a minority of the population (Prochaska et aI. , 1992). 
All these stages have to be passed in sequence by the individual, and the communication channels 
have to be aligned with the stages in the innovation-decision process (Rogers, 1995). More recent 
modifications of the linear model account for the recycling through the stages in order to account 
for relapses (Brownell et al. , 1986). 
Characteristics and adoption of innovations 
There are five main characteristics of innovations which are critical to their adoption (Rogers, 
1995; Greer and Greer, 1996): relative advantage, compatibility, complexity, trialability and 
observability . 
17 
The relative advantage as it is perceived by the user, can be expressed in economic tenns, social 
prestige or other benefits. Economics might be the most important predictor for the rate of 
adoption, but Rogers regards it as unlikely that it is the sole predictor (Rogers, 1 995). For certain 
innovations like clothing fashion, social prestige is almost the only benefit for the adopter. 
Economics were found to be less important to peasant farmers in the Third World, where the 
farmers attached greater importance to social approval than to financial return (Fliegel et aI. , 1968). 
Preventive innovations have particularly slow rates of adoption because individuals have 
difficulties perceiving the advantage, partly because the consequence is somewhere in the distant 
future. The rewards of adoption are not only delayed but also uncertain and therefore also rely on 
the scale of risk perceived by the individual. Someone practising safe sex in order to prevent 
infection with mv I AIDS might not have contracted it even without adopting safe sex (Rogers, 
1995). 
Incentives are often used to accelerate the rate of adoption of innovations by increasing the degree 
of relative advantage. They have been used in many fields, such as agriculture, human health and 
family planning. A wide range of different types of incentives exists (Rogers, 1973 quoted in 
Rogers, 1995): incentives that are paid to adopters directly, incentives that are paid to individuals 
that persuade others to adopt, positive and negative incentives, monetary and non-monetary 
incentives, immediate and delayed incentives. Rogers (1995) drew the following conclusions: 
incentives increase the rate of adoption of innovations, incentives can change the group of 
individuals that adopt the innovations (individuals who would not have adopted otherwise). 
Incentives also increase the number of adopters. However, the quality of adoption might be lower 
than desired, thus limiting the intended consequences, because if individuals only adopt in order to 
get the incentive, there is less motivation to continue to use it. Rogers ( 1995) proposes that the 
effectiveness of incentive policies can be improved if empirical studies evaluate the effects first. 
At a national level, if incentives are not taken up at a rate which Government desires, it can use 
legislation and enforce certain changes, such as the People' s Republic of China having mandated 
the one-child family. 
Compatibility is the degree to which individuals see the innovation as consistent with existing 
sodo-cultural values and beliefs, with past experiences (previously introduced ideas) andlor client 
needs for the innovation. The latter is especially important in Third World Countries. Often the 
indigenous knowledge systems are not recognised or are considered inferior by change agents that 
come from developed countries. Ignoring these knowledge systems can not only miss important 
local knowledge, but also lead to a decreased rate of adoption (Rogers, 1995). The more compatible 
an innovation is, the greater its adoption. 
1 8  
The perception of individuals on the degree of difficulty to understand and use innovations falls 
under the category 'complexity' .  Generally, the more complex an innovation, the slower its rate of 
adoption (Rogers, 1995). 
Trialability is personal experimentation with an innovation to find out how it works under the 
individual's  own circumstances. The easier it is to trial the innovation, the greater the adoption, 
especially for the early adopter category (Rogers, 1995). 
Observability i the degree to which the results of the innovation are visible to others; it i 
positively related to its rate of adoption (Rogers, 1 995). 
Change agent 
A change agent is an individual that influences clients' innovation-decision process towards a 
direction desired by the change agency. A change agent usually tries to ensure adoption of 
innovations. However, they might also attempt to prevent the adoption of certain innovations with 
undesirable effects (Rogers, 1995). They intend to be a communication link between the resource 
system and the client system. This communication has to be two-way. Feedback from the client 
system has to go through the change agent back to the change agency, so that latter one can adjust 
their programmes in order to meet the needs of their clients. Rogers ( 1995) defines seven roles for 
the change agent in the process of introducing innovations: ( 1)  To help develop a need for change 
by pointing out alternatives to existing problems and convince the individual that they are capable 
of confronting the problem; (2) To establish an information-exchange relationship, which can be 
enhanced by being perceived as credible, competent, trustworthy, and by empathising with clients' 
needs and problems. Before clients will accept the innovation, they must accept the change agent 
who promotes it; (3) To diagnose problems, where the change agent has to analyse the client' s 
problems to find out why existing alternatives do not meet their needs; (4) To create an intent in the 
client to change. In this phase the change agent tries to motivate clients' interest in the innovation; 
(5) To translate an intent to action. Peer-networks are the most important influence in this stage, 
and the change agent can only operate indirectly by working with opinion leaders; (6) To stabilise 
adoption and prevent discontinuance, through giving reinforcing messages while the clients are at 
the implementation or confirmation stage of the innovation-decision process; (7) To achieve a 
terminal relationship, where the change agent tries to develop the clients' ability to be their own 
change agents (Rogers, 1995). 
The success of a change agent increases with the amount of effort spent in communicating with 
clients, but the timing of the client contact has to depend on the stage of diffusion the clients are in. 
Another positive relationship was found between agent' s success and the client orientation (versus 
1 9  
a change-agency orientation). Such change agents are regarded a s  having higher credibility and 
obtain a higher feedback. The probability of success is also positively related to the compatibility 
with clients' needs, the change agent' s empathy with the clients' needs, the credibility in the 
clients' eyes and the similarity between the change agent and the client. From the latter relationship 
Roberts concludes that change agents have most contact with clients who are most like themselves 
and therefore change agent contact is positively related to a higher social status, higher education 
and cosmopoliteness amongst clients. This means there is a more effective communication between 
the two parties the more similar they are. However, this leads to a circle of relationships where the 
change agents help those people most that least need their help (Rogers, 1995). 
In order to encourage the less-educated clients, aides are necessary. These people are not fully 
professionals, but help to bridge the gap between the professionals and the clients. They have less 
credibility regarding their competence, but greater credibility as being trustworthy (Rogers, 1995). 
Consequences of innovations 
The consequences of innovations are often inadequately investigated, firstly because often the 
change agencies assumed the consequences to be positive, secondly because they are difficult to 
measure and the usual survey research methods may be unsuitable. The consequences can be 
desirable or undesirable. However, the consequences often have both outcomes. Consequences can 
also be direct or indirect, anticipated or unanticipated. The consequences of diffusion of innovation 
usually widens the gap in socio-econornic status between the early and the late adopters, and often 
the gap between rich and poor in the system (Rogers, 1995). However, this can be avoided by 
specifically targeting certain groups within the system (Rogers, 1 995). 
Decentralised diffusion systems 
In the classical diffusion model, Government officials or technical experts decided on what 
innovations had to be diffused, what channels to be used and to whom the innovation was diffused 
(Fairgray, 1979). The diffusion is from top to bottom, with the individual adopter of the innovation 
being a passive acceptor. In recent decades decentralised diffusion systems were being recognised, 
where innovations originated from the operational levels of a system (Rogers, 1986). These 
innovations were then spread horizontally through peer network with modifications along the line, 
in order to suit their particular circumstances. These decentralised diffusion systems are generally 
not run by technical experts, but are client controlled, where the adopters often serve as their own 
change agents (Rogers, 1986; Rogers, 1995). These systems are designed to stimulate innovation 
because it enhances the diversity of approaches (Sherman and Schultz, 1998). However, 
decentralisation also can lead to decreased efficiency in resource usage. The innovation is divided 
20 
into smaller parts, possibly overlapping, and thus resources might be spent on overlapping projects 
(Banks and Harley, 2000). 
Human behaviour change: the smoking example 
Smoking is  strongly related to the occurrence of lung cancer (Stellman and Garfmkel, 1989), 
increased respiratory illness among children whose mothers smoke (Stoddard and Gray, 1 997). It is 
also a major risk factor for heart diseases and stroke and affects foetuses. In smoking there are four 
stages - initiation, maintenance, cessation and resumption or relapse (Lichtenstein, 1982). 
Psychosocial factors are important in all four stages, while the pharmacological effects of nicotine 
only start to appear in the second stage. 
Because of the negative effects of smoking, many health professionals encourage the cessation of 
smoking. Most smokers know that smoking is harmful and many try to overcome the addiction 
(Leshner, 2000). Individuals overcome the addiction with and/or without the help of professional 
treatment programs. Early investigations of smoking assumed a linear model of behaviour change 
stages, where the individuals pass through the different stages of precontemplation, contemplation, 
action and maintenance (DiClemente and Prochaska, 1982). However, most people who actually 
take action to overcome the addiction, do not maintain it, and therefore have several action 
attempts before they become long-term smoke-free (Norcross and Vangarelli, 1989; Cohen et al. , 
1 989). Prochaska et al. ( 1992) concluded that the linear progression through the stages of the 
model of change is possible, but relatively rare with addictive behaviour. Often the individuals 
relapse and regress to an earlier stage in the model, but not necessarily back to where they started, 
which led to the development of the spiral model, whereby the success of quitting was not related 
to the number of previous attempts (Cohen et al., 1989; Prochaska et al. , 1992). In an earlier study 
they found that 85% of smokers return to the contemplation or preparation stages (Prochaska and 
DiClemente, 1984). However, some individuals feel like failures and refuse to think about 
behaviour change, therefore returning to the precontemplation stage. This transtheoretical model of 
stages has been subject to several critiques, especially as any change is a continuum rather than a 
process through discrete stages (Sutton, 1996; Davidson, 1998). Nevertheless, the model is still 
considered extremely valuable within the addiction field, as it is practical, intuitive appealing to 
several theoretical orientations and as it includes motivation (Davidson, 1 998). 
Quitting smoking is not a single event, but a process and there are many ways to cessation. There is 
much debate about the superiority of self-quitting versus professional treatment programmes for 
smokers. Schachter (1982) suggested that individuals attempting to quit smoking by themselves 
21  
had a higher success rate than individuals who attended fonnal programs. However, evaluations of 
professional programmes only look at a single attempt to quit smoking, whereas retrospective 
studies on self-quitting look at the result of mostly multiple attempts to quit (Cohen et aI. , 1989). In 
a study looking at 10 long-tenn prospective studies on self-quitting it was found that self-quitting 
had a similar or lower success rate than fonnal programs. 
Excellent action-oriented treatment and self-help programmes were designed and available, but 
professionals are disappointed with the percentage of smokers taking these programmes up. 
Schmid et al. ( 1989) reported on four different recruitment strategies for home-based smoking 
control intervention programs, but found that only 0. 1 to 5% of smokers enrolled, indicating that 
the vast majority of smokers are not at the action stage yet. Di Clemente and Prochaska ( 1998) 
estimated from a range of studies on smoking behaviour, that 20 percent of smokers are prepared 
for action, 40 percent are only in the contemplation stage, and 40 percent in the precontemplation 
stage. From these percentages the authors concluded that action-oriented control programmes 
would only target a small percentage. The success of smoking cessation programmes is directly 
related to the stage the smokers were in at the beginning of the intervention, indicating that 
traditional ways of treating all individuals as if they were the same is not achieving maximum 
success (Prochaska et aI. , 1992). Thus they suggested that interventions should be stage-matched in 
order to achieve maximum success in tenns of people entering therapy, continuing therapy, 
progressing in therapy and progressing after therapy (DiClemente and Prochaska, 1998). 
There is a wide range of programmes available to help individuals overcome their addiction or to 
prevent addiction. Smoking prevention programmes that were intended to prevent adolescents from 
becoming cigarette smokers have tried several approaches: providing information on the 
consequences of smoking; concentrating on social influences and resistance skills to withstand the 
temptation of smoking; and general life/social skills (Durell and Buoski, 1984). The latter two 
approaches seemed to be more successful than the fIrst one (Silvestri and Flay, 1989). These skills 
are also incorporated into cessation programs. Programmes available for individuals being addicted 
to smoking range from no-cost self-change programs, to self-help books, media, community 
programs, self-help groups, advice from health professionals, and to expensive commercial or 
clinical programmes (Thompson, 1978; Brownell et aI. , 1986). Nicotine replacement treatment by 
patch is one form of commercial program, which was shown to be effective when comparing with a 
placebo group (Hays et al., 1999), while nicotine gum was only effective on short-term assessment 
(Haaga and Kirk, 1998). Such programmes are mostly combined with consultation and follow-up 
and relapse prevention consultations (Fiore et aI. , 1994; Hurt et aI. , 1994; Silagy et al., 1994). The 
effect of such combined therapies might be higher than either alone (Klesges et al. , 1996). 
Although the training for professional consultants is stressed, some studies did not fInd any 
22 
differences in the effect of self-trained versus workshop-trained professionals (Cameron et aI. , 
1999). For most individuals certain programmes work better than others and screening individuals 
prior to the commencement of a programme might help to match individuals for their optimum 
strategy and would allow professionals to focus on these individuals with the greatest chance of 
success (Brownell et aI. , 1986). 
In order to facilitate long-term maintenance of non-smoking three main methods were traditionally 
available: The first method was the provision of 'booster' sessions, which have been found to be 
ineffective (Lichtenstein, 1982). The second method was the addition of relaxation, contingency 
management and assertion training to the ordinary treatment methods. However, these additions 
cause complexity, and thus less compliance and less effectiveness (Marlatt and Gordon, 1985). 
Lifelong treatment was the third method, an approach taken by Alcoholics Anonymous or by 
Weight Watchers offering lifetime membership. An evaluation of these programmes is difficult 
because of their long-term duration and the problems associated with research on commercial 
groups (Brownell et aI. , 1986). 
Apart from education, anti-smoking policies also include regulation and taxation. It seems that 
taxation regulations are effective, while advertising bans are less effective. Several studies across 
different countries found a decrease in cigarette consumption after increasing taxes, whereby the 
larger the tax, the higher the reduction in tobacco consumption (Peterson et aI. , 1992; Godfrey and 
Maynard, 1995; Stephens et aI. , 1997; Jha and Chaloupka, 2000), however none of these studies 
investigated if the reduced consumption was sustained. A decrease in tobacco consumption can 
also be achieved with extensive bans on tobacco advertising. Partial advertising bans had little or 
no effect (Horgen and Brownell, 1998; Jha and Chaloupka, 2000), mainly because of other forms 
of advertisement. Advertising as posters or in magazines proved to be more cost-effective than for 
example television advertisement (Horgen and Brownell, 1998). One example of a removal of 
advertisement that had no effect on consumption, was when the tobacco industry 'voluntarily' 
removed all advertising from television in the United States. However, this happened after 
legislation was put into place, that required an equal amount of time being spent on anti-smoking 
advertisements. The tobacco industry feared the effect of these anti-smoking messages and 
removed all advertisement from television, but instead concentrated on other media (Horgen and 
Brownell, 1998). 
More recent methods of preventing relapse and cessation programmes include motivation 
enhancement, even involving monetary incentives. In different studies it was found that addiction 
to smoking or other drugs did not seem to be due to lack of suffering, knowledge, education and 
insight. Thus more focus is now put on motivation (Miller, 1998). Motivation is often brought in 
context with the analogy of a carrot and a stick, whereby many people misunderstand the stick as 
23 
some sort of punishment. Yet in the original image the carrot is tied to a stick and therefore 
dangling in front of the donkey, who cannot reach it, but in trying to get to it, the donkey will move 
forward and pull the cart. Thus the carrot and the stick is not a punishment, but an alternative to it 
(Miller, 1998). Motivation does not try to trick people into doing something they do not want to do, 
but it tries to bring people to understand that their addiction is counterproductive to main 
goals/objectives the individual has (Miller, 1998). A motivator for most behaviours is positive 
reinforcement, as people will do what causes a positive feeling (Miller, 1998). This positive 
reinforcement is most effective if it comes frequently and from individuals close to the addicted 
person (Miller, 1998). 
Another form of motivation is often the personalised realisation of pros and cons of smoking and 
its related diseases. It is often postulated that the general knowledge of these pros and cons is 
insufficient unless the smoker experiences them in hislher own situation. A study of post-coronary 
patients resulted in a high cessation rate (Burling et aI. , 1984; Ockene and Zapka, 1997). In contrast 
a study by Silagy et al. ( 1994), who performed a meta-analysis on nicotine replacement therapies, 
found only limited success. In their study the combined results of using nicotine gum in hospital­
based patients only showed a cessation rate of 15% versus 1 1  % in the no-treatment group, which 
the authors considered 'disappointing' ,  as the patients often had coexisting smoking-related 
diseases that might have been an added incentive to quit. 
Motivation not only includes the perception that quitting smoking is a betterlhealthier way, but also 
that the smoker sees him-Iherself capable of quitting smoking (Haaga and Kirk, 1998). It is 
concluded that research and investigation still have to continue in the field of motivation. 
Example from Agricultural Extension 
Behaviour change is needed when there is a gap between where an organisation/individual/ 
community is at present and where the individual/policy makers/market forces/quality assurance 
schemes and others want it to be. These forces encourage behaviour change by three means: 
economic means (penalties, subsidies), legislation (rules, regulations) and by participatory 
processes, such as extension, where people change because they perceive it to be for their own best 
(Greer and Greer, 1996). 
Definition of Extension 
Extension can be defined as a linear process transferring information and knowledge about new 
technology from scientists/industry/other organisations to educators and then to users (Morris et aI., 
24 
1 995). Technology transfer is only one element of extension. Extension is defined differently 
across countries and organisations, causing inherent conflicts (Roling, 1988). There are two main 
elements to extension: providing information so that the individual can clarify and achieve their 
own goals and empowering the poor and thereby achieve structural changes (Roling, 1 988, p.37). 
Often human resource development principles are used in both elements, teaching people how to 
learn, organise, manage, analyse their environment and others. Roling ( 1 988, p.39) notes that 
despite the varying definitions extension comprises following common features: 
it is an intervention 
it uses communication to induce change 
it is only effective through voluntary change 
it focuses on several different processes and outcomes 
it is arranged by an institution. 
Extension is expensive, a reason why it was usually carried out by institutions rather than 
individuals (Roling, 1 988, pA8). All over the world, Governments use extension as a policy 
instrument in order to achieve goals such as export goals, national food security, efficient use of 
national resources. To the Government these goals may be more important than the welfare of the 
individual farmer (Roling, 1988, p.38). 
In the second half of the nineteenth century an increase of agricultural innovations occurred and as 
a consequence the linear model of extension became popular. Scientists defined which aspects of 
farming should be investigated and passed the solutions to farmers through extension by specialist 
educators. This linear model was accepted very rapidly and universally, as the development of the 
'diffusion of innovation' concept showed. Rogers ( 1 995) defines this concept as 'the process by 
which an innovation is communicated through certain channels over time among the members of a 
social system' . A slight variation of this linear model is the multiplier diffusion model, where the 
information diffuses in a secondary diffusion process through the informal farmer to farmer 
network (Fairgray, 1 979). 
Pretty and Chambers ( 1 993; reviewed in Haug, 1999) categorised extension theory according to the 
approach taken and the influence of different disciplines into four stages: 
• Classical top-down, one-way transfer of technology (1 900- 1 975) 
Farmers were seen as recipients of technology. Pioneering work in crop and animal 
breeding. 
• Two-way transfer of technology ( 1 975-1985) 
25 
Farmers were seen as s ources of information and technology design. Pioneering work by 
economists and agronomists. 
• Ecological stage ( 1985-1995) 
Farmers were seen as causes and victims of environmentally unsustainable development. 
Main disciplinary influences from anthropology, agroecology and geography. 
• Institutional stage ( 1995-onwards) 
Farmers are seen as full collaborators in research and extension. Disciplinary influences 
from psychologists, sociologists, political scientists, training specialists and educators. 
Adoption process and efficiency of technology transfer 
In order to be successful a desired change or new innovation has to address farmers' needs or 
create awareness for these needs. Efficiency of technology transfer from scientists to farmers varies 
with the nature of the technology (Raling, 1 998). Easy to use, cost-effective methods will quickly 
be adopted by the majority of farmers often after only creating awareness. In contrast, other 
technologies need more intervention, such as face to face consultation, in order to get them 
adopted. The 'trickle down' or diffusion model assumes that once some individuals have adopted 
the new technology, its use will 'trickle down' or diffuse through the whole target population. The 
model therefore assumes that the better the communication in the fIrst place, the better and the 
more effective the diffusion process will be (Raling, 1 998, p.53). However, Rogers ( 1 995) adds 
that most innovations diffuse at a 'disappointingly slow rate ' .  Anderson ( 198 1 )  cautions on 
overemphasising the 'early adopter' group, as the network of these farmers often contain a high 
proportion of other 'early adopters' . Therefore any information entering this circle is likely to 
remain there, rather than diffuse down to the general farming community. 
In the 1 920' s  studies started looking at the adoption of innovations and a vast number of studies 
supported this 'trickle down' strategy (Weber, 1 974; Kung, 1 984; Ghosh, 1 993; Timmons Roberts, 
1 995 just to mention a few). Rogers ( 1 995) provided a summary of the literature on the technology 
transfer, in which he divides the adoption process into fIve distinct stages: knowledge, persuasion, 
decision, implementation, and confIrmation. The knowledge stage can occur even before a farmer 
feels the need to change. Once the need is there, the change i s  made and the outcome classifIed as 
favourable or unfavourable (persuasion stage). However, a favourable outcome does not 
necessarily mean adoption. The decision to adopt the change depends on its relative advantage, its 
compatibility with existing systems, its ease to use, its risk of implementing, and its observability. 
26 
These five characteristics were found to be the most important factors in explaining the rate of 
adoption of innovations (Rogers, 1 995). 
Rogers ( 1 995) stated that most innovations had an S-shaped rate of adoption and he divided the 
individuals into five groups: innovators, early adopters, early majority, late majority, and laggards. 
He also mentions that the most innovative members of a social system are often perceived as a 
'deviant from the social system' and the average members of the system assign them low 
credibility. Other members of the social system will act as opinion leader or 'knowledge 
influentials' , which is an informal leadership that does not depend on the individual' s  formal 
position or status in the social system, but which is earned through the person' s  technical 
competence, social accessibility, and conformity to the system (Anderson, 198 1 ;  Rogers, 1 995). It 
was also found that these opinion leaders and early adopters were usually more educated, had 
higher social status, got exposed to greater mass media communication, showed greater social 
participation and more cosmopoliteness, and were generally more progressive thinking than 
individuals adopting innovations later (Wassell and Esslemont, 1 992; Rogers, 1 995). Anderson 
( 1 9 8 1 )  observed that the early adopters were mainly belonging into the 'upper' and 'lower middle' 
rather than into the 'upper middle' economic rank. 
Innovations can be adopted or rejected by individuals independent of the decision of others 
(optional innovation-decision), but the social system as a whole can also decide to adopt an 
innovation, either by collective or by authority decision. Collective innovation-decisions are made 
by consensus among the members of the system. Once the decision is made, all units of the system 
have to comply with it. Authority innovation-decisions are made by relatively few individuals, who 
have the power, status or technical expertise. The individual members of the system are then 
required to implement the decision (Rogers, 1995). 
A major role of researches has been the identification of barriers to adoption. When Rogers and 
Shoemaker ( 1 97 1 ,  cited in Rogers, 1995) produced their first review of the diffusion literature they 
were emphasising motivation in the decision to adopt changes, whereas in the later versions they 
increasingly put more emphasis on the role of economic factors (Rogers, 1 995). Ander on ( 1 98 1 )  
found that there is a resistance to change, i n  that farmers have accumulated practical knowledge 
and by resisting change they did safeguard them against mistakes made by advisors over­
enthusiastic for change. Since then other constraints have been identified: if innovations are 
difficult to understand, farmers' motivation, their belief and opinion towards the innovation, their 
perception of the relevance of the innovation and farmers' attitudes towards risk and change 
(Guerin and Guerin, 1994). 
27 
Research into diffusion of innovations contributed importantly to the understanding of human 
behaviour change. However, there is also criticism about the method, such as ignoring or under­
emphasising rejection or discontinuation of innovations, which results in a lower understanding of 
innovation failures than innovation successes (Rogers, 1995). One has to acknowledge that the 
rejection or discontinuance might be rational and appropriate from the individual' s  point of view 
(Reid et aI., 1996). Late adopters and laggards are mostly individually blamed for not adopting the 
innovation, but it could be that the innovation was not as advantageous for them as it was for the 
early adopters. Additionally, farmers do not all have the same or even similar goals and needs, but 
extension programmes are often designed as if they had. Therefore Hannibal and Sriskandarajah 
( 1 996) proposed that extension programmes should be designed at the individual farm level, 
allowing the farmer to use those elements and techniques that suit his situation and mix it with 
existing practices. Nearly forty years ago McMeekan ( 1963) advocated that extension had to focus 
more on the needs of the industry in order to stay as useful for the industry as it was in the past. 
However, until recently, scientists and some groups in agriculture decided the research priorities, 
whose results were handed to farmers (Chambers and Jiggins, 1 987), rather than priorities being 
determined by the individual or the industry itself. 
Another criticism is that diffusion researchers for a long time did not pay sufficient attention to the 
consequences of innovations, especially how the socio-economic benefits of innovations are 
distributed within the social system (Rogers, 1995). Much analysis of the linear technology transfer 
model was done in developing countries and fed back to the developed countries (Raling, 1 988). 
These analyses showed that they mostly benefited the most productive and richest farmers, thereby 
increasing the socio-economic gap between the rich and the poor (Hightower, 1 973; Rogers, 1 995). 
Recent techniques in agricultural extension 
'Farmer first' model 
Out of the increasing criticism of the linear diffusion model, new models of technology transfer 
were developed, where farmers were regarded as co-researchers in a two-way exchange of 
knowledge between farmers and researchers (Morris et aI. , 1995). Often it was the case that 
farmers rejected innovations for practical reasons rather than conservatism. Farmer had detailed 
local knowledge of their environment for their farming system and should be seen as knowledge 
producers not only knowledge receivers (Kloppenburg, 1 99 1 ). Some more recent extension 
systems try to incorporate local resources and match local needs ( 'Farmer first' models) (Chambers 
and Jiggins, 1 987; Chambers et aI. , 1989). These programmes formally investigate farmers'  
circumstances, their goals, and constraints to change. Then they try to involve farmers in the 
28 
research process when designing and evaluating strategies intended to improve farmer well-being 
(McRae et a/., 1993). Instead of starting at the scientists' end, these programmes start at the 
farmers' end. A similar approach is used in the human health field, for achieving changes in 
nutrition behaviour, when using social marketing to find out the consumer' s perspective (Griffiths, 
1994). 
'Farmer first' programmes are also called 'participatory' or 'bottom-up' approaches, where farmers 
participate in research and extension processes. Thus they are also described as 'Agricultural 
Knowledge Information Systems' (AKIS) (Roling and Engel, 199 1 ). They were originally intended 
to be complementary to the conventional transfer of technology (Chambers et aI. , 1 989), while 
others see them as self-sufficient alternatives. Yet, these programmes also have been criticised for 
various reasons, such as new problems for farmers might not be recognised within the local 
knowledge of farmers, little dissemination beyond the group itself and others, little self­
development of the individual group members and others (for details see Black, 2000)). 
Another more recent development is using marketing approaches, whereby new knowledge is seen 
as a product, which is developed and then actively promoted. One such programme was the ' Prime 
Pasture program' in Australia, which aimed to increase the success of pasture establishment (Keys 
and Orchard, 2000). The programme combined different companies (seed, fertiliser, herbicide, 
machinery) to promote an integrated message, using financial incentives, on farm demonstrations, 
high quality brochures and other mass media. These different marketing strategies provided 
continuous reinforcement of the project' s message (Keys and Orchard, 2000). 
Communication 
Extension becomes increasingly multidisciplinary, with more emphasis not only on farmers' 
knowledge, but also on their personal goals, circumstances, and their individual learning styles 
(Paine, 1 993; McRae et aI., 1 993; Fairweather and Keating, 1994). Paine ( 1 993) suggested that 
extension agents can achieve better adoption of innovations by their clients by using 
communication methods that are compatible with their clients' preferred learning strategy, or by 
establishing relationships between farmers with opposite learning styles. 
There are three main types of extension techniques available: mass media channels (radio, TV, 
articles in newspaper, journals, booklets, newsletters), personal contact on a one-to-many basis, 
such as in discussion groups, field days, seminars, workshops, conferences; and there is personal 
contact on a one-to-one basis, such as farm visits. Anderson ( 1 982) found that these farm visits 
mostly involved the 'progressive' farmers, which had better education and larger and wealthier 
farms. He also found that reading as an information source became less important with the age of 
29 
the farmer (Anderson, 1 98 1 ). The one-to-one relationships can also be used to potentially increase 
the adoption of innovations by using such ' focus' farmers as an integral part of the delivery process 
to help 'sell' innovations (Pearson et al. , 2000). 
A potential fourth type of communication could be e-mail and the World Wide Web, which might 
have positive effects, in that the language chosen is more informal and personal, closer to one-on­
one contact than the traditional written information (Fell, 2000). 
Farmers' and consumers' goals 
Farmers' goals are main determinants of their motivation for adoption or rej ection of innovations, 
thus making them an important factor in extension. Without the knowledge of the goals and 
circumstances of farmers it is impossible for extension agencies to define what problems have to be 
solved (McRae, 1 993). The goals themselves depend amongst many other factors on the country 
and other s ocial factors. Gasson (1973) found that farmers in B ritain often mentioned 
'independence' and ' way of life' as personal goals and 'producing good livestock/crop' as business 
goals. For small farmers in New Mexico the 'quality of life' was more important than 'income' 
(Harper and Eastman, 1980), and farmers in Australia (Queensland) found 'safeguarding income 
for the future' more important than 'maximising income' (Cary and Holmes, 1982). However, 
these studies were done in 1973, 1980 and 1 982 respectively. As farmers' goals are dynamic 
(McRae, 1 993) it is questionable if these goals are still the same 20 to 30 years later. 
A recent example of involving farmers in the complete process of research and technology transfer, 
is the work of the Foundation for Arable Research (FAR). This foundation was set up by arable 
farmers to manage farmers' investment in research and information transfer. The involvement of 
farmers was described by Pyke and Johnston (2000) at a recent conference on achieving change. 
Another more recent development is the involvement of consumers' values and goals into the 
agricultural process. Most industrialised economies have an oversupply of agriCUltural products and 
therefore market growth is determined by quality not quantity (Grunert et al. , 1997). Thus 
agriculture has moved from a producer-dominated market to a consumer-dominated market. This 
also means that agriculture has to move away from being product-oriented to being consumer­
oriented, by meeting the needs of consumers. Therefore not only the goals of farmers are important 
but also the goals and values of customers (Frazer, 2000). These values and their application in the 
beef-industry were recently investigated in marketing research using the method ' means-end chain 
theory' (Audenaert and Steenkamp, 1997). A means-end chain intends to explain how a product 
allows the achievement of a desired end state, such as consumers buying products with certain 
attributes (e.g. beef without hormones), that can achieve a desired consequence (e.g. feeling 
30 
healthy) which is of value to the consumer (Gutrnan, 1982). In consumer behaviour three types of 
factors are generally distinguished - properties of the food, person-related factors and 
environmental factors (Steenkamp, 1997). The aim of marketing research is to identify growing 
production systems and thus indicating research areas for improved technologies (Janssen and 
Tilburg, 1997). 
Motivation 
More recently the term technology transfer has been redefined as technological learning and 
knowledge application (Foundation for Research Science and Technolgy, 1998). Within the theory 
of learning, five factors are important: motivation, relevance, interest, environment, and satisfaction 
(Oryden and Vos, 1 993). One of these factors on which much emphasis is put in recent years, is 
motivation (Alderman, 1 990). 
Motivation of the farming community is one of the main roles in agricultural extension (Allison, 
198 1).  However, it is very difficult to investigate motivations for adoption of innovations, as some 
people might not be able to clearly state these, others might not be willing to do so. Economic 
motivation is often important for certain innovations, especially if these involve high expenses. 
Another motivation is the prestige, which people might gain in adopting new technologies before 
their peers (Becker, 1 997). However, for private organisations, such as farming, economics might 
be more important than for public organisations, where prestige reasons might prevail (Rogers, 
1995). Farm profitability and production are often found to be main motivators for practical 
decisions (Carr and Tait, 1991) .  
In some cases, such as conservation issues, legislation and regulation may be the only effective 
way to ensure long-term change, if the dominant beliefs and values of the majority of farmers 
disagree with the change required (Carr and Tait, 1 99 1 ) .  
Computer aided programs 
The more complex farm problems get, the more complex the methods for analysing and solving 
these problems, especially when it comes to such long term problems, for example like land 
degradation problems. In these situations the use of microcomputer-based decision-support systems 
could be efficient and effective in analysing the complex situations, especially when the solutions 
become economically prohibitive (Ludwig and Marsden, 1 993). 
In Australia it was claimed that so far there has been little adoption of intelligent support systems, 
which they attributed in part to the incompatibility with current farming practices, or with attitudes 
3 1  
of farmers towards computers (Lynch et al., 2000). Yet, in New Zealand a relatively high 
proportion of dairy and pig farmers in comparison to such farmers in other countries use computer 
recording and evaluation programs ( 17% of dairy farmers use DairyWin and they manage 25-30% 
of cows, M.Stevenson, pers. comm. 2000). 
Another form of using computers is  by providing internet based two-way communication between 
extension workers and farmers as done in a study group of sheep farmers in the South Island of 
New Zealand (Mulcock et al. , 2000).  
Agricultural extension in New Zealand 
History of extension 
In New Zealand, extension prior to 1985, has been carried out by national non-commercial 
Ministry of Agriculture and Fisheries (MAP), and co-operatives of farmers but rarely by voluntary 
agencies (Morris et aI. , 1995). Most literature on extension is concerned with the adoption of 
individual technologies, the speed of diffusion and the characteristics of adopters and non-adopters. 
New Zealand implemented the diffusion of innovation model with some adaptations for local 
conditions and experienced the same weaknesses as other countries, such as expectations that 
technology provides solutions for all problems, and assumptions that increasing levels of 
technology is beneficial (Morris et aI. , 1995). The linear model of extension focused on a small 
number of farmers and leaves it up to communication amongst farmers to diffuse new innovations. 
New technologies that required minimal extension activities were aerial topdressing, farm bikes 
and some animal health remedies; whereas rotational grazing for example required much more 
intervention. 
From World War IT to the mid 1980s production was the main focus of farm management and 
extension, and the key criterion for judging success (Fairweather and Keating, 1994). During this 
phase, the industry concentrated on increasing the volume of production as a result of Government 
incentives, increased processor capacity, agricultural research and extension (Parminter et al. , 
1993). 
Attitudes towards extension in New Zealand were strongly influenced by the nature of its 
agriculture, with major exports of agricultural products, such as dairy products, wool and sheep 
meat. Government was involved in agriculture since the time of early settlement, first by 
controlling the sale of land, later by controlling pests and some diseases, by importing animals, 
providing credits for farmers, instituting research and providing agricultural regulations. In the 
32 
twentieth century Government established producer boards, quality controls, export regulations and 
sometimes price support (Morris et al. , 1995). The Department of Agriculture, established in 1 892, 
employed instructors in farming techniques (Nightingale, 1 992). Their number increased very 
rapidly, especially after the introduction of mass production of dairy products and refrigeration. By 
1920 face-to-face consultations with farmers, pamphlets and reports were available (Nightingale, 
1992). The department used the diffusion model. It decided what changes should be promoted, it 
trained extension officers in technical and fmancial management, but also in adult education skills, 
and encouraged farmers to make their own decisions and monitor their own progress. Especially  
during the time when agricultural production was rapidly increasing these programmes seemed to 
be highly successful (Parrninter et al. , 1 993; Morris et al. , 1 995) .  
With Great Britain joining the EEC, the guaranteed access to B ritish markets was lost and farmers 
were subsidised for finance, fertiliser and transport until 35% of farmers' income (40% of sheep 
and beef farmers' income) was paid by Government in 1 983 (OECD, 1988; Tyler and Lattimore, 
1990; Walker and Bell, 1 994) By the early 1980s it was apparent that these fmancial 
compensations to farmers did not achieve the intended social or economic results, but only 
contributed to unsustainable national debt (Hawke, 1 987; Rayner, 1990). Additionally New 
Zealand's farming community had decreased to only about 2% of the population at that time (New 
Zealand Department of Statistics, 1940- 1 995) causing decreased justification for large 
Governmental expenditure on agriculture. In 1 984 a new elected Labour Government rapidly 
removed subsidisation, progressively introduced cost-recovery for consultancy, and restructured 
Government's research and development agencies, in order to make them more efficient (Rayner, 
1990; Walker and Bell, 1 994). 
In the 1990s profitability became more important than production, with economic efficiency being 
the key criterion of success (Fairweather and Keating, 1 994). In addition the agricultural extension 
service, previously owned by Government, was commercialised (transition to user-paid 
consultancy) and privatised (transition to privately owned services). Many articles and books 
address the changes to the extension service during the privatisation process (Rayner, 1 990; 
Fairweather, 1992; Ritchie, 1 995). Shortly after the change of Government MAP began to charge 
for selected extension services, after legislation for this was put in place (Tyler and Lattimore, 
1 990). In 1 987 MAP was restructured and in 199 1  the research group within MAP Technology unit 
was transferred to the Department of Scientific and Industrial Research (DSIR), from which time 
full cost recovery was employed for all extension services. In 1 992 the DSIR was commercialised 
as well and restructured into Crown Research Institutes (CRI), which are state owned but operate as 
for-profit business that pay dividends to their shareholders and taxes to Government (Coddington, 
1 993). MAP Technology itself was renamed to Agriculture New Zealand to indicate its commercial 
33 
nature, and was fully commercialised in 1994 as a stand-alone agency within MAF and one year 
later privatised to Wrightson Pty Ltd, a national agriculture supply/service company (Hall and 
Kuiper, 1998). 
Many farmers had to come to terms with having to pay for services they got free previously but in 
the process looked to fmd the best value advisory services. Implementation of innovations was 
found to be greater if farmers had to pay for them, compared with when they were 'free' (Hall and 
Kuiper, 1998). After the commercialisation of the public funded extension, farmer co-operatives 
and commodity boards started to levy their members and in turn provide services to them at low or 
no cost. 
Differences existed and still exist between dairy and sheep and beef farmers in the amount of 
extension available to them, with more extension available to and used by dairy farmers than 
sheeplbeef farmers. Furthermore the information available to dairy farmers is often less conflicting 
than that available to sheeplbeef farmers, as the latter have a more complex farming system. 
Extension to dairy farmers is provided by the New Zealand Dairy Board through a levy on farmers, 
through Agriculture NZ consultants and through The Dairy Research Institute. That institute has 
three times the number of scientists than the meat industry (Stichbury, 1994). There is no such wide 
extension range available to sheep farmers (Morris et aI. , 1995). Rhodes and Aspin (1993) 
described a similar system whereby the Meat Research and Development Council used the levies 
by the Meat Board for project support on 24 Focus Farms. These farms are used for education in 
discussion groups and field days. The difference in the available extension services between the 
farm types is also related to the way products were marketed. In the dairy industry producer owned 
co-operatives were acting since the early 20th century, whereas in the meat industry this is more 
recent (1970s and 1980s) (Morris et al. , 1995). 
It is estimated that during the period when extension services were free of charge to farmers, 
approximately 80% of all New Zealand farmers were being served with some form of advisory 
service. Through the introduction of cost-recovery this number declined to about 10% for beef and 
sheep farmers until 1990/91. Since then the number has increased again, but it is considered well 
below the number prior to commercialisation (Hall and Kuiper, 1998). 
With commercialisation of the advisory services the focus of the service providers also changed. 
Whereas the mission and goal statement of the MAP Advisory Services in 1983 was: 'to help 
farming and horticultural industries to identify and realise potential', it was in 1995-98 'to assist 
strategic growth in agriculture and horticulture by being New Zealand' s foremost provider of rural 
consultancy services' (Ritchie, 1995). This change in the mission statements reflects the change 
34 
from leading farmers to independent decision-making to complementing farmers' strengths by 
advising them in their weak management areas (Hall and Kuiper, 1998). 
Recent extension work has not only focused on the 'traditional' crop and livestock farming 
enterprises, but also on horticulture (Banks and Harley, 2000; Praat et al. , 2000), organic 
production (Kelly et aI. , 2000), sustainable agriculture (Rush, 2(00) and forestry (Bathgate, 20(0). 
Communication channels 
Extension has to be reviewed regularly in order to ensure that they are meeting the changing needs 
and goals of farmers (Mac Clean et aI. , 1997). Only when these needs and goals are understood the 
best methods of communication can be established. Prior to commercialisation MAF Advisory 
Services used mainly discussion groups, field days, meetings and conferences to provide farmers 
with information. Through commercialisation this was shifted towards one-to-one consultations 
(Hall and Kuiper, 1998). Currently extension to farmers is provided by education, individual 
consultations, discussion groups, field days, workshops, conferences, and media (Walker and Bell, 
1994; Morris et aI. , 1995; MacClean et al. , 1997). 
In educational organisations usually up to date information is used to train students, who then can 
take out their knowledge into the wider farm population. However, student numbers for agricultural 
courses have decreased since 1984/85 and less students have contact with farming prior to their 
studies (Wyllie, 1989). These two facts could have contributed to a lesser introduction of new ideas 
(Morris et aI. , 1995). Consultations with farm advisors can help farmers plan for specific goals and 
provides feedback on new innovations in practice (Garland, 1993; Walker and Bell, 1994). Morris 
et al. (1994) found that farmers using consultants were more production oriented. However, due to 
the one-on-one contact it is not possible to reach all farmers (Walker and Bell, 1994). 
Discussion groups amongst farmers usually involve farmers with similar interests and provide 
knowledge about new ideas and their implementation to progressive farmers (Rwenyagira, 1985; 
Walker and Bell, 1994; Greer and Greer, 1996). Their establishment can be very demanding, but 
they allow the joining of farmers' needs with science information (Tarbotton et aI. , 1997). In a 
study of farmers in the North Island of New Zealand they were the most commonly attended events 
(Gavigan and Parker, 1996). Discussion groups also play a role in providing a social network 
(Wegener et aI. , 2000). They also have an indirect influence on non-participating farmers, who 
observe and copy the improved techniques (Morris et aI. , 1995). However, often there is help 
available in setting up groups, but little training or skill development in managing group dynamics 
and providing on-going support (Oliver et aI. , 2000). In an Australian study of discussion groups it 
. was found that they are often without specific direction and the extension workers were unsure 
35 
about the practical implications of the topics discussed (Wegener et aI. , 2000) .  Another factor is, 
that discussion groups may not appeal to all farmers (Morris et aI. , 1 995). The same applies to 
conferences. However, conferences have the potential of incorporating interactive workshop and 
seminar sessions that allow extension agents and agricultural researchers to receive feedback from 
those that will apply the knowledge (Mac Clean et al. , 1 997). 
Workshops, using small groups, intensive face-to-face contacts, are increasingly used to provide 
learning opportunities and can achieve that most participants gain the targeted knowledge and skills 
(Stantiall, 1 999; Stantiall, 2000; Frazer, 2000).  
Mass media is least linked to changing practices. They principally provide awareness on new 
methods and innovations and also reasons for change (Walker and Bell, 1 994). Nevertheless, 
publications of this sort are read regularly by farmers (Gavigan and Parker, 1996). As most farmers 
are not enthusiastic readers, articles written in newsletters and agricultural j ournals have to be 
written in a concise readable form, and in a style that is familiar to farmers (Garland, 1 993). 
One-on-one consultations provide an opportunity for whole-farm analyses, looking at the specific 
situation of the farm, and establishing professional and personal goals and plans to achieve these. 
However, they provide fewer opportunities for farmers not using advisory services (Hall and 
Kuiper, 1998). 
Advisors found that extension programmes were most successful if they involved the industry in 
the planning process, if they gave farmers simple measuring techniques to allow them to measure 
their own progress on their own farm., if they used the full range of extension techniques, if 
advisors were trained in adult education and if they assisted farmers in the learning process, rather 
than telling them what to do (Walker and Bell, 1994). Walker and Bell ( 1 994) quoted the 
philosophy "tell me what to do and I will remember about 10%, show me what to do and I will 
remember about 40%, involve me and I will remember most of it". These authors also found that 
the hardest part of extension programmes was the creation of motivation for change (Walker and 
Bell, 1 994). 
Factors influencing change in farm practices 
The effectiveness of extension programmes has been investigated since the 1 960s (Allison, 198 1 ;  
Greer, 1982). Only few farmers resist change, but it is necessary to know the costs and benefits of 
new innovations on each farmer' s own farm in order to understand why adoption or rejection 
follows (Morris et aI. , 1 995). Factors that are believed to influence adoption of new technologies 
are age, education, management skills and communication between farmers (Morris et aI. , 1995) .  
36 
Higher income and farm size was also positively associated with the adoption of new technologies 
(Stewart, 1979). The goal of increased farm income, less time having to be spent on farm work, 
more available time for leisure are some most common motivations for changes (Greer, 1982). 
Verkerk et al. (2000) reduced the number of factors that influenced the acceptance of new 
technologies to three key performance factors: the success achieved, the fmancial return and the fit 
to values and beliefs. 
Farmers' goals and underlying valueslbeliefs are crucial for any change. Although many farmers 
rank their farm p oduction and profitability as their highest goal, they also have a whole range of 
other goals (Gavigan and Parker, 1996; Parminter and Perkins, 1997). If this range of goals is 
accounted for in extension services there is a higher likelihood for voluntary adoption of 
innovations (Parminter and Perkins, 1997). Not only business goals are important, but also personal 
goals (McRae, 1993; Fairweather and Keating, 1994), core enterprise beliefs (Parminter et al. , 
1993) and long-term goals of the farming business, such as expansion and less personal 
involvement (Valentine et aI. , 1993). Fairweather and Keating ( 1994) categorise farmers according 
to their business and personal goals into three management style groups: dedicated producers, 
flexible strategists, and environmentalists. These authors suggest that distinctive models could be 
developed for each of the management styles, taking into account the proportion of each of the 
types. 
Greer ( 1982) noted that many farmers that had rejected new innovations actually had accepted 
inaccurate information, on which they based their decision, indicating a flaw in the linear extension 
model. 
Greer ( 1982) also concluded from reviewing New Zealand literature that because of 
methodological problems little of the research into motivation of farmers could be linked directly to 
their adoption behaviour, but that communication between farmers was crucial. This conclusion 
was also made by several other researchers, such as Fairgray ( 1979) in his study of farmers 
adopting rotational grazing, or Panninter et al. ( 1993) in investigating the adoption of new beef 
breeding cow technologies. Morris et al. (1994) also emphasised the farmer network in the spread 
of new technology. Once farmers considered new technologies they sought information on their 
performance on farms in the local area and the most convincing proof of an innovation is seeing 
another farmer successfully using it (Garland, 1993; Morris et aI.,  1995). This reference to the local 
area was of major importance (Morris et al. , 1995). 
'Farmer First' research programmes started in the early nineties that showed that farmers were 
aware of new technologies, but either considered them unsuitable for their particular situation or 
were financially not able to adopt them (Reid et al. , 1993; Brazendale et al. , 1994). 
37 
In a study interviewing farmers on their reasons for changing their farm practices several main 
factors influencing the decision to change farm practices were identified (Morris et aI. , 1995). The 
goal for most changes was to increase profitability, either by reducing labour or increasing 
efficiency. Another factor influencing the adoption of new technology was their trialability. 
Farmers preferred to try an innovation on part of their farm and then decide to adopt or reject if for 
the whole farm. Past experiences were especially important for sheeplbeef farmers in their decision 
making. Farmers wanted to farm safely with less associated risks, which also meant that farmers 
were more likely to adopt changes if they assisted in attaining control over factors which might 
impact upon production (e.g. soil testing, herbage testing). Sheeplbeef farmers were more likely to 
prepare personally for industry crisis than dairy farmers. Younger farmers were more willing to 
change and take risks than farmers who farmed at the time of the downturn ( 1980s) (Morris et al. , 
1995). 
Most farmers were aware of new technology (with more or less detailed knowledge), and non­
adoption of new innovations was after a deliberate decision was made not to change (Morris et aI. , 
1995). Some farmers had fewer options to adopt new technologies, due to financial (high debts, 
small farms) or topographical constraints (Morris et al. , 1995). 
Morris et al. ( 1995) reported that dairy farmers were more likely to adopt new innovations than 
sheeplbeef farmers, while Journeaux ( 1990) five years earlier found that sheep and beef farmers, 
given enough motivation, often relating to fmancial incentives,  were as quick to adopt new 
technologies as dairy farmers. However both authors state that it was easier for dairy farmers to see 
the impact of new technologies, as their farming system is simpler and only one outcome (milk) is 
produced. Furthermore, more extension activities were conducted on dairy farms than on beef 
farms and farm income for dairy farms has increased even during the time of the rural downturn of 
the 1980s. Adoption of new technologies in sheeplbeef farming systems usually took longer than in 
dairy farming. Most sheeplbeef farmers know of farmers that have lost their farm in the downturn, 
despite acting on the advice they had been given. The confidence and trust in the industry and in 
the advice had been undermined for these occurrences (Morris et aI. , 1995). 
Commercialisation of advisory services 
With commercialisation of advisory services an improved adoption rate for innovations was 
achieved, but on a smaller number of farms, indicating that the farmers who use the service are 
more likely to adopt it (Hall and Kuiper, 1998). The reasons for this were two-fold: Firstly farmers 
paid for services and therefore demanded very specific advice, which was immediately applicable 
to their situation. Secondly, farmers valued service they had to pay for more than the free service 
(Hall and Kuiper, 1998) It is a characteristic of human nature to perceive something that is received 
38 
free, as  having little or no value (Kerr, 2000). It might be that the fanners willing to pay for advice 
are also the ones more willing to change and to adopt innovations, therefore the number of fanners 
adopting innovations might be still the same as before, but the total number of fanners using the 
service is lower, thus increasing the percentage of fanners receiving and adopting advice (Hall and 
Kuiper, 1998). Other advantages of commercialisation! privatisation of extension include a greater 
responsibility to clients, greater emphasis on benefits and results, rather than purely service 
activities and the shift from a 'technology push' to a 'demand pull' orientation (Rivera, 2000). 
Yet, other authors stress the fact, that there is a need for co-operation between advisors and farmers 
to develop new ideas and design new systems. This information flow will be cut if all advisory 
services have to be paid for by farmers (McArthur, 1 987; Butcher, 1 987). The concept of user-pays 
depends on the benefits that incur to the farmer and/or to the public (McArthur, 1987; Butcher, 
1987). However, in farming it is sometimes difficult to decide who is the main beneficiary. 
Although additional knowledge is first beneficial to the farmer acquiring it, yet, over time, 
knowledge becomes a public good, as one person using it does not exclude others using it. On the 
other hand, the distribution of this knowledge benefits mainly the farmers getting the advice 
directly, so therefore farmers should pay for this. However, these authors ask the question who is 
paying whom if the advisor comes on to the farm and a two-way information flow exists between 
farmers and advisors. Often 'new' technologies are investigated where farmers actually played a 
creative role in the generation of these new technologies (Chambers et al. , 1989). Another 
disadvantage of commercialisation of extension is that it becomes a service only for producers that 
can afford to pay for it (Rivera, 2000). 
The privatisation of extension services meant also that environmental and sustainable-agriculture 
education programmes ceased. This necessitated more regulations, fmes and legislation to 
'encourage' farmers to adopt such practices (Hall et al. , 1999). 
Relevance to the hypothesis researched in  this thesis 
Bovine tuberculosis (TB) is of major importance to New Zealand' s agricultural economy with 
potential trade restrictions, if the level of infected herds cannot be reduced from currently 2% to 
0.2% (Oliver et al. , 2000). Three main control methods are employed in the control of TB: stock 
testing, stock movement restrictions, and control of infected wildlife. One of the three key 
objectives of the 1996-2001 National Pest Management Strategy is to 'encourage individuals to 
take action against TB on their properties and in their herds' (Animal Health Board, 1995). It is 
recognised that there is a need to complement official control efforts with farmer effort. 
39 
If farmers in New Zealand are expected to assist in the control of tuberculosis more than they have 
done so far (Animal Health Board, 2000), it is necessary to provide them with new ideas/methods 
of controlling tuberculosis on the farm level. The insights gained from this literature review were 
combined with epidemiological knowledge of tuberculosis in livestock and wildlife. Knowledge of 
habitat factors, TB hot-spots (McKenzie and Morris, 1995) and behavioural studies of interactions 
between livestock and wildlife (Paterson and Morris ,  1 995; Sauter and Morris, 1 995) led to the 
development of on-farm control measures (see Chapter 3). These measures had to be implemented 
and evaluated on study farms. From the literature review it was known that the one-on-one contact 
provided the best way of transferring knowledge between the farmers and the scientists in a two­
way manner. As participation of farmers in the development and implementation of innovations is 
stressed by several authors as a mean to achieve greater adoption (Chambers et aI. , 1 989; Pyke and 
Johnstone, 2000; Verkerk et aI. , 2000), farmers were involved throughout the project and 
responsibility for vector control handed over to farmers in the third year of the project. The 
literature also provided evidence that market forces or other external imperatives, such as for 
example Animal Health Board policies, can have a strong influence on the adoption of new 
ideas/innovations. 
Conclusion 
In this chapter I have briefly described the field of changing human behaviour, the processes 
involved and some mechanisms that try to achieve this change. Two examples were described in 
more detail, the field of smoking cessation and the field of agricultural extension with the focus on 
the New Zealand situation. In both examples the belief is that the change in behaviour will lead to a 
specific goal, better health in the smoking example and more productivity or achievement of 
personal goals in the field of agricultural extension. 
• The key process for changing human behaviour is: 
need recognition '" search for information '" evaluation '" choice (Rogers, 1 995) 
therefore the need to change has to be instigated/accepted, before any of the other steps are 
following. 
40 
• Key factors influencing the adoption of innovations are: 
Farmers' perception of the relevance, compatibility, complexity, trialability and 
observability of innovations, financial costs, farmers' beliefs and opinions towards the 
innovation, farmers' attitudes towards risk and change and farmers' motivation (Guerin 
and Guerin, 1994; Rogers, 1995). 
In agricultural extension farmers are not just a passive part in the technology flow, but extension 
personnel' s  main purpose is to help farmers help themselves. Therefore, the well-known Chinese 
proverb can also be applied to agricultural extension: 
"If you give a hungry man a fish, he will be fed for one day. 
But if you teach him how to fish, he will be fed for life." 
The main emphasis presently is on empowering farmers to take responsibilities themselves. This 
applies to farm productivity, but also disease control. 
Bibl iography 
Alderman, M.K. ( 1 990). Motivation for at risk students. Educational Leadership, 48 ( 1 ) :  27-30. 
Allison, AJ. ( 1 981). The effectiveness of farm advisory services in New Zealand: The scientist's 
view. New Zealand Agricultural Science, 15 ( 1) :  15-20. 
Anderson, A. M. ( 1 98 1 ). Farmers' Expectations and Use of Agricultural Extension Services. 
Hawkesbury Agricultural College, Richmond, NSW, Australia. 
Anderson, A. M. ( 1 982). Processes and implications of knowledge transmission in Australian 
Agricultural Extension. Hawkesbury Agricultural College, Richmond, NSW, Australia. 
Animal Health Board ( 1 995). National TB Strategy: proposed pest management strategy for bovine 
tuberculosis. Animal Health Board, Wellington, New Zealand. 
Animal Health Board (2000). Bovine Tuberculosis National Pest Management Strategy 200 1-20 1 1 ;  
Towards a TB free NZ; A discussion paper on the future options. Animal Health Board, 
Wellington, New Zealand. 
41 
Ashby, A.W. ( 1926). Human motives in Farming. The Welsh Journal of Agriculture, 5-12.  
Audenaert A., Steenkamp J.E.M. ( 1997). Means-End Chain Theory and Laddering in Agricultural 
Marketing Research. In: Wierenga B. ,  van Tilburg A., Gruntert K., Steenkamp J.E.M.,  Wedel 
M. (eds). Agricultural Marketing and Consumer Behavior in a Changing World. Kluwer 
Academic Publishers, Boston, USA, pp. :2 17-230. 
Ballantyne, R.R. and Packer, J.M. ( 1996). Teaching and Learning in Environmental Education: 
Developing Environmental Conceptions. Journal of Environmental Education, 27 (2): 25-32. 
Banks, N. and Harley, L.  (2000). Technological innovation in the Kiwifruit industry. In: Stantiall 
ID, McDiarmid CM (eds.), Achieving Change Through Improved Knowledge Systems. 
Centre for Agricultural and Veterinary Continuing Education, Massey University, Palmerston 
North, New Zealand, pp. 27-33. 
Barr, J. (2000). Opportunities for Growth: Developing Skills in the Facilitation of Learning. In: 
Stantiall ID, McDiarmid CM (eds.), Achieving Change Through Improved Knowledge 
Systems. Centre for Agricultural and Veterinary Continuing Education, Massey University, 
Palmerston North, New Zealand, pp. 133-137. 
Bathgate, J. (2000). Innovation for competitive advantage - a forestry case study. In: Stantiall ID, 
McDiarmid CM (eds.),  Achieving Change Through Improved Knowledge Systems. Centre for 
Agricultural and Veterinary Continuing Education, Massey University, Palmerston North, 
New Zealand, pp. 243-247. 
Becker, M.H. (1997). Factors Affecting Diffusion of Innovations among Health Professional. 
American Journal of Public Health, 60: 294-305. 
Black, A.W. (2000). Extension theory and practice: a review. Australian Journal of Experimental 
Agriculture, 40 (4) : 493-502. 
Botha, N. (2000). Group Functioning: A Theoretical Overview. In: Stantiall ID, McDiarmid CM 
(eds.),  Achieving Change Through Improved Knowledge Systems. Centre for Agricultural and 
Veterinary Continuing Education, Massey University, Palmerston North, New Zealand, pp. 
67-75. 
42 
Brazendale, R., Reid, J.I. and McRae, A (1994). The on-farm Impact of Beef Production 
Technologies. Proceedings of the New Zealand Society of Animal Production, 54: 413-416. 
Brownell, K.D., Marlatt, G.A, Lichtenstein, E. and Wilson, G.T. ( 1986). Understanding and 
preventing relapse. American Psychologist, 4 1 :  765-782. 
Burling, T.A, Singleton, G.E., Bigelow, G.E., Gottlieb, W.F. and Gottlieb, S .H. ( 1 984). Smoking 
following myocardial infarction: A critical review of the literature. Health Psychology, 3 :  83-
96. 
Butcher G. ( 1987). User Pays Revisited. In: Wallace L.T., Lattimore R. (eds). Rural New Zealand 
what next? Agribusiness and Economics Research Unit, Lincoln College, Canterbury, New 
Zealand, pp. :29. 1 -29.9. 
Cameron, R., Brown, K.S., Best, J.A. et al. ( 1999). Effectiveness of a Social Influences Smoking 
Prevention Program as a Function of Provider Type, Training Methods, and School Risk. 
American Journal of Public Health, 89 ( 12): 1 827- 183 1 .  
Carr, S .  and Tait, J .  (1991) .  Differences in the Attitudes of Farmers and Conservationists and their 
Implications. Journal of Environmental Management, 32 (3): 28 1 -294. 
Cary, lW. and Holmes, W.E. ( 1982). Relationships among farmers' goals and farm adjustment 
strategies: some empirics of a multidimensional approach. The Australian Journal of 
Agricultural Economics, 26: 1 14-1 30. 
Chambers, R. and Jiggins, J. (1987). Agricultural research for resource-poor farmers. Part 1 :  
Transfer of technology and farming systems research. Agricultural Administration and 
Extension, 27: 35-52. 
Chambers, R., Pacey, A. and Thrupp, L.AE. (1989). Farmer First: Farmer Innovation and 
Agricultural Research. The Bootstrap Press, New York, NY. 
Coddington, D. ( 1993). Turning pain into gain - the plain person's guide to the transformation of 
New Zealand 1984 - 1993. Alister Taylor Publications, Auckland, New Zealand. 
43 
Cohen, S., Lichtenstein, E., Prochaska, I.O. et al. ( 1989). Debunking Myths About Self-Quitting, 
Evidence from 10 prospective studies of persons who attempt to quit smoking by themselves. 
American Psychologist, 44 ( 1 1) :  1355-1 365. 
Davidson R. ( 1998). The Transtheoretical Model, A Critical Overview. In: Miller W.R., Heather N. 
(eds). Treating Addictive Behaviors. Second edition, Plenum Press, New York, USA, pp. :25-
38. 
DiClemente, c.c. and Prochaska, J.O. ( 1982). Self-change and therapy change of smoking 
behavior: a comparison of processes of change in cessation and maintenance. Addictive 
Behaviors, 7 (2): 133- 142. 
DiClemente C.C., Prochaska J.O. (1998). Toward a Comprehensive Transtheoretical Model of 
Change. In: Miller W.R., Heather N. (eds) .  Treating Addictive Behaviors. Second edition, 
Plenum Press, New York, USA, pp.:3-24. 
Dryden, G. and Vos, J. ( 1993). The Learning Revolution. A lifelong learning programme for the 
world's finest computer: your amazing brain. Profile Books, Wellington, New Zealand. 
Durell, J. and Buoski, W. ( 1984). Preventing Substance Abuse: The State of the Art. Public Health 
Report, 99 ( 1 ): 23-3 1 .  
Fairgray, J. D .  M. ( 1979). The Influence of Farm Advisory Officers i n  the Diffusion o f  Agricultural 
Innovations. Unpublished PhD Thesis, University of Auckland, Auckland, New Zealand. 
Fairweather, J. R. ( 1992). Agrarian restructuring in New Zealand. Lincoln University Research 
Report 213 .  Lincoln University, Canterbury, New Zealand. 
Fairweather, J.R. and Keating, N.C. ( 1994). Goals and Management S tyles of New Zealand 
Farmers. Agricultural Systems, 44: 1 8 1 -200. 
Fell, L.R. (2000). Time to converse: the importance of language, conversation and electronic media 
in agricultural extension. Australian Journal of Experimental Agriculture, 40 (4): 503-509. 
Fiore, M.C. ,  Smith, S.S., Jorenby, D .E. and Baker, T .B .  ( 1994). The effectiveness of the Nicotine 
Patch for S moking Cessation: A Meta analysis. Journal of the American Medical Association, 
271 (24): 1940-1947. 
44 
Aiegel, F.C., Kivlin, I.E. and Sekhon, G.S. ( 1968). A Cross-National Comparison of Farmers' 
Perception of Innovation as Related to Adoption Behavior. Rural Sociology, 3 3 :  437-499. 
Foundation for Research Science and Technology ( 1 998). Technological Learning and Knowledge 
Application Review. FRST, 
Frazer, A. (2000). Seeking a different view through workshops - a meat industry experience. In: 
Stantiall JD, McDiarmid CM (eds.), Achieving Change Through Improved Knowledge 
Systems. Centre for Agricultural and Veterinary Continuing Education, Massey University, 
Palmerston North, New Zealand, pp. 305-307. 
Garland, C. ( 1 993). Technology transfer: systems used by a Wairarapa farm consultancy firm. 
Proceedings of the New Zealand Grassland Association, 55: 7-9. 
Gasson, R. ( 1 973). Goals and Values of Farmers. Journal of Agricultural Economics, 24: 5 2 1 -537. 
Gavigan, R.G. and Parker, W.J. ( 1 996). Information for wool growers. Rural Business, 2 :  30-32. 
Ghosh, M. ( 1 993). Test of trickle-down hypothesis in rural West Bengal. Indian Journal of 
Agricultural Economics, 48 (2): 216-225. 
Godfrey C., Maynard A. ( 1 995). The economic evaluation of alcohol policies. In: Holder H.D., 
Edwards G.E. (editor). Alcohol and public policy: Evidence and issues. Oxford University 
Press, Oxford, England, pp. :238-259. 
Greer, I. ( 1 982). Motivation and Other Factors Influencing the Adoption of Practices by Sheep 
Farmers in Oxford County. Unpublished Master of Agricultural Science Thesis, Lincoln 
College, New Zealand. 
Greer, I. and Greer, G. (1996). Experiences of Agricultural Extension. MAF Policy Information 
Paper No. 1 2, Wellington, New Zealand. 
Griffiths, M. ( 1 994). Social marketing: Achieving changes in nutrition behaviour, from household 
practices to national policies. Food and Nutrition Bulletin, 1 5  ( 1 ):  25-3 1 .  
Grunert K.G., Harmsen H., Larsen H.H., S�rensen E., Bisp S. ( 1 997). New Areas in Agricultural 
and Food Marketing. In: Wierenga B., van Tilburg A., Gruntert K., Steenkamp J.E.M., Wedel 
45 
M. (eds). Agricultural Marketing and Consumer Behavior in a Changing World. Kluwer 
Academic Publishers, Boston, USA, pp. :3-30. 
Guerin, L.J. and Guerin, T.F. ( 1994). Constraints to the adoption of innovations in agricultural 
research and environmental management: a review. Australian Journal of Experimental 
Agriculture , 34 (4): 549-571 .  
Gutman, J .  ( 1982). A Means-End Chain Model based on Consumer Categorization Processes. 
Journal of Marketing, 46 (2): 60-72. 
Haaga D.F., Kirk L. ( 1998). Strategies for Tobacco Cessation. In: Miller W.R., Heather N. (eds). 
Treating Addictive Behaviors. Second edition, Plenum Press, New York, USA, pp.:245-258. 
Hall, M.H. and Kuiper, D. (1998). Commercialization and Privatization of Agricultural Extension: 
The New Zealand Experience. Journal of Production Agriculture, 1 1  ( 1 ): 135- 140. 
Hall, M.H., Morriss, S .D. and Kuiper, D. ( 1999). Privatization of Agricultural Extension in New 
Zealand: Implications for the Environment and Sustainable Agriculture. Journal of Sustainable 
Agriculture, 14 ( 1 ) :  59-71 .  
Hannibal, M.  and Sriskandarajah, N .  ( 1996). Systemic change: Rational and emotional elements. 
Proceedings of the New Zealand Grassland Association, 57: 5 -8. 
Harper, W.M. and Eastman, C. ( 1980). An Evaluation of Goal Hierarchies for Small Farm 
Operators. American Journal of Agricultural Economics, 62: 741-747. 
Haug, R. ( 1999). Some leading issues in international agricultural extension, a literature review. 
The Journal of Agricultural Education and Extension, 6 (2): 263-274. 
Hawke G. ( 1987). Overview of New Zealand agriculture. In: Wallace L.T., Lattimore R. (eds). 
Rural New Zealand what next? Canterbury, New Zealand Agribusiness and Economics 
Research Unit, Lincoln College, pp. :  1 . 1  - 1 . 1 1 .  
Hays, J.T., Croghan, LT., Schroeders, D.R. et al. ( 1999). Over-the-Counter Nicotine Patch Therapy 
for Smoking Cessation: Results from Randomized, Double-Blind, Placebo-Controlled, and 
Open Label Trials. American Journal of Public Health, 89 (1 1) :  1701 - 1707. 
46 
Hightower, J. (1973). Hard Tomatoes, Hard Times: the Failure of America's Land-Grant College 
Complex. Schenkman, Cambridge, Massachusetts. 
Horgen K.B., Brownell K.D. (1998). Policy Change as a Means for Reducing the Prevalence and 
Impact of Alcoholism, Smoking and Obesity. In: Miller W.R, Heather N. (eds). Treating 
Addictive Behaviors. Second edition, Plenum Press, New York, USA, pp.: 105-1 18. 
Hurt, RD., Lowell, D.C., Fredrickson, P.A. et al. (1994). Nicotine Patch Therapy for Smoking 
Cessation Combined with Physician Advice and Nurse Follow-up: One Year outcome and 
Percentage of Nicotine Replacement. Journal of the American Medical Association, 27 1 (8): 
595-600. 
Jackson, L. and Caffarella, RS. (1994). Experiential learning: a new approach. Jossey-Bass 
Publishers, San Francisco. 
Janssen W.G., Tilburg A.V. (1997). Marketing Analysis for Agricultural Development: 
Suggestions for a new Research Agenda. In: Wierenga B., van Tilburg A., Gruntert K., 
Steenkamp J.E.M., Wedel M. (eds). Agricultural Marketing and Consumer Behavior in a 
Changing World. Kluwer Academic Publishers, Boston, USA, pp. :57-74. 
Jha, P. and Chaloupka, FJ. (2000). The economics of global tobacco control, clinical review. 
British Medical Journal, 321 :  358-361 . 
Journeaux, P. (1990). Adoption of sheep and beef improvement technologies relative to the 
adoption of dairy beef improvement techniques, Intelligence report. MAF Rural Policy Unit, 
Wellington, New Zealand. 
Kelly, T., Kuiper, D., Macgregor, N. and Renquist, R (2000). Facilitating ecological agriculture in 
New Zealand. In: Stantiall JD, McDiarmid CM (eds.), Achieving Change Through Improved 
Knowledge Systems. Centre for Agricultural and Veterinary Continuing Education, Massey 
University, Palmerston North, New Zealand, pp. 93-99. 
Kelman H.C. (1967). Compliance, Identification, and Internalization: Three Processes of Attitude 
Change. In: Fischbein M. Ed. Readings in Attitude Theory and Measurement. John Wiley & 
Sons, Inc., New York, USA, pp.:469-476. 
47 
Kerr, P. (2000). That, that is free, is perceived to have no value. In: Stantiall ID, McDiarmid CM 
(eds.), Achieving Change Through Improved Knowledge Systems. Centre for Agricultural and 
Veterinary Continuing Education, Massey University, Palmerston North, New Zealand, pp. 
269-273 . 
Keys, MJ. and Orchard, P.W. (2000). The Prime Pasture Program - using marketing concepts to 
achieve technology transfer. Australian Journal of Experimental Agriculture, 40 (4): 541-546. 
Klesges, R.C., Ward, K.D. and DeBon, M. ( 1 996). Smoking cessation: A successful behaviorall 
pharmacological interface. Clinical Psychology, 45 : 361 -366. 
Kloppenburg, J. ( 1 99 1 ). Social Theory and the Deconstruction of Agricultural Science: Local 
Knowledge for an Alternative Agriculture. Rural Sociology, 56: 5 1 9-548. 
Kolb, D. A. ( 1 984). Experiential Learning. Prentice-Hall, New Jersey. 
Kung, E. ( 1984). Rural development in the Third World. Universitas, 39 (2): 1 37-147. 
Leshner A.I. (2000) NIDA Director. Addiction research can provide scientific solutions to the 
problem of cigarette smoking, NIDA Notes, Volume 1 3, No. 3, 
http://165 . 1 12.78.6 1INIDA NoteslNNVo1 13N3lDirRepVo1 1 3N3.html 
Lichtenstein, E. ( 1 982). The smoking problem: A behavioral perspective. Journal of Consulting and 
Clinical Psychology, 50: 804-8 19. 
Ludwig, lA. and Marsden, S. (1993). Adopting new Technologies for grazing lands: the role of 
decision systems. Proceedings of the XVII International Grassland Congress, 635-637. 
Lynch, T., Gregor, S. and Midmore, D. (2000). Intelligent support systems in agriculture: how can 
we do better? Australian Journal of Experimental Agriculture, 40 (4): 609-620. 
MacClean, J., Penno, J. and Howse, S. ( 1 997). Current information and technology priorities of 
dairy farmers - a challenge for agricultural researchers. Proceedings of the New Zealand 
Grassland Association, 59: 2 1 5 -220. 
Marlatt, G. A. and Gordon, J.R.E. ( 1 985). Relapse prevention: Maintenance strategies in addictive 
behaviour change. Guilford, New York. 
48 
McArthur A ( 1987). Innovation, Infonnation and Communication. In: Wallace L.T., Lattirnore R. 
(eds). Rural New Zealand what next? Canterbury, New Zealand Agribusiness and Economics 
Research Unit, Lincoln College, pp.:3 . 1  - 3. 10. 
McKenzie, J. and Morris ,  R. S .  ( 1995). Development of spatial predictors of tuberculosis infection 
risk in wildlife populations and their application within a decision support framework. In: 
Griffin F, de Lisle G (eds.) Tuberculosis in Wildlife and Domestic Animals - Otago 
Conference Series No. 3 .  University of Otago Press, Dunedin, pp. 168- 170. 
McMeekan c.P. ( 1963). Research and its application. in: Moriarty MJ. (editor). New Zealand farm 
production and marketing. Oxford University Press, London, UK, pp. :30-45. 
McRae, AF. ( 1993). As applied agricultural scientists we serve . . . .  ? Proceedings of the New 
Zealand Society of Animal Production, 53: 107- 1 10. 
McRae, AF., Anderson, F.M. and Brazendale, R. ( 1993). Applying farmer first research in New 
Zealand. Proceedings of the XVII International Grasslands Congress, I: 637-639. 
Miller W.R. ( 1998). Enhancing Motivation for Change. In: Miller W.R., Heather N. (eds).  Treating 
Addictive Behaviors. Second edition, Plenum Press, New York, USA, pp. : 12 1 - 132. 
Morris, C., Loveridge, A. and Fairweather, J.R. ( 1995). Understanding why farmers change their 
farming practices: the role of orienting principles in technology transfer. Research Report 
No.232. Agribusiness and Economics Research Unit, Lincoln University, Canterbury, New 
Zealand. 
Mu1cock, c., Ogle, G. and Aubrey, B. (2000). Improving Methodologies for Farmers to use 
Environmental Infonnation. In: Stantiall ID, McDiarmid CM (eds.), Achieving Change 
Through Improved Knowledge Systems. Centre for Agricultural and Veterinary Continuing 
Education, Massey University, Palmerston North, New Zealand, pp. 101-103. 
New Zealand Department of Statistics ( 1940- 1995). New Zealand agriCUltural statistics 1940- 1995. 
NZDS, Wellington, New Zealand. 
Ni Mhurchu, c., Margetts, B.M. and Speller, V.M. ( 1997). Applying the Stages-of-Change Model 
to Dietary Change. Nutrition Reviews, 55 (1) :  10- 16. 
49 
Nightingale, T. ( 1992). White Collars and Gumboots: A History of the Ministry of Agriculture and 
Fisheries 1892-1992. Dunmore, Palmerston North, New Zealand. 
Norcross, J.e. and Vangarelli, DJ. ( 1989). The resolution solution: Longitudinal examination of 
New Year's change attempts. Journal of Substance Abuse, 1 :  1 27-134. 
Ockene, J.K. and Zapka, J.G. ( 1997). Physician-based smoking intervention: A rededication to a 
five step strategy to smoking research. Addictive Behaviors, 22 (6): 835-848.  
Oliver, J . ,  AlIen, W . ,  Kilvington, M. and Gilbert, M. (2000). The Role of Animal Health Board's 
Locally Initiated Programmes in Changing Farmer Attitudes. In: Stantiall ID, McDiarmid CM 
(eds.), Achieving Change Through Improved Knowledge Systems. Centre for Agricultural and 
Veterinary Continuing Education, Massey University, Palmerston North, New Zealand, pp. 
343-349. 
OECD ( 1988). Agricultural policies, markets, and trade: monitoring and outlook. Organisation for 
Economic Cooperation and Development, Paris. 
Paine, M.S. ( 1993). Extension agents can perform more effectively through an appreciation of 
individual learning styles. Proceedings of the New Zealand Society of Animal Production, 53:  
1 15- 1 19. 
Parminter, T.G. and Perkins, A.M.L. (1997). Applying an understanding of farmers'  values and 
goals to their farming style. Proceedings of the New Zealand Grassland Association, 59: 107-
1 1 1 . 
Parrninter, T.G., Power, M.P.B. and Shaw, D.e. ( 1993). The effects on adoption of beef breeding 
cow technologies of selected farm business standards, farmer demographics, and farmer 
beliefs. Proceedings of the New Zealand Society of Animal Production, 53 :  12 1 - 125.  
Paterson, B .M. and Morris, R . S .  ( 1995). Interactions between beef cattle and simulated sedated 
possums on pasture. New Zealand Veterinary Journal, 43 (7): 289-293. 
Pearson, G., Barugh, L, Morel, P.,  Smallwood, J.  and Kerr, L. (2000). Achieving through 
participation. In: Stantiall ID, McDiarmid CM (eds.), Achieving Change Through Improved 
Knowledge Systems. Centre for Agricultural and Veterinary Continuing Education, Massey 
University, Palmerston North, New Zealand, pp. 309-3 14. 
50 
Peet, K. and Peet, 1. ( 1995). Global learning for Sustainability - with people's wisdom. New 
Zealand Journal of Adult Learning, 23 (2): 55-69. 
Peterson, D.E., Zeger, S .L., Remington, P.L. and Anderson, H.A. ( 1992). The effect of state 
cigarette tax increases on cigarette sales, 1955 to 1988. American Journal of Public Health, 82: 
94-96. 
Praat, J.P., Bollen, F., Yule, I. and Empson, C. (2000). Applied infonnation management systems. 
In: Stantiall JD, McDiarmid CM (eds.), Achieving Change Through Improved Knowledge 
Systems. Centre for Agricultural and Veterinary Continuing Education, Massey University, 
Palmerston North, New Zealand, pp. 28 1-287. 
Prochaska, 1. O. and DiClemente, C.C. ( 1984). The transtheoretical approach: Crossing traditional 
boundaries of change. Dorsey Press, Homewood, ll.." USA. 
Prochaska, J.O., DiClemente, c.c. and Norcross, J.c. ( 1992). In search of how people change: 
applications to addictive Behaviors. American Psychologist, 47 (9): 1 102- 1 1 14. 
Prochaska, 1.0., Vlicer, W.F., Rossi, 1.S. et al. ( 1994). Stages of Change and Decisional Balance 
for 1 2  Problem Behaviors. Health Psychology, 13  ( 1 ): 39-46. 
Pyke, N. and Johnstone, 1. (2000). Knowledge in the arable industry - responding to need. In: 
Stantiall JD, McDiarmid CM (eds.), Achieving Change Through Improved Knowledge 
Systems. Centre for Agricultural and Veterinary Continuing Education, Massey University, 
Palmerston North, New Zealand, pp. 35-37. 
Rayner T. (1990). The seed of change. In: Sandrey R., Reynolds R. (eds). Farming without 
subsidies. Wright and Carman, Upper Hutt, New Zealand, pp. :  1 3-23. 
Reid, J.I., Coulson, J.L. and Cameron, E.A. (1996). A Framework for Understanding the Adoption 
and Use of Technologies by Dairy Farmers. Third Annual Conference "New Zealand 
Agriculture in the 2 1 st Century" ;  New Zealand Agricultural Economics Society, Discussion 
Paper No. 144, 209-216. 
Reid, 1.1., McRae, AJ. and Brazendale, R. (1993).  Farmer First Research: A review of phase one 
results in relation to farmers' willingness and ability to change. Proceedings of the New 
Zealand Grassland Association, 55: 17-21 .  
5 1  
Rhodes, A.P. and Aspin, M.D. ( 1 993). MRDC Monitor Farms - using information. Proceedings of 
the New Zealand Grassland Association, 55: 23-26. 
Ritchie, 1. ( 1 995). From the public to the private sector - The Agriculture New Zealand story. 
Agricultural Science, 8 (4): 29-3 1 .  
Rivera, W.M. (2000). The Changing Nature of Agricultural Information and The Conflictive 
Global Developments Shaping Extension. The Journal of Agricultural Education and 
Extension, 7 ( 1 ) :  3 1 -4 1 .  
Rogers, E .  M .  ( 1986) .  Models of knowledge transfer: critical perspectives. In: Beal, G.M., 
Dissanayake, W., Konoshima, S. (eds.), Knowledge generation, exchange, and utilization. 
Westview Press, B oulder, Colorado, USA, pp. 3 7-60. 
Rogers, E. M. ( 1 995). Diffusion of lnnovations (4th ed). The Free Press, New York. 
Rating, N. ( 1 988). Extension Science. Cambridge University Press, Cambridge, UK. 
Raling, N. and Engel, P. G. H. ( 1 99 1 ). (Chapter12:) The Development of the Concept of 
Agricultural Knowledge Information Systems (AKIS): Implications for Extension. In: Rivera, 
W. M. and Gustafson, D. J. (eds.), Agricultural Extension: Worldwide Institutional Evolution 
and Forces for Change. Elsevier, Amsterdam, pp. 1 25- 1 37 .  
Rush, M. (2000). Beyond the information delivery trap: Designing effective change programmes 
for sustainable agriculture. In: Stantiall JD, McDiarrnid CM (eds.), Achieving Change 
Through Improved Knowledge Systems. Centre for Agricultural and Veterinary Continuing 
Education, Massey University, Palmerston North, New Zealand, pp. 105- 1 2 1 .  
Rwenyagira, B .  ( 1 985). Farmers' Group Organisation and Operation for Effective Agricultural 
Extension . Unpublished MA Thesis, University of Canterbury, University of Canterbury, 
New Zealand. 
Sauter, C.M. and Morris, R.S. ( 1 995). Behavioural studies on the potential for direct transmission 
of tuberculosis to farmed livestock from feral ferrets (Mustela foro) and possums (Trichosurus 
vulpecula). New Zealand Veterinary Journal, 43 (7): 294-300. 
52 
Schachter, S. ( 1982). Recidivism and self-cure of smoking and obesity. American Psychologist, 37: 
436-444. 
Schmid, T.L., Jeffrey, RW. and Hellerstedt, W.L. ( 1989). Direct mail recruitment to home-based 
smoking and weight control programs: A comparison of strengths. Preventive Medicine, 18 :  
503-5 17.  
Schramm W. ( 1973). Channels and Audiences. In: Pool LD.S., Frey F.W., Schramm W., Maccoby 
N., Parker E.B. (eds.) Handbook of Communication. Rand McNally College Publishing 
Company, Chicago, USA, , pp. : 1 16-140. 
Sherman, H. and Schultz, R (1998). Open boundaries - creating business innovation through 
complexity. Perseus Books, Reading, Massachusetts. 
Silagy, c., Mant, D., Fowler, G. and Lodge, M. ( 1994). Meta-analysis on efficacy of nicotine 
replacement therapies in smoking cessation. The Lancet, 343: 1390142. 
Silvestri, B. and Flay, B.R (1989). Smoking Education: Comparison of Practice and State-of-the­
Art. Preventive Medicine, 18 :  257-266. 
Sporny, L.A. and Contento, LR (1995). Stages of Change in Dietary Fat Reduction: Social 
Psychological Correlates. Journal of Nutrition Education, 27 (4): 191- 199. 
Stantiall, J. (2000). Affecting the rate of change: Awareness of information or quantum learning? 
In: Stantiall JD, McDiarmid CM (eds.), Achieving Change Through Improved Knowledge 
Systems. Centre for Agricultural and Veterinary Continuing Education, Massey University, 
Palmerston North, New Zealand, pp. 145- 158.  
Stantiall, J. D. ( 1999). Facilitating learning: enhancing dairy farmer competence through 
workshops. Unpublished Master of Applied Science in Agricultural Extension thesis , Massey 
University, Palmerston North, New Zealand. 
Steenkamp J.E.M. ( 1997). Dynamics in consumer behavior with respect to agricultural and food 
products. In: Wierenga B., van Tilburg A., Gruntert K., Steenkamp J.E.M., Wedel M. (eds). 
Agricultural Marketing and Consumer Behavior in a Changing W orId. Kluwer Academic 
Publishers, Boston, USA, pp.: 143- 188. 
53 
Stellman, S .D. and Garfinkel, L. ( 1989). Lung Cancer Risk is Proportional to Cigarette Tar Yield: 
Evidence from a Prospective Study. Preventive Medicine, 1 8: 5 18-525. 
Stephens, T., Pederson, L.L., Koval, 1.1. and Kim, C. ( 1997). The Relationship of Cigarette Prices 
and No-Smoking Bylaws to the Prevalence of Smoking in Canada. American Journal of 
Public Health, 87 (9): 1519- 152 1 .  
Stewart, D .  E. (1979). Innovativeness and Communications: a Study of Behaviour and Attitudes 
Among Farmers in Otago, New Zealand. Unpublished PhD Thesis, University of Otago, 
Dunedin, New Zealand. 
Stichbury J. ( 1994). Better dairy farming - the Consulting Officer Service. In: Davis G.H. (editor). 
Lab coats to gumboots. Wellington, New Zealand New Zealand Society of Animal 
Production, pp. :  1 13-140. 
Stoddard, J.J. and Gray, B. ( 1997). Maternal Smoking and Medical Expenditures for childhood 
Respiratory Illness. American Journal of Public Health, 87 (2): 205-209. 
Sutton, S. ( 1996). Can 'stages of change' provide guidance in the treatment if addictions? A critical 
examination of Prochaska and DiClemente's model. In: Edwards,  G. and Dare, C. (eds.), 
Psychotherapy, Psychological Treatments and the Addictions. Cambridge University Press,  
Cambridge, UK, pp. 189-205. 
Tarbotton, I.S., Paine, M.S., Wedderburn, M.E. and Carlyon, M. ( 1997) .  A method for improving 
farm practice: a case study of pugging. Proceedings of the New Zealand Grassland 
Association, 59: 79-85. 
Thompson, E.L. ( 1978). Smoking Education Programs 1960-1 976. American Journal of Public 
Health, 68 (3): 250-257. 
Timmons Roberts, J. (1995). Trickling down and scrambling up: the informal sector, food 
provisioning and local benefits of the Carajas mining "growth pole" in the Brazilian Amazon.  
World Development, 23 (3): 385-400. 
Triandis, H. C. ( 1971) .  Attitude and Attitude Change. John WHey & Sons, Inc., New York. 
54 
Tyler L., Lattimore R ( 1 990). Assistance to agriculture. In: Sandrey R ,  Reynolds R (eds). 
Farming without subsidies. Wright and Carman, Upper Hutt, New Zealand, pp. :60-79. 
Valentine, I., Hurley, E.M. and Glass, W. ( 1 993). Goals and management strategies of dairy 
farmers. Proceedings of The New Zealand Society of Animal Production, 53 : 1 1 1 - 1 1 3 .  
Verkerk, G . ,  Burke, c.,  Paine, M. and Jolly, P .  (2000). Factors influencing the uptake of controlled 
breeding technologies on dairy farms. In: Stantiall ID, McDiarmid CM (eds.), Achieving 
Change Through Improved Knowledge Systems. Centre for Agricultural and Veterinary 
Continuing Education, M assey University, Palmerston North, New Zealand, pp. 1 83- 1 85 .  
Walker, A .  and Bell, B .  ( 1994). Aspects of New Zealand's Experience i n  Agricultural Reform since 
1984. MAF Policy Technical Paper 94/5, Wellington, New Zealand. 
Wassell, T.R and Esslemont, RJ. (1992). Herd health schemes, their scope and use by dairy 
farmers. Farm Management, 8 (4): 194-200. 
Weber, B.A. ( 1 974). Trickling down: the responsiveness of rural and rural poor family income and 
labour supply to regional economic growth. Dissertation Abstracts International, 34 (9) : 5428. 
Wegener, P., Frank, B. and Maroske, M. (2000). A Planning and Self-Evaluation process for dairy 
farmer discussion groups. In: Stantiall ID, McDiarmid CM (eds.), Achieving Change Through 
Improved Knowledge Systems. Centre for Agricultural and Veterinary Continuing Education, 
Massey University, Palmerston North, New Zealand, pp. 77-86. 
Wyllie, D. ( 1 989). Vocational Agricultural Education in New Zealand. Agricultural Progress, 64: 
8 1 -88. 
55  
Beef farm in  the Wairarapa 
Beef farm in summer (photograph courtesy Fiona Dickinson) 
56 
57 
Chapter 2 
Analysis of Wairarapa farmer perceptions of tuberculosis 
and management options for control 
58 
59 
Abstract 
The Australian brushtail possum is the main vector for bovine tuberculosis in New Zealand. Thus 
control and eradication of tuberculosis is complicated, both livestock and vector control 
populations have to be controlled. The main control method to eradicate TB from vector 
populations is poisoning of possums, which is expensive and cannot be conducted in all areas with 
infected vector populations. Therefore, more active involvement of farmers in the control of TB is 
desired and expected. 
A survey of 27 farmers in the Wairarapa, whose cattle herd were infected with TB, was conducted 
using an open-ended questionnaire. Grounded theory was used to identify key factors in the farm 
management relating to TB and its control. The key theories found were: Farmers knew or had 
suspicions about potential areas on their farms where their cattle were likely to become infected 
with TB. Cattle herds that grazed in such areas had a higher TB reactor incidence than farms where 
cattle were excluded from grazing these areas. A major role in decision making regarding on-farm 
TB control were economics. On-farm TB control was only sporadic and motivation was lacking or 
wore off quickly. Information sources were mostly people that visited the farm on animal health 
business (livestock officers, Regional council staff and veterinary practitioners). Farmers differed 
in their perception of the TB related knowledge of livestock officers, some farmers found them 
very knowledgeable, while for others they did not meet the requirements. Grazing management 
was flexible on beef farms, less on dairy farms. It was influenced by the number of stock and 
pasture shortage. These findings were used to develop on-farm control methods that exclude cattle 
from grazing TB hot-spot areas during certain high risk times. 
Introduction 
The control and eradication of bovine tuberculosis in New Zealand is  complicated by the existence 
of a wildlife-reservoir, the Australian brushtail possum (Trichosurus vulpecula) (Morris and 
Pfeiffer, 1995). Therefore, tuberculosis control has to comprise disease and vector control (Animal 
Health Board, 1 995). The main control methods for tuberculosis currently are farm animal testing, 
and reduction of infected possum populations on or near farms, involving trapping, ground 
poisoning and aerial poisoning operations. 
The Animal Health Board, the national organisation responsible for TB control strategies, produced 
a national pest management strategy (PMS) to operate for the period 1995/96 to 2000/01 ,  which is  
60 
currently under review. Priority is given to "encouragement and assistance to farmers in infected 
areas to improve their disease status" (Animal Health Board, 1 995). To develop TB control 
programmes for farmers to improve their disease situation, it is necessary to have detailed 
knowledge of existing farm management practices and to assess the potential for incorporating TB 
control methods into these management practices. 
It is essential that the initial hypotheses in any study be based on appropriate assumptions, in order 
to avoid statistically significant but irrelevant results (Boland and Morris, 1 988). Especially when 
human behaviour is involved, assumptions have to be checked for appropriateness. Much scientific 
research has been based on methodologies in which pre-existing knowledge and theories about the 
subject matter are the base for formulating hypotheses, which are then tested experimentally. In 
contrast, the research methodology 'grounded theory' investigates the subject of interest without 
requiring previously fonnulated hypotheses. It aims to generate theoretical concepts rather than to 
verify or invalidate them (Glaser and Strauss, 1 967). 
The present study was the baseline for designing the on-farm TB control programme described in 
Chapter 3. The first objective was to obtain details of farmers' observations on TB in both cattle 
and wildlife on their farm, identify farm grazing strategies, examine relationships between farm 
management practices and TB infection in cattle, and gain insights into farmers' attitudes towards 
TB and its control. The second objective was to identify the sources of infonnation that were used 
by the majority of farmers. The third objective was to develop hypotheses for using farm and 
grazing management for TB control purposes. These hypotheses were then tested for effectiveness 
in on-farm TB control methods (see Chapter 3). 
Materials and Methods 
Grounded theory 
This section justifies the selection of 'grounded theory' as an analytical method for this study. 
'Grounded theory' was developed by social scientists for behavioural research to reduce the 
subjectivity of data and the inherent uncertainty of its interpretation. Glaser and Strauss ( 1 967) 
described two such tools; a specific technique for content analysis of in-depth interview data, and a 
procedure for constant comparison of interview events. These tools are then used to create a 
substantive theory grounded in the fieldwork data (Boland and Morris, 1 988; Strauss and Corbin, 
1 990). In other words, the researcher does not start with a pre-fonned hypothesis to test about 
behaviour, but rather listens to the responses from subjects and builds an understanding of how and 
61  
why decisions are made, which is derived directly from the views presented by subjects. These 
qualitative techniques can be used appropriately in the initial investigation of disease control and 
management. Boland and Morris ( 1988) used this technique to study the way in which veterinary 
practitioners search for information with regard to veterinary innovations. 
The 'grounded theory' approach is a style of qualitative analysis that includes its own distinct 
methods and techniques, such as theoretical sampling and the use of specific coding procedures .  It 
also has methodological guidelines, such as the use of constant comparison and the use of a coding 
paradigm (Glaser and Strauss 1 967, Strauss 1987, Strauss & Corbin 1990). Creativity and insight 
are vitally important for developing a sensitive interpretation of the data as the number of interview 
subjects grows (Glaser and Strauss, 1 967) .  
Generating a theory from data means that most hypotheses and concepts are not only derived from 
the data, but also they are all systematically worked out in relation to the data during the course of 
analysis by a process of constant comparison, where generation of hypotheses, concepts and 
categories are continuously 'checked' with the data itself (Glaser and Strauss, 1967). 
Interviewing and the use of questionnaires are the primary technique of data collection when using 
grounded theory methods (Strauss and Corbin, 1990). Data collection, analysis, and interpretation 
occur simultaneously (also called 'theoretical sampling') .  Analysis often involves techniques such 
as open, axial and selective coding (Strauss and Corbin, 1990; Adams et a!., 1999). Categories are 
developed by finding and comparing code words that describe similar phenomena. Testing is built 
into every step of the process and negative cases are important, as they contribute to possible 
variation (Strauss and Corbin, 1990). Generally sampling is continued until 'theoretical saturation' 
is reached in each category. This occurs when no new data or new concepts relating to the central 
problem, emerge with additional interviews (Strauss and Corbin, 1990). Emphasis is put on the 
validity of the data collected in representing the issue of interest, rather than its statistical 
significance. 
Analysis of qualitative data 
Qualitative data is often complex, because the goal is to learn new things, and/or to understand the 
complexity of situations. Efficiency of the researcher can be supported by computers through 
coding, storing information, counting, and searching text. Creativity, such as data exploration, 
reflection on data, construction of categories, and theory building are the real challenge for any 
qualitative analysis, and hence for a software package, which might be used with such data. 
Flexibility (such as several open windows), quick retrieval of data for iterative processes in theory 
building and linking of data (e.g. data to categories, data to ideas, memos to theories or text to other 
62 
interview documents) are all main components of grounded theory facilitated by computers 
(Richards and Richards, 1 994). 
Qualitative computing is not only a new way of doing the same things as previously done 
manually, but the use of computers has also changed the methods, by adding new techniques and 
offering additional features for old techniques, such as illustrating interpretation of data on the 
computer screen (Richards, 1 997). 
Richards (1997) concludes that doing "qualitative research without a computer now would be 
rather like doing statistical research with an abacus". She compares qualitative software to an 
innovative cookbook, which should be used as a collection of possibilities, not requirements. 
As a fIrst step it was necessary to identify farm management practices, grazing schemes and stock 
policies, which were apparently associated with TB in cattle, and therefore identify potential risk 
factors. The study concentrated on the three main farm types: dairy (DH), beef breeding (BB), and 
beef fInishing (beef dry stock, BD) farms in the Wairarapa. As a second step it was important to 
evaluate if grazing management was flexible enough to be altered for TB control purposes. The 
Wairarapa region was chosen for its known M. bovis infection in wildlife and because the 
intervention study to evaluate on-farm TB control progranunes (Chapter 3) was conducted in this 
region. 
Selection of farms 
No specifIc sample size or composition was set at the beginning of the investigation. As one of the 
aims of the study was to evaluate grazing management in relation to TB infection, purposive 
selection was employed. Two areas in the Wairarapa region were chosen for the study (Feathers ton 
and Tinui). In both areas the Regional Council staff were conducting wildlife surveys, either as part 
of their annual vector control programme or specifIcally to identify TB hot-spots. As the on-farm 
control progranunes were intended to be employed by fanners having a possum-related TB 
problem, it was essential to assess the fann situation under circumstances where TB was known to 
be present in the wildlife. With the help of AgriQuality (formerly MAF Quality Management) 
veterinary officers, all fanners in these two areas were approached for the study. Only one fanner 
refused to participate in the study. One additional fann in the Wellington region was chosen, as TB 
hot-spots were already known on this fann. This purposive sampling approach ensured that all the 
fanns in the study had a high likelihood of their cattle being infected by possums. 
63 
Interview process 
Between March 1994 and February 1996, 27 farmers were interviewed about their TB herd history 
and grazing management. The farmers were interviewed with an extensive, open-ended 
questionnaire (see Appendix I, p. 327), which was tested on two fanners ftrst. Most interviews 
were organised in an open-ended way around topics such as grazing management, general farm 
management and tuberculosis. Farmers were thus able to give their views as they wished and refer 
to several questions at the same time. Mostly the questions were posed by raising key-words, rather 
than speciftc questions in order to obtain the farmers' perceptions/opinions in a comprehensive way 
on these subjects. All interviews were recorded on to audio tapes, to ensure minimal data loss and 
to preserve detail of responses. 
Content analysis 
The tapes were transcribed into text documents (using MS Word) and then imported into 
WinMAX98 (Copyright 1998, Udo Kuckartz, BSS, Berlin, www.winmax.de). a software 
programme for qualitative data analysis. The study was intended to generate rather than to test or 
verify theories and therefore the analysis focused on an exploratory and qualitative investigation of 
the TB situation and grazing management. 
The text documents were examined ftrst by the process of 'open coding' ,  where the data was 
broken into different labelled concepts (e.g. farmer possum control) (Strauss and Corbin, 1990). 
These concepts were then sorted to establish major themes and categories. The concepts were then 
grouped into these categories (e.g. TB control) (Phillips and Rempusheski, 1995). The categories 
were expanded, sub-categories developed and a tree-structure built (see Appendix IT, p. 33 1). 
Figure 1 shows a screen of the programme WinMAX98 in use, with the open text document on the 
right window and the codes used for TB history in the left window. The codes were arranged 
within a tree structure, indicating broad categories and their sub-categories. The lines and boxes in 
the text window show the coding that was assigned to the different text segments. Each section in 
the text was analysed for its meaning and assigned to at least one code. 
64 
, 
9-----l!} TB [3: 7] 
Contro! [1 :3} 
E Ii Grazing (2,7) � MotlVaton [1 1 :38 Possum [0:0. 
��d ��0�8 [1 2:2 
on wildlife ( 
farmer control [75: 
RC control [S 7: 1 55 
View on methods f 
$l Cost [0,0) 
1-- 1 control [20:57] 
L non-control releled (36: 1 ,  
farmer's percepuon [78:285] t=fii regerding farm monagerr 
regarding MAF;RC contre 
fermer's theories (1 6:42] 
$l history [1 5:24] 
hot-spots [4:1 4] �til areas [32:1 05J 
grazmg [1 4:53] 
TB cattle (8:26J � . " ... .... ... -
Bill cnd brother (T ed) were in partnership on the farm in the 
' 1 late 960$ end were gettlng M hard with TB at thoJ time, Tested 9 ! every 3 months, Were gettmg 1 0-1 5 reodors eech test. Highest 
9 number of reedors ot 0 test wes 27, which neorly put them out of 
i J business, In eorly 70s when Bill ond his wife took over the le.rm 
6 they were still getting M hard Wlth TB, Pest Board did some 'I poisomng which knocked the TB bock a bit. Form has alweys been e dairy farm, 
• In the late 50s - eorly 60s, Jock (local vet) put them onto the I role of possums spreading T8 PJ this bme MAF didn't believe 91 Jack that possums were carrying TB, Knew that pigs were ce.rrying 
TB os olot of pig hunting in the area. Derek. who was ferming 
(dOlry) John's property at the lime, wos losing huge numbers 
('dozens') of cows 0.\ eoch test when it first becom,e compulsory 
to TB test. He and Jack did big surveys of the possums on Derek's 
ferm and believed that possums were spreading T8 to the cettle. 
A cousin of Bill's wos dairy farming near Cross C and cetchmg 
2000-3000 possums 0 year os part of their living, (\Nera e.lso 
�TB.fe.rmer's perceptionjtheiderm), When one 01 the COUSinS left the 
farm, they stopped tropplng possums and the TB inadence in their 
cows put them out of bUSiness within 2 yeors - lost their herd, 
TB gredually moved sOI,.'th end their term bege.n to get hit Took 0 
COD1NGS f S :  ht C: OR 
Figure 1:  Text coding windows for TB history of one sample farm using the qualitative 
software programme WinMAX98, 
With this technique of coding text, similarities in the answers could be found without requiring 
exact match in tenninology as in a method using key word count. 
Impressions gained by the researcher and the significance of comments were noted in 'memos' , .  
which were text notes linked t o  specific comments made by fanners, or loosely linked with the 
whole text document. 
Within WinMAX98 all text segments relating to one or more specific codes, categories and/or 
memos could be retrieved. For each fann the most significant features were entered into 
MindManager® (version 3.5.5.; copyright 1999; www.mindmanager.com) for graphical display. 
Figure 2 presents the codes of the TB situation on a fann as an example of interrelating and 
displaying the events of the different categories. These Mind Maps together with the retrieved 
coded segments and memos were used to compare separate interviews. 
IlO real TB since then 
a with hardly TB 
rabbits. possums. hedg 
fem!tSistoatsleats 
not known 
sick possums in daytime see" infected 
potential time bomb 
... akest point in control 
coneemod that hedgehogs transmit TB TB . ' "  unwilling to 1"'0 WIth tt risk to nem. mar1< 
65 
70' huge 25'1. 01 herd 10 cows test clear. 4 in _s riddk!d 
lots 01 poisoning 
� /  control not gm. near bush 
above road 1'TlOf'e than in bush e!ose to gorse 
doesn� use bushPaddoeks for cattle 
Figure 2: Example of a Mind Map with categories and events used in describing the TB 
situation and TB perception of one farmer in the sample. 
Because the data set for this study was already complete at the start of the analysis, several long 
interviews were analysed first. Then the other interviews were used to challenge or expand the 
theoretical grasp, a method used by other researchers in the field of grounded theory (Kearney et 
aI. , 1 995). The final analysis therefore included the full range of variation represented in the study 
sample. 
The TB reactor situation for each of the farms was assessed using the interviews and the official 
herd testing history as recorded by AgriQuality, the national field veterinary service. More detailed 
information on regulations regarding TB testing, herd status, and movement of cattle can be found 
in Appendix III (p. 334). 
Results 
Descriptive analysis of study farms 
General farm characteristics 
Table 1 presents some characteristics of the study farms. Eleven of the farms were beef breeding 
farms, three beef finishing farms, 1 1  dairy farms and two farms had both a dairy and a beef 
breeding herd. For analysis purposes these latter two farms were included in the beef breeding 
66 
fanns, as their TB problem appeared to be more associated with the beef breeding herd than the 
dairy herd. 
Table 1. Mean [and range] of some characteristics of study farms in the Wairarapa. 
Dairy (n=l1) Beef breeding (n=13) Beef fInishing (n=3) 
Size of farm (ha)a 1 1 6.82 878.85 452.00 
[28 - 364] [48 - 31 03] [247 - 769] 
Effective area of farm 86.25 79.91 83.07 
(%)b [55.49 - 100] [45.54 - 1 00] [77.35 - 90.90] 
Size of run-off (ha)C 30.07 (n=13) 80.33 (n=3) 42.00 (n=1 ) 
[7 - 67] [40 - 146] 
Cattle Stock units (CSU) 1710.59 31 87.93 1 298.25 
[1 063.5 - 3054.5] [61 .5 - 1 2808.2] [1 094.5 - 1 588.0] 
Sheep Stock units (SSU) 562.5 (n=2) 3380.55 2023.33 
[321 - 804] [52.1 - 6501 .0] [1 245.0 - 3025.0] 
No. cattle � 2yr 21 9.45 391 .08 1 40.00 
[136 · 409] [1 1 - 1728] [98 - 1 82] 
a excludes run-offs or other leased areas 
b effective size indicates what percentage of the land area is used for grazing, the remainder is bush or 
unproductive land area; it excludes run-offs or other leased areas 
C run-offs are areas that generally are located some distance from the main farm. Some farms had multiple 
run-offs - hence the number is larger than the number of farms. 
Stock units were calculated from the number of animals present on the farm at the end of June, 
using the conversions provided by Flerning ( 1 996). Dairy cows were assumed to weigh an average 
of 450 kg and produce 165kg rnilkfat per year (the average value for dairy cows in the Wairarapa). 
TB history from existing records kept by AgriQual ity 
Records on reactor numbers and lesioned animals were those contained in the National Livestock 
Database (NLDB), kept by AgriQuality. 
The average cumulative incidence for five years ( 1990- 1994) was calculated for each farm type 
from TB herd history records kept by AgriQuality New Zealand, using the sum of animals with 
tuberculous lesions at slaughter (lesioned culls and lesioned skin-test positive animals) and the total 
number of animals tested over the five years. Beef breeding herds had the highest cumulative 
incidence with an average of 0.026 (range 0.00 1 2  to 0. 1 74), whereas dairy farms had the lowest 
with an average of 0.008 (range 0.00 to 0.024). Beef finishing herds had on average a cumulative 
incidence of 0.01 2  (range 0.0006 to 0.02). However, cumulative incidence was not significantly 
67 
dependent on farm type. One beef breeding farm with only 1 1  animals over 2 years of age, but a 
cumulative incidence of 0. 174 and a studentised residual of 4.61 was classified as an extreme 
outlier - all other BB farms had a cumulative incidence of less than 0.05. After removal of this 
outlier the ANOVA test of cumulative incidence and farm type yielded an F-statistic of 0.941 with 
p=O.4l .  This non-significance is presented in the overlapping ranges in the box-plot of the 
cumulative incidence for the three farm types (Figure 3). Cumulative incidence was also not 
dependent on the total effective farm size, nor on the cattle stock units on the farm (p>O.40) . 
0 .06.,---------------� 
G> 0 � .... 
G> u 0 .04 c G> "0 'u 
.s 
m 
I- possible outlier G> 0 0 > 
� 0 .02 75% "5 quartiles E ::s 25% u 
median 
I range 0 
Beef breeding Beef finishing Dairy 
(n=1 2) (n=3) (n= 1 1 )  
Farm Type 
Figure 3. Box-plot of five-year cumulative TB incidence rates (lesioned animals 1990-1994) 
for the three main farm types (excluding one extreme outlier in the beef breeding group with 
an incidence of 0.17). 
The cumulative incidence of TB lesioned animals was positively associated with the total number 
of whole herd tests conducted within the period 1990 to 1994 (F=2.761 ,  p=0.042, 6 and 19 df, after 
removal of the extreme outlier). B eef breeding farms had an average of 5.83 whole herd tests 
during the five-year period, beef finishing farms an average of 1 .67 and dairy farms an average of 
7.27 (F=12. 14, p=O.OOO, 2 and 23 df). 
The beef breeding farms in the study were infected for an average of 4.67 years within the five-year 
period, the beef finishing farms an average of 3 .00 years and dairy farms an average of 3.64 years 
(F=2.28, p=O. 125, 2 and 23 df). The number of years a herd was infected was not associated with 
68 
the five-year cumulative incidence for TB (F=O.95, p=0.472, 2 and 23 df), nor with the number of 
whole herd tests per herd over the five year period (F=2.63, p=O. 1 20, 2 and 23 df). 
Building theories using the interviews 
Many interview questions related to all three aspects of the study: tuberculosis, farm management 
and grazing management. The division into the three separate groups was more for logistical 
reasons and many points mentioned are interrelated with each other. Not all farmers commented on 
each of the point described in the following, explaining the discrepancies in the numbers. 
Tuberculosis related observations by farmers 
Cattle affected 
Three farmers mentioned that the reactor animals were mostly amongst the highest producing 
animals. In dairy herds these were the high producing milk cows and in beef herds or groups of 
young animals the biggest animals of the group. Nine farmers stated that all age groups of cattle 
were affected in their herds, whereas 1 2  farmers stated that confirmed TB infection was mostly in 
the older animals (cows and rising two-year heifers). 
Tuberculosis hot-spots 
In the following the term TB hot-spot is used for localised high tuberculosis risk areas for 
livestock. 
Nineteen farmers (ten BB, three BD, six OH) had either a suspicion or knowledge about certain 
habitat factors or localities on their farm where their cattle might have become infected with TB 
from possums. Sometimes no clear habitat factor was mentioned, only that the cattle were 
presumably infected in certain parts of the farm, such as the run-off. Suspected vegetation included 
bush, gorse (Ulex europaeus), lake edges, willows (Salix species), pine (Pinus radiata) plantations, 
dams, and swamp. Dairy farmers mostly suspected their run-off areas, which were often close to 
rivers and willows. Three farmers mentioned the possibility of having had infected possums in their 
hayshed, and cattle getting infected either by grazing around the hayshed or being fed with infected 
hay. 
"Had two bulls react several years ago when we were fattening cattle, and thought they'd 
got infected from hay - probably infected possums in the hayshed. Because the cattle had 
lesions in lymph nodes in the head. " 
69 
More beef farmers than dairy farmers (eight BB, one BD, two DH) considered a specific area of 
their farm as a TB hot-spot area, either because they had reactors out of this area, or because 
tuberculous possums were found there. Only four farmers (three BB, one DH) stated that many 
areas of the farm were considered hot-spots. These farms mostly experienced reactor cattle out of 
one area one year, and out of another area the following year. 
A greater proportion of dairy farmers than beef farmers (five DH, three BB) had no idea where 
their cattle picked up TB. These farmers could only hypothesise that their only areas with bush or 
trees on the farm could be possible hot-spots. Only one farmer stated that he also might have 
bought TB infected cattle. 
The TB incidence data suggested that farms with hot-spots within paddocks grazed by cattle 
experienced a higher tuberculosis reactor incidence than farms where there was no clear suspected 
hot-spot area or where hot-spots were fenced off from grazing areas (F=2.976, p=O.097, 1 and 24 
df). Eight farmers stated that they had a higher number of TB reactors in the group of animals that 
grazed the hot-spot area, than in other groups of animals. 
Transmission of tuberculosis to cattle 
Perceptions about transmission of TB not only included possum-ta-cattle transmission, but also 
cattle-ta-cattle transmission. Cattle density was mentioned as a possible influencing factor, such as 
when cattle congregated around water. Cattle-ta-cattle transmission was mentioned by five farmers, 
three of whom had experienced 'anergic' animals in the past. These were mostly older animals that 
had tested clear several times using the intradermal tuberculin test, but when sent to slaughter 
showed extensive lesions (mentioned on five of the farms). 
" Years ago TB was with the herd itself, that means we had one of those cows what they 
call 'anergic ' cow that wouldn 't react but would spread it round. A bad problem with dairy 
fanns. " 
Possum control 
All except four farms had received possum control by the Wellington Regional Council. Most of 
the 23 farmers who had received control work were pleased with the staff and the work they did, 
despite six of these farmers stating that the Regional Council should control their farms more often. 
The farmers that had received possum control by the Regional Council stated that as a consequence 
the possum population was reduced quite severely and that the TB reactor rate in cattle decreased. 
70 
"Since the Regional Council poisoned the possum in the last three years we have had very 
little TB, as has next door. " 
Eight farmers explicitly stated that they regarded the Regional Council programme as being the 
best option for possum control, partly as it had experienced staff. Some farmers were even 
reasoning that they themselves should not have to do any control, as they paid rates to the Regional 
Council. 
"We are paying the rates to have them do it [possum control] " 
The degree of on-farm possum control conducted by the farmers themselves varied considerably. 
Ten farmers used to do plenty of possum control themselves or had trappers in to do it, but their 
efforts had stopped over the years. Only nine farmers had strategic possum-control programmes in 
place at the time of the interviews. Their reasons were grouped under 'awareness',  comprising 
economical and preventive reasons. Eighteen farmers only controlled possums sporadically, either 
by setting a few traps or going shooting occasionally. The reasons for their lack of control or for 
ceasing their control efforts were mostly grouped under the category 'convenience' .  Ten farmers 
stated the good success of the Regional Council control, others mentioned very low possum 
numbers, not enough time to do control, that it was too difficult, that they would/could not do it 
over the whole farm, that they could not keep up their motivation if they did not catch any 
possums, or that they stopped after they knew that the Regional Council will come on to the farm 
the following year. 
"We didn 't trap much any more after the Regional Council was on the place, as we felt the 
possums were under control. " 
"I 've had traps and bait stations out there and haven 't caught anything in the traps for 
ages. One bait station we brought home just the other day and it still had bait in it. So 
there aren 't many possums out there. " 
One farmer stated that he stopped because he got frustrated by having conducted pos urn control 
and still had TB in his herd. 
Eleven farmers pointed out that individual farmer efforts could not be effective, but co-ordinated 
schemes were necessary, as possums did not recognise farm boundaries. The fact that farmers were 
not in full control of the situation (wildlife moving around) and the threat that TB posed to the 
international market and therefore to the whole 'economy were both mentioned as reasons for the 
involvement of the Government. Three farmers explicitly mentioned that the scheme as it was in 
71  
place at the time of  the interviews, should remain, and that compensation should not be  lowered, as 
economic margins of cattle farming are already low. 
For seven farmers economics played a major role in deciding whether to conduct any possum 
control themselves. With decreasing profit margins in cattle farming it would become less and less 
economical to spend money on possum control.  
"Current cattle prices must be a concern given that, as the returns drops, then so must the 
concern and the concerted effort to control TB. " 
One farmer questioned if the amount paid to the Regional Council could be spent more 
economically by doing the control themselves on the farm. Another suggested that possum control 
could be more economically feasible if farmers were supplied with free poison by the Regional 
Council. 
A vailable methods of possum and tuberculosis control 
Twenty-five farmers stated that control methods available to them were shooting, trapping or 
poisoning of possums, but mentioned low motivation and no time as main constraints. 
"We could put bait stations up, but there is always another job I have to do. If I am not 
looking after the stock I am doing something else that needs doing. " 
In addition six farmers mentioned regular cattle testing and careful stock purchases as a way of 
controlling TB in their herds. Grazing management was only mentioned by three farmers. One 
farmer wished for a combined programme with his neighbours by employing a trapper or using big 
mobile bait stations. Another farmer expressed the wish for training on the use of poisons. 
Wildlife observed by farmers 
Twelve of the beef farmers thought that they had a high number of possums on their farm, whereas 
all dairy farmers and four beef farmers stated that they saw few or no possums on their farms. 
Farmers were asked if they knew of any tuberculous wildlife found on their farm, or if they had 
observed sick possums during daylight hours. Fourteen farmers had not observed any sick possums 
and did not know of any infected wildlife. Nine farmers remembered that they had seen sick 
possums during daylight, some several years ago, some in their suspected TB hot-spot areas, but 
also in some cases on previously managed properties in areas where hardly any TB existed in 
cattle. Most of these farmers did not make the connection between a sick possum and the risk of 
tuberculosis transmission or hot-spot areas. Tuberculous possums, ferrets, cats, deer and pigs had 
been found previously on six of the farms. 
72 
Perceptions about tuberculosis testing and control in the herd 
Farmers' concern about the TB control programme as it was in place at the time of the interviews 
included that possum control would never be adequate as there were too many places to control, 
that possum control had to be a continuous effort, and that farmers had started to live with it and 
did not expect to be off Movement Control for any length of time. 
The less than 100% sensitivity of the intradermal tuberculin test was mentioned by eight farmers, 
one farmer even expressing the view that false positive animals would be acceptable, but false 
negative would not, as they would allow cattle-to-cattle spread. However, five farmers were 
apparently unaware of the possibility of false-negative animals in the test, as they believed that 
buying 'white-tagged' cattle (cattle from infected farms, tagged with an official white tag for 
identification) would not represent any risk of buying infected animals. 
"If you were buying cattle that were infected they would show up on the test. " 
The specificity of the skin test was only mentioned by one farmer, who believed that one of his 
animals that reacted positively to the skin-test, but did not show any lesions at slaughter ( 'NVL' for 
'no-visible lesion' animal), was actually infected with M.avium rather than M.bovis. Another 
farmer had two animals reacting positively to the skin test, which subsequently reacted to M.avium 
in the comparative skin test. The farmer believed the animals got infected from wild birds when 
they were grazing by a lake. 
Cost of tuberculosis as perceived by farmers 
The cost of TB control was mainly associated with the extra handling and management required for 
TB testing, and removing reactors from the herd at times when culling would not be normally done. 
Some financial loss was also due to compensation not being 100%, to having to sell young animals 
with official white-tags at a discount, and through lost production in dairy cows. Costs due to 
possum control, such as Regional Council rates and poison used on the farm were also mentioned. 
Only five farmers stated that TB had a considerable impact on farm income, two of whom had stud 
cattle. Eighteen farmers considered the cost of TB as not major, rather as an inconvenience. 
"[TB J hasn 't had a major financial impact for us, even though we had a high number of 
TB animals la.!.:t year, only 6-8 were condemned and we received full payment on the 
others. I guess we lost the P; .:Jrluction from the 18 heifers. " 
"[The cost of TB is J not that much, it's just m��,;; of f'ln inconvenience. " 
73 
For the two farmers with stud cattle in the sample the costs of losing individual animals was 
considerable, but the loss of the market was financially more detrimental. 
" We wouldn 't want to live with TB because of economic reasons. It has major financial 
impact on us, being a stud. We even have to consider whether to carry on with the stud. 
Once we regain accredited status, it will be difficult to get old clients back as they tend to 
be very loyal and will be committed to a new supplier. That means for us, we will have to 
find new markets. There will also be a tarnishing affect for a while, and the concern that it 
may come back again, which may put potential customers off. It will take a long time to 
build back up again . . .  We calculated a $40,000 loss in one year due to TB, that is the 
difference between the potential value and the value we get in the works [slaughterhouse}" 
If stock was going to slaughter anyway, the effect of TB was not considered important. 
"Other stock were going to the works anyway, so didn 't have a major effect on returns. " 
In order to minimise the cost of TB ten farmers changed or considered changing from selling young 
stock live to selling to slaughter only after finishing. However, only two of these farmers changed 
mainly because of TB, the others stated TB being only part of the reason and general farm 
economics, and farm management being the other part. One farmer changed his enterprise type 
from deer to cattle, as he believed farming cattle would be associated with less risk of contracting 
TB than farming deer. 
Apart from these direct costs of TB seven farmers mentioned that TB limits the general farm 
management options, such as not being able to use certain off-farm grazing options, not being able 
to sell their animals to anywhere they choose, or easily move to other properties in the case of 
being a share-milker' . 
Sources of information 
Several sources of information regarding tuberculosis, TB control and possum control were 
brought up in the interviews. Most commonly AgriQuality veterinarians and livestock officers 
(during TB testing of cattle), Regional Council workers, local veterinary practitioners, discussion 
groups and media were mentioned. Especially Regional Council staff were often seen as a 
particularly valuable source of information. Six farmers expressed the view that they benefited very 
much from meetings organised by the local RAHC (Regional Animal Health Committee) or other 
organisations with invited speakers from research groups. Twenty-three farmers mentioned a range 
1 A share-milker owns or manages a herd of milking cows in partnership with the owner of the farm on an 
agreed profit-share basis. 
74 
of sources, partly actively seeking information. Five farmers regarded themselves as 
knowledgeable enough, without the need to seek new information. 
The perception of the knowledge of Livestock officers (LOs) varied. Three farmers found them 
very helpful and knowledgeable, but four farmers were concerned about the LOs' lack of technical, 
factual knowledge. They found that LOs would only express their own beliefs, but did not have 
new information at hand. The farmers who commented in more detail about the LOs considered 
that they had great potential in providing information in that they visit many farmers and can talk to 
them while testing cattle. The farmers thought the LOs could be used more in disseminating 
knowledge and to help farmers help themselves. 
"LOs see more farmers than anyone else and provide a great opportunity to provide 
farmers with info on TB. They could at least have a brochure, which they dropped off. 
Farmers may be inclined to read this, if talking to other farmers who also get the 
brochure. " 
Willingness to accept Tuberculosis 
Sixteen of the 27 farmers would not want to live with tuberculosis permanently, for economic 
reasons (both for the herd and the international market), health reasons and having flexible 
management options. However, eight farmers stated that they have lived with it for long periods, 
got used to it, and that they could not see tuberculosis eradicated from their farms, especially under 
the current economic situation. 
"We have lived with TB for 50 years now. Would love to be free of TB given the choice, but 
there isn 't any. " 
Three farmers mentioned that the economics of the farming situation had to be considered, losing a 
few animals to TB might still make it more economical to graze hill country, rather than avoid TB 
hot-spots and staying clear of TB. 
"TB is not something I would live with on the farm but it does depend on how much it is 
going to cost to get rid of it against the cost of it at low levels -for example the reduced 
price of sale per head. " 
Only one beef finishing farmer was content to deliberately accept TB in his herd. As all of his 
cattle were going to slaughter he made it a farm policy to buy only 'white-tagged' animals, as these 
were cheaper. 
75 
General farm management 
In general farmers did not regard off-farm grazing, grazing other owners' cattle or their purchasing 
behaviour as risky practices for getting their cattle infected with TB. Farmers who considered it 
risky had changed their practices where possible, and other farmers would change if the risk in the 
individual circumstances was considered high and if they considered it feasible. 
Purchasing behaviour 
Fourteen farmers reared their own offspring and bred or bought in their breeding bulls. Thirteen 
farmers bought in 'white-tagged' cattle, mainly because they were cheaper and farmers did not 
regard them as high-risk animals, as they had been tested clear with the skin test. All except one 
farmer stated that they did not think this purchasing behaviour had influenced the TB situation on 
the farm. 
Seven farmers did not consider TB at all when purchasing cattle. Some changed this attitude once 
they had TB in their herd. 
Selling behaviour 
Sixteen farmers said that TB had no influence on their selling policies, whereas seven farmers (5 
BB, 2 DH) indicated that they had changed their selling or farm practices, by reducing the number 
of cows, moving from selling weaners towards finishing their stock, or by sending stock to 
slaughter rather than to live markets. Reasons for these changes were of moral (not wanting to 
spread TB) and economic nature ( ,white-tagged' animals yielded lower prices, and only had 
limited markets). 
"The last few years we've reared bull calves which go to first sale in November. Not doing 
it this year. Having white tags in the calves has reduced the price received by $30-$40. " 
Nine of the farms where TB had no influence on selling policies were already selling their cattle 
directly to slaughter, while the others kept their regular selling policies and accepted lower prices 
for their cattle. 
Pasture shonage as a limiting factor in grazing management 
Two main periods were mentioned for pasture shortage; winter due to lack of pasture growth and 
summer due to dry weather. Pasture shortage depended on the area the farm was in, lasted typically 
for two months, ranging from June to October for winter and December to March for summer. 
76 
Management of pasture shortage involved feeding supplements (hay, silage or feed crops), using 
less productive paddocks (bush areas) for cattle, grazing stock off-farm and/or selling stock. 
Grazing other owners ' cattle 
Seven farmers in the sample grazed other owner' s cattle on their property. They either grazed cattle 
from properties related to their own farm (share-farming or relatives; n=5), or only did this 
sporadically (n=2). None of these farmers considered this practice as high risk for contracting TB 
infection . 
Grazing management 
General grazing management 
Grazing management on beef breeding farms was flexible, based on tradition and often on past 
personal experience. It was changed on demand to meet immediate needs. The only thing fixed on 
most farms were certain calving paddocks, which were used year after year for calving. These 
paddocks mostly were less productive and had vegetation that provided cover during the winter 
months. Thereafter cows were used to clean up rough patches in paddocks all over the farm and 
were put into less productive areas, whereas the young fattening stock were kept on the better 
pasture. Most beef breeding farmers described their grazing management as being responsive to 
feed requirements of cattle and feed availability, which meant that grazing management could not 
be fixed, but cattle had to be put in any areas where feed was available. 
"Can 't have too fixed a pattern, we have to play it by ear a bit therefore. It 's that sort of a 
fann - often short of feed. " 
For dairy farms the grazing was much more fixed, with set routines of grazing rotations, using run­
off areas at certain times and for certain stock groups. These farms often used the same system for 
many years. Young stock (after weaning) were mostly grazed on a run-off or other off-farm grazing 
arrangements, and only the cows were grazed on the home farm during the milking period. During 
the winter months, when cows were dried off, they too were shifted to the run-off in order to avoid 
damage to the high-producing paddocks on the main farm. 
All except two of the 27 farmers tried to keep their cattle throughout the year in their respective 
mobs (e.g. based on age and sex). Only three of the beef farmers had their mobs in separate parts of 
the farm, on all other beef farms all animals would graze over most of the farm within one year. 
Three farmers stated that their large number of stock classes and cattle mobs made grazing 
management complex and difficult. 
77 
Bush grazing 
Bush areas were mostly used by beef farms. Of the 1 3  beef farms that had bush on their farm, ten 
used these areas at least for some of the year. In contrast, only four of the dairy farms had bush on 
their property. On two farms these bush areas were fenced off and excluded from grazing and on 
one farm these areas were only used in exceptional circumstances. 
Most beef breeding farmers used the bush areas for calving their cows (cover, plenty of space, 
drier), wintering their cattle (on the less productive pasture, in order to save the better, more 
productive pasture) or as a last resort for times of pasture shortage. 
"Generally we don 't graze in bush. We have tried all sorts of things out there and we 
finally got down to the fact that we need to feed the cows. So this year we put them in [the 
bush area] and they did very nicely. " 
However, two beef farmers were very conscious of the TB risk and did not even put cattle close to 
the bush areas. 
Three farmers ensured that the time animals spent in the bush areas was short and/or stocking 
density was low, and therefore the animals were not pushed hard for feed when cattle were put into 
the bush paddocks. Four farmers used their bush area without taking TB into consideration, 
because they believed that the vegetation was not conducive to possums, or because they 
considered the TB risk economically less costly than feeding supplements. 
"In the wintertime they might as well eat that [grass in bush a rea] rather than feed them 
ha " y. 
Several farms had patches of bush all over the farm, and therefore they did not consider it possible 
to exclude cattle from these areas. 
Hot-spot grazing 
Of the 19  farmers who had knowledge or suspicion of a TB hot-spot area on their farm, four 
farmers fenced these areas off and seven farmers tried to consider it when setting up their grazing 
management. Of the eight farmers that did not take their potential hot-spot areas into consideration, 
three explained this by not being able to exactly identify such areas, and two farmers by economic 
reasons, where they did not see any other option without huge expense such as grazing cattle off­
farm or feeding supplements. These farmers considered the consequences of TB infection in their 
cattle herds as being low. 
78 
"I was aware of the TB risk at this stage, but it was more economical to graze off and run 
the risk of one or two reactors rather than winter the cows at home and feed them 
supplementary feed. " 
Seven farmers tried to incorporate TB hot-spots into their grazing management, by excluding 
breeding cattle from these areas and only grazing them with younger fattening stock, or by grazing 
cattle in these areas only after possum control had been conducted. However, two farmers stated 
that the options for the grazing management were limited. 
"Tried to take this [TB hot-spot] into consideration in the grazing programme, but options 
were limited, particularly as there wasn 't as much cleared land on the farm at that time. 
We didn 't put cows there, but used to put young stock on the hills and got a lot of TB in this 
group. " 
Off-farm grazing 
Sixteen farmers (seven beef farmers, nine dairy farmers) used to graze some of their cattle off­
farm, either on their own run-offs (n=1 1) ,  on share-farms (n=4), or on other unrelated farms (n=2) . 
Beef breeding farms mostly used off-farm grazing to finish their cattle before selling to slaughter, 
whereas dairy farms used these areas all year round for young stock and during winter for cows. 
Unless beef farmers were linked into a share-farming business, they used less off-farm grazing than 
dairy farms (n=3 vs. n=9), and often it was only in exceptional circumstances (e.g. drought). 
Farmers, who had to find off-farm grazing opportunities for their cattle each year experienced more 
constraints because of the resistance to 'white-tagged' animals, than farmers having their own run­
off or being associated with other farms. These farmers paying for grazing opportunities found it 
difficult to find other properties that were willing to take 'white-tagged' cattle on for grazing. 
Consequently many of these farmers bought a property for use as a run-off, in order to avoid this 
dependence. 
"We went to one place [for grazing] 2-3 years in a row. But [it] has been getting increasingly 
difficult finding grazing if on MC. Generally we took whatever was going. " 
"We have always been careful with choosing where the heifers go grazing - avoid any farm 
currently on MC. But now it's getting more difficult as graziers are becoming more cautious 
about taking cattle that are on MC. In the future we will have to go to a place on MC, which I 
am reluctant to do, or change policy and winter cattle on the home farm and the newly 
acquired run-off. " 
79 
Discussion 
Descriptive analysis of study farms 
Comparing the three farm types studied (BB, BD and DH) a number of differences were found. 
The ranges of the cumulative incidence of lesioned tuberculous cattle between the three farm types 
were overlapping, but a trend was still apparent. Beef breeding farms were infected on average for 
more than 4.5 years, whereas dairy farms were infected on average less than four out of the five 
years under investigation. The beef breeding farms had a higher incidence of tuberculosis, with 
cumulative incidences of up to 0.05 . One beef breeding farm was extreme with a cumulative 
incidence of 0 .17 .  However, this farm was in a different location to the other farms (Wellington 
region) and had one of the lowest percentage of effective land (only 50% of the farm area was used 
for grazing; half the farm was in bush, gorse and trees and unsuitable for grazing). Additionally this 
farm had the lowest number of cattle, and these were allowed to graze over large areas. It might be 
suspected that due to the large area of un-used land the possum population was high, creating 
ample opportunities for direct contact between possums and cattle. This contact was reported to be 
a strong potential pathway for transmitting the disease (Sauter and Morris, 1995a). 
The beef finishing farms in the present study had an intermediate cumulative TB incidence. A 1985 
to 1990 study of TB in the Wairarapa showed beef finishing farms had a higher incidence of 
tuberculosis than dairy or beef breeding enterprises (Zewdie, 1997). However, the sample size of 
only three beef finishing farms in the present study was very low. The dairy farms had the lowest 
range of cumulative TB incidence amongst the three farm types. 
The differences in the TB incidence rates between farm type might also be related to the vegetation 
on the farm. Dairy farms are usually intensive and do not have a lot of bush, whereas beef breeding 
farms are often high-hill country farms with plenty of bush and scrub, potential possum habitat 
(Ministry of Agriculture, 1998), a situation that was clearly the case in the present study. 
In the present study there was a statistically significant positive association between the incidence 
of TB lesioned cattle and the number of whole herd TB tests over the five-year period. It is possible 
that farmers with a high incidence in their cattle herds were more interested in controlling the 
disease and therefore were prepared to test their cattle more frequently. 
The number of whole herd TB tests and the number of years during which the cattle herds were 
infected within the five-year period in the present study differed also between the three farm types. 
Dairy farms had on average a greater number of whole herd tests than beef breeding farms. This 
could be a consequence of the different animal management. Beef breeding farms generally run a 
80 
higher number of cattle on a much larger area than dairy farms. Thus, it requires more effort to get 
all animals into the yards for TB testing. The dairy farms in contrast have to bring in their cows 
daily for milking, making TB testing much easier. However, the higher number of whole herd tests 
in dairy farms could also be related to a possible greater awareness of TB and concern about it in 
dairy farms. The low number of whole herd tests in beef finishing farms can be explained by the 
legislation in place. Until the early 1 990s beef finishing farms were not required to test their 
animals, as most animals were going to slaughter. The number of years infected within this five­
year period showed a trend that beef breeding farms were infected for longer than beef finishing 
and dairy farms, but this association was not statistically significant. 
There was no statistical association between the number of whole herd tests and the number of 
years the herd had been infected. However, the time period of five years is very short and it is 
expected that it will take some time for frequent testing to have an effect on the TB status of the 
herd. 
Methodology 
The present study focused mainly on two areas in the Wairarapa, where TB infection in the possum 
population existed and wildlife surveys were being conducted. The Wairarapa region was chosen 
because this study was part of a greater research project being carried out in this region. 
Furthermore the intervention study (see Chapter 3) to evaluate on-farm control programmes, was to 
be carried out in the same region. The conclusions drawn from this interview study regarding the 
tuberculosis situation and grazing management are directly applicable only to the study population 
in the Wairarapa, although it is likely that the principles derived from these interviews will also 
apply to other areas in the country. 
The data gathered during the interviews regarding tuberculosis and grazing management were less 
suited for purely quantitative methods, as they were individual and subjective for each farm. 
Furthermore, as one of the intentions was to gain knowledge about the management on farms 
without pre-conceived assumptions, qualitative methods were used to develop theories (Strauss and 
Corbin, 1 990) about farm and grazing management, the drivers for specific actions and principles 
of TB management used on farms. These included the relationship of grazing management to the 
probability of cattle getting infected with tuberculosis, and its potential for use in conjunction with 
other TB control measures. Using open-ended, unstructured questions allowed farmers to describe 
their concepts, TB situation, perceptions of TB, grazing management and reasons for using the 
specific grazing management, expressed in their own vocabulary. Analysing these unstructured 
interviews using the 'grounded theory' method, enabled the researcher to find similarities and 
trends in highly variable and subjective data. The software programme WinMAX98, developed for 
8 1  
qualitative data analysis and mostly used in sociology and related fields, greatly facilitated data 
coding, structuring, retrieval and development of theories. Another method for analysis would have 
been one based on 'key words' ,  where the individual interview documents are searched for key 
words. Each occurrence of a key word is then analysed for its combined occurrence with other key 
words. However, grazing management and the tuberculosis situation on each farm were very 
complex and therefore described by the individual farmers with a high variability. 'Grounded 
theory' was the preferred method for analysis. 
The number of subjects used in 'grounded theory' studies on human behaviour varies widely from 
less than ten (Conrad, 1 978) to several thousands (Ekins, 1 993). However, more important than the 
number of subjects used is the diversity. Glaser and Strauss ( 1 967), the early developers of 
'grounded theory' suggested the use of as many divergent subgroups of the population as possible, 
including individuals who were likely to challenge the developing behavioural patterns. In the 
present study nearly complete coverage in the two study areas was obtained, only one farmer was 
unwilling to participate in the study. Therefore the findings can be considered to represent the 
views of these communities. 
One limitation in the present study was that all interviews were completed by the time the analysis 
started. This arose because the study was undertaken as a component of a larger, multifaceted TB 
research programme and therefore its timing had to conform to the needs of the overall study. Thus, 
the interview procedure could not be progressively adapted in the light of results in order to make 
full use of the constant comparative method proposed to enhance the effectiveness of ' grounded 
theory' method. The constant comparative method aims to reassess the emerging theories by 
changing the interviews and asking different questions. Another characteristic of the constant 
comparative method is that the collection of events is stopped when no further new concepts are 
found, a state where 'grounded theory' is said to have reached saturation (Strauss and Corbin, 
1 990). In the present study, the number of interviews available was small in relation to the variety 
of answers found and more ideas might have emerged if more interviews were available. However, 
most of the key theories emerged very quickly and were confirmed by the majority of subsequently 
analysed interviews. Therefore it is likely that the additional new ideas would not have contributed 
significantly to the main theories. Nevertheless, for the purpose of designing on-farm TB control 
methods, which could be used widely, this was sufficient, but it has to be acknowledged that more 
ideas and new concepts could have led to additional insights or even to new TB control ideas. 
Interview content analysis 
The following key theories emerged from the study and were considered important in setting up 
on-farm TB control programmes:  
82 
• Many farmers know or have suspicions about potential TB hot-spots on their farm. Dairy farms 
have less definite knowledge of hot-spots but they also have less high-risk possum habitat than 
beef farms. 
• Farms where TB hot-spots are located within paddocks have a higher TB reactor incidence 
than farms where TB hot-spots are excluded from grazing by fencing etc. 
• Economics plays an important role in decision making by farmers about TB control, both for 
possum control and for grazing strategie . TB generally is believed not to ha e a major impact 
on farm income. 
• Farmers are supportive of TB control as a national priority to protect New Zealand' s trading 
position, but, due to economic considerations, vary in their degree of commitment to 
implementing control measures on their own farms.  
• Farmer initiated possum control is sporadic, and the motivation to do it under current 
circumstances is often lacking or wears off very quickly due to competing work pressures. 
• Grazing management on beef farms is flexible, whereas dairy farms have more set routines. 
Pasture shortage is a limiting factor in setting up grazing regimens. Increased number of stock 
classes reduces flexibility of grazing management. 
• Practices such as grazing off-farm, grazing another owner' s cattle on the home farm and 
purchasing cattle are not regarded as high risk by farmers; if such practices were regarded as 
risky in the past, then farmers changed them if possible. 
• AgriQuality veterinarians and people who visit the farm on animal health business (Livestock 
officers, Regional Council staff and veterinary practitioners) are the most common source of 
information. They were seen as having great potential for being an optimal medium for 
knowledge transfer. Some farmers found Livestock officers very knowledgeable, while they 
did not meet the requirements of other farmers. 
In the following, these key theories and related points are discussed in more detail. 
Farmers generally had a good idea about tuberculosis and their specific situation. As dominance 
was found to be positively related to age and weight (Dickson et aI., 1967), the comment of some 
farmers that the reactor animals were mostly the highest producing, biggest animals in the group, is 
in agreement with dominance studies performed with cattle/deer and possums (Sauter and Morris, 
1 995b; Sauter, 1996). Sauter and Morris ( 1 995b) found that animals high in the dominance order 
83 
were the ones investigating sedated possums, which simulated the behaviour of terminally ill 
tuberculous possums. 
Amongst the farmers in the present study there was variation in the age at which cattle were 
affected by tuberculosis, some farmers mentioned that TB was occurring in all age groups, whereas 
others reported that older animals were more affected. However, this would depend very much on 
where each group of animals was grazed and from the small sample size in this study no 
generalisations can be made. 
With regard to potential TB hot-spots, most farmers (more beef farmers than dairy farmers) had 
some suspicions or knowledge that certain areas of their farm might be TB hot-spots. However, 
most dairy farmers did not have definite knowledge of where their cattle picked up TB .  This could 
be related to characteristics of dairy farms, as they do not have many areas with trees or scrub, and 
often the remaining bush areas are fenced off. However, cattle on dairy farms are usually grazed in 
high numbers in small paddocks, which means if there was a TB possum in the paddock, then most 
animals would see it and have the opportunity to interact with it. The cows are also rotated around 
all paddocks as a group on a fairly short cycle, so there is little scope to discriminate where 
infection occurred. This is in contrast to the beef breeding farms, which generally graze 
extensively. Beef breeding farmers were more specific about their potential TB hot-spots. The 
vegetation mentioned by these farmers as suspected TB-areas, such as gorse, willows or bush, was 
confirmed to be of high TB risk in a later habitat study (McKenzie, 1999). 
The finding that farms with known hot-spots within paddocks experienced a higher tuberculosis 
incidence than farms without known hot-spots or with fenced off hot-spots, suggests that excluding 
cattle from hot-spots could have major implications in controlling TB . The farmers who knew 
about TB hot-spots and excluded them from grazing with cattle, either by fencing off or by not 
grazing these paddocks with cattle any more, still used the paddocks adjacent to these hot-spot 
areas. This could indicate that to reduce TB incidence, it is not necessary to keep infected possums 
out of cattle grazing areas, but that it is sufficient to keep cattle out of these hot-spots based around 
possum denning areas. Using this in conjunction with the seasonal pattern of tuberculous possum' 
deaths (see Chapter 3) this leads to potential major on-farm TB control methods: TB in livestock 
may be reduced by either fencing off hot-spots permanently or temporarily, or by excluding 
paddocks with hot-spots permanently or seasonally from grazing with cattle. One disadvantage of 
fencing off hot-spots permanently is the potential expansion of bush area, which might lead to an 
increased number of hot-spots. However, the main advantage of these methods is that they are of 
low cost and readily available to farmers, two characteristics that make them powerful tools in the 
control of the disease. 
84 
Farmer-initiated possum control was found to be only sporadic, not continuous. Several reasons 
were mentioned: low possum density, time constraints, difficulties in controlling possums, and 
paying Regional Council rates. More dairy farmers than beef farmers mentioned low numbers of 
possums as a reason for not conducting or continuing their possum control efforts. The reasons for 
the difference in possum numbers between dairy and beef breeding farms could be two-fold. 
Regional Council possum control operations started earlier in the main dairy areas and later in the 
beef areas, but also the vegetation on dairy farms is much less possum-prone than on beef farms, 
which therefore have a much larger proportion of the farm as possum habitat. 
The reasons given for the lack of possum control conducted by farmers, indicates that any further 
education and agriCUltural extension of farmers must be targeted towards this field. If farmers are 
expected to assist in the control of possums, their motivation must be raised. Motivation has to be 
kept up and reinforced once possum numbers become low. Several farmers in the present study 
stated that motivation had suffered, and they had actually stopped their possum control efforts as 
soon as it became apparent that the Regional Council would be undertaking control work on the 
farm. Others felt that they did not have any responsibility for controlling possums themselves, 
because they paid possum control rates to the Regional Council. Some farmers also mentioned that 
individual farm control efforts could not be successful as possums range across farm boundaries, 
therefore necessitating co-ordinated efforts. This may often be a 'convenient' way to justify 
avoidance of possum control, by assuming that the neighbours would not do their part of the 
control and therefore hampering the farmer' s own possum control efforts. 
Several farmers also mentioned economics as a reason for their lack of TB control. They regarded 
it as uneconomic to spend money on TB control (equipment, material and labour) under the current 
low profitability of their farms. In general, farmers did not regard tuberculosis as of major financial 
importance, unless they were breeding stud cattle.  In that case the disease had a major financial 
impact, as animals had to be sold to slaughter at below their monetary breeding value, as the 
breeders lost their stud cattle markets. Most other farmers reported only a small loss due to 
compensation for reactor animals not being 100% or having to sell young stock with white-tags. 
Most farmers regarded TB and its control mainly as an inconvenience due to the extra animal 
handling required and/or due to it limiting overall farm management. If farmers sent the maj ority of 
their stock to slaughter anyway, TB did not have severe financial consequences. This was also the 
reason for some of the farmers to change their stock policies towards finishing all of their stock and 
sending prime cattle to slaughter, rather than selling young stock. By moving to slaughter-only, the 
cost of TB could be minimised. Thus, if farmers do not associate a significant cost with TB ,  they 
are unlikely to see any need for controlling possums and TB on their farms. As more and more 
farmers move to slaughter-only, cost-effective TB control programmes are required. Therefore, 
85 
further investigation needs to consider both the costs of TB control programmes and the cost of the 
disease to the farm. The cost analysis could also investigate changes in the regulation that increase 
the perceived cost of TB to farmers, such as removal of compensation or direct payment for TB 
testing. 
All these reasons given for the lack of farmer-initiated TB control and for ceasing control efforts by 
farmers indicate that further investigation should focus on motivational incentives that raise the 
willingness and motivation to carry out TB control on farm, on incentives that keep and reinforce 
this motivation and that apply to most farmers, in order to achieve widespread adoption of control 
programmes. If motivational incentives could be put in place that resulted in widespread adoption 
of TB control programmes, it might be hypothesised that peer pressure amongst farmers might 
even lead to complete coverage over time. As economic considerations are a major driver for 
farmers it is likely that these incentives will have some financial influence, such as decreased 
Regional Council rates, subsidised poison and bait stations or an insurance scheme for TB (see 
Chapter 6). 
Apart from the ·field of motivation of farmers, education and extension should also focus on TB 
testing related issues. Several farmers mentioned issues, such as 'anergic' animals, the wish for a 
better test, their belief that 'white-tagged' cattle are 'un-infected' or that cattle become infected by 
birds. Such 'anergic' animals with advanced stages of disease, infected cattle that did not react to 
the intradermal tuberculin test, but showed lesions (sometimes extensive) at slaughter, are 
considered to be seldom now with frequent testing, as animals are removed before extensive 
lesions can develop. It was notable that farmers who had anergic animals in the past were much 
more aware and put a much higher risk assessment on cattle-to-cattle spread of TB than farmers 
who had never had an anergic animal. 
The less than 100% sensitivity of the intradermal tuberculin test was also mentioned by several 
farmers as a constraint to successful TB control. Investigations from Australia and New Zealand 
showed that the sensitivity of the caudal fold test was between 65.6 % and 95.6% (Lepper et ai. , 
1 977; Francis et aI. , 1 978; Lepper et al. , 1 979; Wood et al. , 1 99 1 ;  Wood et aI. , 1 992; Ryan, 1 992). 
In the Wairarapa MAF veterinarians use a ' working' figure of 85% ± 5% (G.Pannett, pers. comm. 
1 999). This would indicate that 1 0-20% of infected animals are not identified by the intradermal 
test. In an analytical model of Kean and Barlow ( 1999) it was found that this Iow sensitivity was of 
little consequence within the herd TB testing programmes. The desire of farmers for a better or 
perfect test has to be met with education programmes. This also applies to the fact that many 
farmers felt that intradermal tested animals from infected herds ( 'white-tagged' animals) were 
identified as 'un-infected' .  
86 
Any education and extension of farmers regarding motivation or TB testing issues could be built on 
existing knowledge transfer methods. The farmers in the study were aware of a wide range of 
sources of information on TB. However, staff from AgriQuality and the Regional Council, coming 
directly on to the farm, were mostly mentioned as the fIrst and most important contact, especially 
the livestock offIcers (LOs). The impression farmers got from these LOs differed very much 
amongst farmers, some fInding them very knowledgeable, others fInding them too subjective and 
not able to provide information which met the requirements of farmers. These LOs were seen by 
several farmers as an ideal medium of knowledge transfer, as they were in regular contact with the 
farmers and often spent several hours on the farm. With adequate training these LOs could educate 
farmers on issues such as test sensitivity or on-farm control programmes. Veterinary practitioners 
were also mentioned as a source of information, although not by the majority of farmers, indicating 
an opportunity for veterinary practitioners to expand their role. 
From a national perspective, with the Pest management strategy aim of controlling and eradicating 
TB, it is encouraging that most farmers would not want to live with TB in their cattle herd 
permanently. However, the substantial number of farmers who stated that they had got used to it, 
that they have lived with it for a long time, and that they cannot see TB being eradicated, indicates 
a strong need for education and agricultural extension. 
Economic considerations were also of major importance in general farm management and grazing 
management. Half of the farmers in the study bought in 'white-tagged' animals, mainly because 
they were cheaper. These farmers stated that this purchasing behaviour had not affected the TB 
situation on their farms. Only one farmer believed that TB was introduced by infected cattle rather 
than having a TB problem on the farm itself. The minor role of cattle movement in creating new 
TB breakdowns was also reported by Carter et al. ( 1 995). Nevertheless these authors stated that the 
risk could be minimised by adopting a policy of preventing movements from infected herds when 
reactors were found at the pre movement TB test, a change that has been implemented since June 
1 997. Farmers also did not consider their off-farm grazing or grazing other owner's cattle as a 
high-risk practice. They stated that if they had the feeling that these practices caused TB infection 
in their animals, then they would change them, as long as it was economic to do so. However, to 
change these practices is often only possible for farmers who have to fInd new grazing 
opportunities for their cattle each year. If they had purchased extra land, such as a run-off, it is 
more diffIcult and associated with costs to change this. Therefore it is more likely that farmers 
would continue to use these areas despite the TB risk to their livestock. 
One of the aims of the present study was to develop hypotheses for on-farm TB control 
programmes. Resulting from the key theories of TB hot-spots, the approach of excluding hot-spot 
areas from grazing by cattle can be developed. To test if this method is practicable, it is necessary 
87 
to know if the grazing management on farms does allow such an alteration, and incorporation of 
TB control methods into standard farm management. On beef farms grazing management was 
found to be flexible, whereas on dairy farms it was more fixed. Beef farmers stated that cattle were 
put in paddocks ad hoc where grass was available, whereas dairy farmers had set routines, with 
young stock being mostly off farm grazing. Some of the farmers in the study, who tried to take TB 
hot-spots into consideration, mentioned that only limited options were available to them. This was 
especially the case in times of pasture shortage, when farmers were more likely to ignore their 
concerns about possible TB infection from possums, as their main aim was to feed their cattle. This 
also indicates that economic considerations were one of the most important factors. If it is more 
economic to risk TB infection by grazing cattle in hot-spot areas rather than purchasing 
supplements, it is likely that not many farmers will consider TB during these times. 
It is therefore proposed that on beef farms, there is potential for excluding TB hot-spots from 
grazing with cattle for certain high risk periods. For dairy farms this might be more difficult. In 
addition beef farms generally also have sheep, which could use these hot-spot areas. Dairy farms 
do not have any other species but could use the grass for supplements. However, first it has to be 
evaluated whether such a control method of excluding cattle from TB hot-spots during high risk 
periods is effective or not (see Chapter 3).  Secondly the costs of such a control programme and the 
cost benefits of TB control to the farm have to be established, which might then lead to 
motivational incentives that make it more economical for farmers to control possums and TB (see 
Chapter 6). Because beef farms generally present more of a TB control problem than dairy farms, 
individual farm strategies may well be more effective and necessary on such farms, and this needs 
to be examined. 
Bibl iography 
Adarns, c.L., Bonnet, B.N. and Meek, A.H. ( 1 999). Owner response to companion animal death: 
development of a theory and practical implications. Canadian Veterinary Journal, 40 ( 1 ) :  33-
39. 
Animal Health Board ( 1 995). National TB Strategy: proposed pest management strategy for bovine 
tuberculosis. Animal Health Board, Wellington, New Zealand. 
Animal Health Board ( 1 998). Five Year National Bovine Tuberculosis Pest Management Strategy 
to 30 June 2001.  July 1998 to June 2001 .  Animal Health Board, Wellington. 
8 8  
Boland, c.J. and Morris, R.S.  ( 1 988). Grounded Theory - Collection and Analysis of Qualitative 
Data in a preliminary research investigation. Acta Veterinaria Scandinavica - Supplement, 84: 
493-495. 
Carter, C., Crews, K., Hoyle, F. and Ryan, T. ( 1 995). The movement of cattle from herds in which 
Mycobacterium bovis has been diagnosed. In: Griffin F, de Lisle G (Tuberculosis in Wildlife 
and Domestic Animals - Otago Conference Series No. 3 .  University of Otago Press, Dunedin, 
pp. 298-9. 
Conrad, C.F. ( 1 978). A grounded theory of academic change. Sociology of Education, 5 1 :  1 0 1 - 1 1 2 .  
Dickson, E.P., Barr, G.R. and Wieckert, D.A. ( 1 967). Social relationships of dairy cows in  a feed 
lot. Behaviour, 29: 1 95-203. 
Ekins, R. ( 1 993). On male Femaling: A grounded theory approach to cross-dressing and sex 
changing. Sociological Review, 1 ( 1 ) :  1 -29. 
Fleming, P. E. (1996). Farm Technical Manual. Department of Farm and Horticultural 
Management, Lincoln University, Canterbury, New Zealand. 
Francis, J., Seiler, RJ., Wilkie, LW., O'Boyle, D., Lumsden, MJ. and Frost, AJ. ( 1 978). The 
sensitivity and specificity of various tuberculin tests using bovine PPD and other tuberculins. 
Veterinary Record, 103 ( 19): 420-425 .  
Glaser, B.  G. and Strauss, A.L. (1967). The discovery of Grounded Theory: Strategies for 
Qualitative Research. Aldine de Gruyter, New York. 
Hathaway, S .c., Ryan, T. J., Johnstone, A. C. and de Lisle, G. W. ( 199 1 ). The role of 
slaughterhouse inspection in the diagnosis and control of tuberculosis. In: Dr. R. Jackson 
(Convenor), Symposium on Tuberculosis. Foundation for Continuing Education of the New 
Zealand Veterinary Association, Palmerston North, pp. 227-34. 
Kean, J.M. and Barlow, N.D. ( 1 999). Evaluating potential sources of bovine tuberculosis infection 
in a New Zealand cattle herd. New Zealand Journal of Agricultural Research, 42 ( 1 ): 1 0 1 - 1 06 .  
89 
Keamey M.H., Murphy S . ,  Rosenbaum M. ( 1995). Chapter 18 :  Mothering on crack cocaine: a 
grounded theory analysis. Glaser B .G. Editor. Grounded Theory 1984 - 1994. Volume One. 
Mill Valley, CA, USA Sociology Press, pp. :257-279. 
Lepper, A.W.D.,  Carpenter, M.T., Williams, 0.1. et al. ( 1979). Comparison of the efficiency of two 
doses of bovine PPD tuberculin in single caudal fold tests on Australian cattle. Australian 
Veterinary Journal, 55  (6): 25 1-256. 
Lepper, A.W.D., Newton-Tabrett, D.A., Corner, L.A., Carpenter, M.T.S.W.A., Williams, 0.1. and 
Helwig, D.M. ( 1977). The use of bovine PPD tuberculin in the single caudal fold test to detect 
tuberculosis in beef cattle. Australian Veterinary Journal, 53  (5): 208-213 .  
McKenzie, J. S.  (1999). The use of habitat analysis in  the control of wildlife tuberculosis in  New 
Zealand. Unpublished PhD thesis, Massey University, Palmerston North, New Zealand. 
Ministry of Agriculture ( 1998). Farm Monitoring Report - North Central Region. MAP Policy, 
Ministry of Agriculture, Palmerston North, New Zealand. 
Morris, R.S. and Pfeiffer, D.U. ( 1995). Directions and issues in bovine tuberculosis epidemiology 
and control in New Zealand. New Zealand Veterinary Journal, 43 (7): 256-265. 
Phillips, L. R. and Rempusheski, V. F. (1995). Making Decisions about Elder Abuse. In: Glaser, B.  
G. (Ed.), Grounded Theory 1984 - 1994. Sociology Press, Mill Valley, CA, USA, pp. 405-21 .  
Richards L. ( 1997). Computers and Qualitative Analysis. In: Keeves J.P. (editor) Educational 
Research, Methodology, and Measurement: An International Handbook. second edition, New 
York Pergamon, pp. :286-290. 
Richards T.J., Richards L. ( 1994). Chapter 28: Using Computers in Qualitative Research. In: 
Denzin N.K., Lincoln Y.S. (eds.) .  Handbook of Qualitative Research. Thousand Oaks Sage 
Publications, pp. :445-462. 
Ryan, T. ( 1992). AHB Project Code: 57/90. Ministry of Agriculture and fisheries, MAP QUality 
Management, Ruakura, Hamilton. 
90 
Sauter, C. M. ( 1 996). Epidemiological investigations of risk factors in the behaviour of livestock 
for transmission of tuberculosis from possums and ferrets in New Zealand. Unpublished 
Thesis, Freie Universitat Berlin, Berlin. 
Sauter, C.M. and Morris, R.S. (1995a). Behavioural studies on the potential for direct transmission 
of tuberculosis to farmed livestock from feral ferrets (Mustela furo) and possums (Trichosurus 
vulpecula). New Zealand Veterinary Journal, 43 (7): 294-300. 
Sauter, C .M. and Morris, R.S. ( 1995b). Dominance hierarchies in cattle and red deer (Cervus 
elaphus): Their possible relationship to the transmission of bovine tuberculosis. New Zealand 
Veterinary Journal, 43 (7): 301 -305. 
Strauss, A. and Corbin, 1. (1990). Basics of Qualitative Research. Grounded Theory Procedures 
and Techniques. Sage Publications, Inc., 1990. 
Strauss, A. L. ( 1987). Qualitative Analysis for Social Scientists. Cambridge University Press, New 
York. 
Wood, P.R., Corner, L.A., Rothel, J.S. et al. ( 199 1) .  Field comparison of the interferon-gamma 
assay and the intradennal tuberculin test for the diagnosis of bovine tuberculosis. Australian 
Veterinary Journal, 68 (9): 286-290. 
Wood, P.R., Corner, L.A., Rothel, J.S. et al. (1992). A field evaluation of serological and cellular 
diagnostic tests for bovine tuberculosis. Veterinary Microbiology, 3 1  ( 1 ) :  7 1 -79. 
Zewdie, S. (1997). Epidemiological analysis of tuberculosis in cattle herds. Massey University, 
Palmerston North, New Zealand. 
9 1  
Deer interacting with simulating moribund tuberculous possum 
Typical 'interaction' between sheep and simulated moribund tuberculous possum 
92 
Chapter 3 
Effectiveness of on-farm tuberculosis control 
programmes: Farms located in  the Wai rarapa 
93 
94 
,,:t �\).,. 
. . ... � Jf 
95 
Abstract 
The Australian brushtail possum is the main vector for bovine tuberculosis to cattle and deer in 
New Zealand. Due to this wildlife reservoir and vector species, control of the disease is 
complicated and has to involve both livestock and wildlife. In about 23% of the country TB is 
endemic in possum populations and 3% of all cattle and deer herds are classified as infected with 
TB . Vector control operations, mainly large scale poisoning programmes conducted by aerial or 
ground baiting, are expensive and cannot cover all areas with TB in possum populations. And even 
in some areas where control was conducted, TB in livestock has not been eradicated so far. Thus, 
increased responsibility for control will be placed on farmers themselves. Therefore it was 
necessary to evaluate the effectiveness of on-farm control measures available to farmers. 
An intervention study involving 67 farms in the Wairarapa was conducted. Two main on-farm 
control measures were employed on 34 farms: targeted vector control in spring and autumn, and 
grazing management that excluded cattle from grazing TB hot-spot areas during July and August, 
and again from November to January. Vector control was conducted by the research team during 
the first two years of the project, while farmers took over sole responsibility for the continuation of 
the control measures in the third year. A higher proportion of focused control farms than standard 
control farms achieved TB clear status for their herds. A lower proportion of focused control than 
standard control farms had multiple TB-positive animals in the final project years, and the two-year 
cumulative TB incidence was reduced more in focused control than in standard control farms. Most 
results did not reach significance at the p=O.05 level, however, the small sample size meant that 
statistical power was inescapably low. Nevertheless, the study provided evidence that specifically 
targeted possum control, combined with stock management practices that minimise or prevent 
contact with infected vectors are effective methods of reducing the herd incidence of TB and the 
rate at which herds are declared TB-free. These control measures are of low cost and they should 
easily be implemented and should contribute significantly to the AHB strategy to reduce the herd 
prevalence of TB . 
Introduction 
In New Zealand, the Australian brushtail possum (Trichosurus vulpecula) is nQt o�y a pest to 
native flora and fauna, but also a reservoir host for bovine tuberculosis and a vector of the disease 
to livestock (Morris et al. , 1994). Areas where tuberculosis (TB) is endemic in possum populations 
are classified as Vector Risk Areas (VRAs). In 1 998 about 23% of New Zealand's land area was in 
96 
zones classified as VRAs (Animal Health Board, 1 998a). Thus control of the disease has to be 
twofold - control in livestock through eradication measures within herds, and control of infected 
vector populations. This has resulted in complicated and expensive control programmes (Animal 
Health Board, 2000). Since 1996, New Zealand's  TB eradication programme has been carried out 
under a National Pest Management Strategy administered by the Animal Health Board (Animal 
Health Board, 2(00). The main objective of this strategy is to reduce the number of TB infected 
cattle and deer herds to achieve internationally recognised freedom from TB within 10 years, in 
order to protect export market access (Animal Health Board, 2000). This freedom of TB is 
achieved if 99.8% of all herds have been tested negative for at least three years. At the beginning of 
2000 New Zealand has had approximately 97% of herds that have achieved this requirement 
(Coleman and Livingstone, 2000). 
Approximately 10% of New Zealand' s land area is currently under some form of continuous TB 
vector control (Coleman and Livingstone, 2000). The majority of possum control is done by 
poison-baiting, using aerial and ground application of poison baits (Morgan and Hickling, 2000). 
Since 1 987 all possum control programmes include annual or biennial follow-up programmes that 
are intended to maintain the possum population at low levels (Coleman and Livingstone, 2000). 
Most vector control is conducted by Regional Councils and financed in part through the Animal 
Health Board, which receives levy funding from the cattle and deer industries, and funding from 
the Crown, complemented by direct funding from the region through rates paid by landowners 
(Anon, 1998). In 1 995 $ 1 8  million was spent on vector control (Animal Health Board, 1995), and 
this expenditure rose to $28.4 million in 1998/99 (Coleman and Livingstone, 2000). These possum 
control operations have contributed to reducing both the number of cattle and deer herds under 
quarantine for TB control purposes and the incidence of TB (Pannett, 1995 ; Livingstone, 1997; 
Animal Health Board, 2(00). However, in many areas eradication of TB from the possum 
population has not been achieved, necessitating continuous vector control efforts to keep the 
disease in livestock at low levels (Anon, 1998; Coleman and Livingstone, 2000). Additionally, it is 
expected that the cost of maintaining the areas already controlled will remain high, and with current 
technology, will continue long into the future (Animal Health Board, 2(00). 
As the incidence of TB in cattle and deer herds within VRAs decreases, the marginal return on 
funds invested in possum control becomes smaller and the (perceived?) justification for spending 
large amounts of taxpayers' money in VRAs becomes less (Cullen and Bicknell, 2000). Therefore 
increasing pressure and responsibility will be put on individual farmers to assist in the control of 
the disease, especially on properties with persistent TB. One of the objectives of the present five­
year strategy of the Animal Health Board is 'to encourage individuals to take action against TB on 
their properties and in their herds' (Animal Health Board, 1 995). It is unlikely that farmers can 
97 
provide the same service currently offered by the Regional Council in terms of possum control 
(whole farm blanket control), but other complementary methods, assisting the Regional Council 
control strategies, could be available to farmers . 
TB is not evenly distributed within an infected possum population, but is clustered particularly in 
certain denning areas, typically covering an area with a diameter of approximately 40 meters, 
ranging from 20 to 1 20 meters (Pfeiffer, 1994; Hickling, 1995) .  These areas, commonly referred to 
as 'TB hot-spots ' ,  may persist over many years at the same location or be more sporadic and only 
persist for relatively short periods (McKenzie et aI., 1997; McKenzie, 1999). It appears that 
environmental risk factors are important in maintaining a TB hot-spot for long periods by 
influencing the transmission rate of TB (McKenzie et aI. , 1997). In a study by McKenzie ( 1999), 
multiple TB possums were more likely found in flatter land with multiple enclosed dens. 
Accumulating knowledge about the spatial epidemiology of TB in possums, suggests that 
specifically targeted farm management practices may reduce the incidence of TB in livestock. This 
should lead to cost-effective control methods, that result in more rapid eradication of TB from 
herds. The objective of the present study was to evaluate the effectiveness of practical, low-cost, 
on-farm TB control measures, which consisted of targeted vector control and management 
practices to reduce the risk of transmission from vector to livestock. This chapter presents the 
effectiveness of these practices in reducing the herd incidence of TB and achieving eradication of 
TB from the herd, as measured by revocation of TB Infected Status. 
Materials and Methods 
An intervention study was conducted with 70 farms that were under Movement Control (MC) 
restrictions at the time of selection. These were randomly allocated into equal numbers of 'focused 
control' farms (receiving specific advice and control effort in addition to standard official control) 
and 'standard control' farms (receiving standard control measures only). 
Study area 
All farms selected for the study were located in the Wairarapa (Figure 4), a region where TB in 
cattle has been endemic since at least the 1 950s (Shortridge, 198 1 ) . Traditional test and slaughter 
strategies were successfully employed until the late 1960s, when there was initial evidence that 
possums had become a reservoir for TB (Anon, 1 998). Since 1 968 tuberculous possums have been 
found in over 140 different locations in the Wairarapa. Other tuberculous wildlife such as ferrets, 
98 
stoats, feral deer and pigs have also been found in the Wairarapa during the last two decades 
(Lugton, 1 997; Anon, 1998). 
• 
• • • 
• • 
• 
, . . �l' 
•• • .J' 
./ · : r 
.. y. ' .. .. • • • Focused conlrol 1arms 
.. Standard conlrol 1arms 
Figure 4. Location of study farms in the Wairarapa in the North Island of New Zealand. 
Possum control efforts by the Regional Council 
Large scale possum control operations were conducted in Eastern and Southern Wairarapa between 
1 976 and 1 980 (Anon, 1998) . The TB incidence in cattle decreased by 80% within two years of the 
initial control operation. Follow-up control work was not conducted, unless the initial operation 
was found to be ineffective. The TB incidence in cattle started to increase and reached pre-control 
levels 8- 1 0  years after the initial control operation. During the 1980s only small scale possum 
control operations were conducted, and TB increased. From 1989 onwards more possum control 
operations were put in place, increasing the area under control from 5-7,000 hectares to 265,000 
hectares in 1997/98 (Anon, 1 998). These on-going control efforts resulted in a steady decrease in 
TB incidence in cattle and deer (Livingstone, 1997). 
Farms involved in this study (both focused control, and standard control farms) that were located 
within a vector control area received control as usually conducted by the Regional Council. Initial 
vector control programmes by the RC are termed ' Initial operation', any follow-up operations 
conducted annually or biennially are termed 'maintenance operations' .  
99 
Farm selection process 
To select the 35 focused control and 35 standard control farms it was necessary to follow several 
steps. In June 1 996 there were 2,702 cattle herds and 1 67 deer herds in the Wairarapa, of which 
9.5% and 1 1 .3%, respectively, were on MC for TB (Anon, 1 998). Farms for this project were 
selected with the help of AgriQuality staff in Masterton between March and June 1 996. In a first 
step all cattle farms on Movement Control were identified. Deer farms were excluded in order to 
have a more homogeneous group of farms. The second step involved selection of all dairy ( 'DH') 
farms and all beef fInishing ( ,beef dry stock' , 'BD ' )  farms amongst the farms identified in the first 
step. Of the beef breeding ('B B ' )  farms only a selection that was judged by AgriQuality 
veterinarians as being representative of the Wairarapa, was taken into consideration. In the third 
step, all farms that had only recently come on to MC, small farms,  and farms where the 
veterinarians in AgriQuality felt TB was introduced solely by bought-in animals, were excluded. 
Additionally, all farms in the Northwest of the Wairarapa were excluded as this area was managed 
by a different Regional Council. Parts of the area under consideration were subject to Regional 
Council possum control programmes with various commencement dates, while other parts were not 
subject to such programmes at the time the study commenced. 
This selection process left 170 farms available for the project. A letter asking farmers if they were 
prepared to participate in the study was mailed at the end of May 1 996 with a follow-up letter four 
weeks later to 40 non-responding farmers. 1 27 farmers were prepared to participate in the study, 1 1  
farmers stated reasons why they considered their farms as 'unsuitable' for the study and 32 did not 
reply. 
Of the 1 27 farms whose owners/managers were prepared to participate, 22 came off Movement 
Control between the frrst approach (May 1 996) and June 1 996. The remaining 1 05 farms consisted 
of 54 BB farms, 22 BD, 26 DH and 3 Mixed dry stock herds. Of these 105 farms, seventy were 
randomly selected for consideration as either focused control or standard control farms, using the 
random number function in MS Excel. Initially a distribution of 30 BB, 20 BD, and 20 DH, in 
equal numbers within and outside the Regional Council (RC) control area, was desired. However, 
only 30 farms ( 1 4  BB, 8 BD, and 8 DH) outside the RC control area were available. These were 
randomly assigned to the focused control and standard control groups, keeping numbers per group 
balanced. Then 40 farms (16 BB, 1 2  BD and 1 2  DH) were randomly selected from those within RC 
control area, and randomly assigned in equal numbers to the focused and standard control groups. 
The 35 farmers selected for the focused control group were contacted by phone and the nature of 
their expected commitment explained. One farmer indicated that he planned to cease farming 
cattle, one farmer was not prepared to co-operate, and six farmers considered their farm less suited 
100 
for the project, as they believed they either had bought in their tuberculous animals (rather than 
being infected from possums) in the past (n=2), or only had a small number of animals on their 
property (n= l), or had changed their off-farm grazing arrangements as a result of TB infection on 
previously used land (n=3). For each of these farms a replacement farm was selected randomly 
from the remaining farms of this herd type and control area, if available. As there were no 
replacement farms available in non-vector-control areas, a farm assigned as standard control farm 
for this combination of herd type and control area, was also removed from the list, in order to keep 
the distribution between focused control and standard control farms equal. As more BD farms had 
to be removed and no other farms of this herd type were available, more beef breeding farms were 
selected to keep the total number at 70. 
During the programme one focused control farm (BB in non-control area) proved too difficult to 
work on (location, farm management, farmer co-operation) and was dropped from the study. Two 
beef finishing farms in the standard control group were dropped out of the study due to following 
reasons: One farm (BD in non-vector control area) ceased farming cattle between the time of first 
approach and the start of the programme and a second farm (BD in control area) was on Movement 
Control at the time of selection, but its herd history was inconsistent and mainly based on 
'veterinary directions' rather than evidence of TB. No other beef fInishing farms were available for 
replacement. Therefore it was not possible to balance numbers of farms in each category (herd type 
and control area). The fmal distribution of farms available in December 1 996 is shown in Table 2. 
Information on farm characteristics, such as farm size, stock numbers etc. are presented in Chapter 
5. No significant differences were found in these variables between focused and standard control 
farms. 
Table 2. Distribution of focused and standard control farms by herd type and Regional 
Council vector control area (December 1996). 
Focused Standard 
control farms control farms 
Beef breeding in vector control area 1 0  1 0  
Beef breeding i n  non-vector control area 6 7 
Beef finishing in vector control area 5 4 
Beef finishing in non-vector control area 3 2 
Dairy herd in vector control area 8 7 
Dairy herd in non-vector control area 2 3 
Total 34 33 
10 1  
Cattle TB data 
Six BB farms (three focused control, three standard control) and one standard control BD also had 
deer on their properties. However, these herds had a clear TB status when the farms were selected, 
and they were not subject to any interventions. Therefore no deer TB testing data was collected. 
Cattle TB testing data was obtained from the National Livestock Database (NLDB), which is held 
by AgriQuality New Zealand (the equivalent of a state field veterinary service) (Ryan, 1 997). All 
records relating to TB testing on farms is recorded and managed in this database on a herd basis. 
Since 1996, the NLDB has also been linked to Agribase, a national geographically referenced farm 
database. This allows the linkage between herds and farms to be established by a unique farm 
identification number. In the Wairarapa, TB testing data held in NLDB dated back to 1979 for most 
farms. 
For each farm the NLDB also records the year in which the Regional Council conducted a possum 
control operation on this farm. Most operations are conducted between October in that year and 
June of the calendar year following the recorded date in the database (1st September). It is assumed 
that these possum control operations only start to have an effect on the TB incidence in livestock 
herds in the year following the operation. 
Confirmation of TB status 
All cattle in the Wairarapa are subject to annual TB testing, using the caudal fold skin tuberculin 
test, which is conducted by livestock officers employed by AgriQuality (Animal Health Board, 
1998b). Any cattle that reacted positively to this test was considered TB positive (Animal Health 
Board, 1 998b) unless the animal was serial tested negative thereafter with an ancillary test. If the 
animal did not get tested with an ancillary test, but went to slaughter, it was considered rB 
positive, irrespectively of whether visible lesions at slaughter were found or not. The TB status of 
cull animals, that went to slaughter without being tested first and in which lesions, indicative of 
TB, were found at slaughter, was determined according to the results of the histological 
examination. Any positive animal was termed 'TB animal' .  
A cattle herd was considered TB positive a s  long a s  it was under MC restrictions. Only i f  the whole 
herd had obtained two clear TB tests with a minimum of six months in between, the MC 
restrictions were revoked and the herd considered TB clear. The TB status of one beef dry stock 
standard control herd was being determined through 'Works monitored' , an option where no cattle 
have to be tested on the farm, provided over 90% of the cattle go to slaughter each year (Animal 
Health Board, 1998b). Herds that take up this option cannot achieve a Clear TB status, but keep 
1 02 
their Works Monitored status, unless infection is found in an animal, thus leading to an Infected 
status. 
In possums, any gross lesions suspect of tuberculosis found at post mortem examination were 
collected and cultured. If Mycobacterium bovis was cultured, the animal was considered 
tuberculous. 
Farm Visits 
The research team consisted of two persons, the author/epidemiologist and one field worker. The 
author provided the background and technical knowledge to the farmer, which was incorporated 
through the activities of the field worker into practical, individual farm control measures. There 
was a close working relationship between the two team members which ensured that uniform 
information was given by both members to farmers. 
The interventions were designed to run for three years. It was intended that the possum control be 
conducted by the research team during the first two years, whilst the farmers would take 
responsibility for control in the third study year. In the first year TB hot-spot areas were identified 
on many farms, using vegetation and slope data (McKenzie, 1 999), wildlife surveys and grazing 
records if available. Once these hot-spots were identified, possum control was specifically targeted 
towards these areas using trapping and poisoning, supplemented by occasional shooting. As only 
one staff member was available for possum control, not all farms could be controlled at the same 
time so possum control was not only conducted in spring and autumn. In the third year of the 
project, the research team ceased their control effort and encouraged farmers to keep bait stations 
running and to continue with other measures. 
To identify hot-spot areas, grazing records and cattle TB testing results were analysed, in 
conj unction with analysis of farm-specific vegetation and slope data. Results from TB hot-spot 
studies indicated that habitat influences the risk of a tuberculous possum being present in particular 
locations on farms. These locations can be predicted with acceptable sensitivity and specificity 
using satellite vegetation mapping and analysis of slope data (Hickling and Efford, 1996; 
McKenzie, 1999). This procedure was used on most participating farms. In many cases the TB 
reactors occurred in one specific age group or grazing group, therefore indicating that this group 
was exposed to one or more tuberculous possums in paddocks grazed since shortly before the last 
TB test. The paddocks used six weeks prior to the previous TB test were also taken into account, in 
order to allow for animals that became infected prior to the previous test but did not react then. Any 
suspicions which farmers put forward of likely TB hot-spot areas were also taken into 
consideration. Traps and bait stations were then set in these likely or suspected hot-spot areas. 
103 
In addition to these trapping visits, regular visits on the focused control fanns were made by the 
author, to conduct interviews and review the TB situation. During the fIrst visit to the farm in 
July/August 1 99 6  the project and planned interventions were explained and information regarding 
basic farm characteristics, TB situation and high TB risk areas collected. At the second visit the 
planned interventions were modilled and adapted to each individual farm with the help of the 
farmers and/or manager. During the third and fourth visit the questionnaire (Chapter 5 )  was 
conducted with the farmer/manager. The subsequent visits focused mainly on the review and the 
achievement of the imposed control measures .  If unexpected TB related issues occurred, farm visits 
were performed outside the regular schedule to discuss and review the situation and the possible 
control methods. The fIrst four visits were made within a period of 15 to 18 months. If herds came 
off MC early on in the project, only two visits were made within the fIrst 1 8  months, whereby 
farmers were interviewed during the second visit. In the second year no specifIc time table was set 
up for farm visits, they mostly arose out of TB testing sequences and their results on the farms. 
Phone calls were scheduled in between the farm visits in order to maintain information flow 
between research team and farmer. Their frequency varied considerably between farms, some 
farms only receiving less than fIve, others more than ten. 
During the whole duration of the project, farmers were actively encouraged by the epidemiologist 
and the fIeld worker to use grazing management to keep cattle out of TB hot-spots during winter 
and summer months. 
In May 1 997 a group meeting was arranged, where all farmers were invited to discuss the project 
up to that point and to visit the study site of the longitudinal study at Castlepoint (Pfeiffer and 
Morris, 199 1 ; Pfeiffer, 1 994) and hear about vaccination trials being conducted concurrently by 
other researchers. 
TB control measures employed in this study 
Basis of the hypothesis 
Previous research showed that tuberculous possums are not evenly distributed on a farm, but 
clustered in hot-spots (McKenzie, 1 999). The results of an interview analysis regarding farm 
management (see Chapter 2) indicated that farms with recognised TB hot-spots within paddocks 
had a higher incidence of TB in cattle grazing these areas than farms that specifIcally grazed their 
cattle away from TB hot-spots or did not have any TB hot-spots on their property. Additionally, 
behavioural studies with possums/ferrets and cattle/deer strongly indicated that TB is transmitted 
from wildlife species directly to cattle and deer grazing close by, e.g. sniffIng and licking 
terminally ill tuberculous possums (Sauter and Morris, 1995). It was also found that a single 
1 04  
intensive possum control operation in an area is  not sufficient to eradicate TB permanently from 
livestock in that area (Coleman and Livingstone, 2(00). Even after conducting vector control for 
several years, infection in the possum population can persist. The persistence of infection within 
TB hot-spot areas, reinforced by the immigration of infected possums, is considered to be the 
source of the continuing infection (Caley et aI. , 1 999). 
Since 1989 data on TB in possums had been collected in a longitudinal study on the epidemiology 
of TB in possums conducted on a 23 ha study site in Castlepoint (Pfeiffer and Morris, 199 1 ;  
Pfeiffer, 1 994). Data from these studies has been extracted to produce graphs of the average 
monthly point prevalence (Figure 5) and the time when tuberculous possums died (Figure 6), which 
forms the basis for the seasonal targeted control measures. 
For calculating the average monthly point prevalence of tuberculous possums in Figure 5, all 
possums with clinically detectable tuberculous lesions were taken into account. Palpation of 
superficial lymph nodes is the only practical way in which TB can be detected in live possums 
(Pfeiffer and Morris, 1 99 1 ;  Jackson, 1 995), and suspect lesions detected in this way were 
confirmed by culture of lymph node aspirate. From this graph it can be seen that the highest 
prevalence of tuberculous possums occurs during the Southern Hemisphere summer, with a slight 
peak in winter. 
004 
0.035 
003 
Q) t.> 0025 � Q) 
«l > 002 Q) ... c.. 
0.015 
0.01  
0.005 
0 � � et Q, t C) c � ;; .t c '1S III • :::J :::I • ., .., ... � � .., .., 4( {I) 0 Z Q 
Figure 5. Average monthly point prevalence of TB in possums (data obtained from the 
longitudinal study in Castlepoint). 
In Figure 6 the seasonality of tuberculous possums' deaths is shown, calculated by the number of 
TB-possums found dead per month, divided by the Jolly-Seber estimate (Ibrahim and Trpis, 1 986) 
105 
of number of possums living on the study site. There is a much stronger winter peak than summer 
peak. As tuberculous possums show abnormal behaviour mostly only in the late stages of the 
disease, these dying possums are the critical factor in the transmission of TB to cattle. Normally 
nocturnal possums can be seen out in daylight, attracting the attention and investigation of cattle 
and deer and thus providing opportunities for close contact between the species (Sauter and Morris, 
1 995). 
0.018 
0.016 
0.014 
0.012 
c 0 0.01 1: 0 0. 0 li:. 0.008 
0.006 
0.004 
0.002 
0 c "' ""\ 
Figure 6. Proportion of possums dying from tuberculosis per month of population at risk. 
These fmdings led to the hypothesis that if the direct contact between terminally ill tuberculous 
possums and livestock can be prevented or reduced, then the TB incidence in cattle or deer herds 
should also be reduced. 
Two main methods were investigated for their effectiveness: targeted localised possum control, and 
grazing management of cattle in hot-spot areas. In the following the combination of these two 
methods is called 'on-farm control measures' . 
Targeted localised possum control 
The targeted approach of possum control is a novel method evaluated in this study. The official 
vector control conducted by RC was a 'blanket' approach, covering the whole area of a farm and 
usually an entire locality, typically 20,000 to 50,000 ha. A targeted approach is more likely to be 
more cost-effective, particularly for farms with persistent TB problems. Therefore it was important 
to identify TB hot-spots on the individual farms. 
106 
Localised possum control in TB hot-spot areas was intended to reduce the number of tuberculous 
possums in these areas. As the monthly prevalence of tuberculous possums showed a strong 
summer and a lesser winter peak (Figure 6), possum control was targeted to occur in spring and 
autumn, to reduce both the number of infectious TB possums over summer, and the successful 
establishment of young immigrant possums on farms in autumn, and on a smaller scale in spring. 
These young possums, some of which are infected with TB during rearing by an infected mother, 
disperse from their natal area into new areas principally in autumn, smaller numbers disperse in 
spring. These juveniles are very important in spreading infection, but are particularly susceptible to 
control when they are establishing in a new location. Possum control during this time will therefore 
reduce the establishment of juvenile possums, but also the number of possums in the area and with 
it the number of possible contacts between the possums, which subsequently should result in a 
lower number of tuberculous possums. It was intended to conduct localised possum control on all 
of the farms twice every year, however, due to time constraints this was only possible on four of 
the farms. The other farms were only controlled once a year. 
During the ftrst two years of the project, all carcasses of possums were recovered and submitted to 
a post mortem examination (Jackson et al. , 1995a). Therefore localised possum control was done 
by using leg-hold traps (Victor No. 1 '12 Soft Catch (Montague and Warburton, 2000)) or by bait 
stations, using a cyanide poison (Feratox®), which acted instantly (Eason et al. , 2000).  
Up to 80 traps and 20 bait stations were used at any one time on a farm. These were shifted around 
the farm, covering likely and possible hot-spot areas. Traps were set along locations that showed 
evidence of territorial marking, feeding activities or signs of possum movement ( 'runs' ). Traps 
were set every night and checked the next morning. Traps were lured with apples coated in flour 
and cinnamon, while bait stations were ftlled with special feed pellets, in which encapsulated 
cyanide pellets were mixed at a rate of approximately eight poison pellets to one bait station. 
Trapping and poisoning usually continued for 10 to 14 days, farms with large farm sizes and many 
or large suspect hot-spot areas were trapped for up to four weeks. Thereafter the traps and bait 
stations were moved to the next farm. Especially in the dairy farms, which only had small areas to 
be covered by possum control, it was possible to conduct possum control on two farms at the same 
time. On two focused control farms, where RC control involved trapping of possums, instead of 
poisoning, the RC staff members collected the carcasses and submitted them to the research team 
for post mortem examination. 
Possums, hedgehogs and ferrets caught in leg-hold traps were humanely killed (using a sharp blow 
to the head or by lethal injection of 20% barbiturate). All animals caught were identifted with the 
1 07 
trap number they were caught in and transported to a central place for detailed post mortem 
examination. Any gross lesions suspect of TB were collected and cultured. 
In the third year of the project, five to 1 5  bait stations, depending on number and size of hot-spot 
areas were set up permanently in likely or identified hot-spot areas and fIlled with a slow acting 
brodifacoum poison (Talon®) (Eason et al. , 2000). Additional poison was left with the farmer, who 
was expected to check the bait stations and if necessary to refill them. 
Livestock grazing management practices 
Livestock grazing management practices were intended to keep cattle/deer out of TB hot-spot 
areas, mainly during summer and winter, when the proportion of dying tuberculous possums was 
greatest (November to January and July and August, see Figure 6). The aim was to prevent the 
direct contact between sick tuberculous possums and livestock during these critical times. 
Farmers were advised to change their grazing routine in such a way, that during these critical 
months, cattle did not graze in any paddocks with suspected or known hot-spots. If paddocks 
themselves did not have such hot-spots, but were within 30 to 50 metres of one, the advice was not 
to graze cattle in these paddocks either, or if grazing is necessary in this paddock to put up a 
temporary fence that would keep cattle outside this distance of the hot-spot. On beef breeding 
farms it was recommended to put sheep in these paddocks during these times, as they were less 
likely to interact with possums (Sauter and Morris, 1995) .  
However, the fmal decision regarding grazing management lay with the farmers and they had to 
judge if it was possible or not. If it was impossible to exclude high risk paddocks from grazing it 
was suggested that farmers continue possum control during the time of grazing, either by trapping 
or by shooting. 
Farmers were also encouraged to keep grazing records in order to facilitate the later identification 
of hot-spots if TB positive animals were found. Biannual TB testing was also strongly 
recommended, firstly to help identify hot-spots, and secondly to eliminate any tuberculous cattle as 
soon as possible from the herd. 
Analysis of data 
The frequency and time of vector control operations conducted by the Regional Council on focused 
and standard control farms was analysed in order to establish any differences between the two farm 
groups, that could have influenced the outcome of the present study. 
108 
It was not possible to detennine exactly how many farmers complied with the management advice 
given to them, the research team had to rely on the information given by farmers. In addition, when 
targeted possum control and grazing management was employed, it was not feasible to distinguish 
between the individual effects. However, as the possum control had been done in the fIrst two years 
of the project by the research team, all farms within the focused control farm group were regarded 
as having received TB control measures, independent of whether management advice was followed 
or not. 
TB testing data was analysed for the time on MC, the herd TB status of farms and thus the number 
of farms on MC at the nd of the study, the number of TB animals on the farm irre pe ti e of the 
herd size, the two-year cumulative TB incidence and its reduction versus the pre-project levels of 
1 995/96. Differences between focused and standard control farms were tested for signifIcance 
using the Mann-Whitney U test for categorical data and Chi-squared tests for continuous data 
(Bortz, 1 993). 
The database management software Microsoft Access 97 (Microsoft Corporation, Redmond, W A) 
and spread sheet software Microsoft Excel 97 (Microsoft Corporation, Redmond, W A) were used to 
store and manipulate the data. Statistical analyses were conducted using NCSS 2000 (Number 
Cruncher Statistical Systems, Kaysville, Utah, U.S.A.), and SPSS for Windows version 9.0. 1 
(SPSS Inc. Headquarters, Chica�o, lllinois, U.S.A.). The power analysis was done in Power and 
Precision, release 1 .20 (Borenstein, Rothstein, Cohen, U.S.A.). 
Results 
Power analysis 
Retrospective power analysis was used to provide insight into the scale of study which would have 
been required to achieve signifIcance for each of the variables evaluated. 
In order to establish the statistical power to detect any difference between focused and standard 
control farms, a power analysis was conducted, using data gathered in this study to provide 
accurate estimates of variance. Figure 7 shows the relationship between sample size and power to 
detect differences in the proportion of farms remaining on Movement Control. With 30% of 
standard control farms remaining on Movement Control and a sample size of 35 farms in each 
group, the desired power of 80% would have been obtained, if only 5 %  (i.e. two instead of the 
seven observed) of the focused control farms had remained on Movement Control. If 20% of the 
focused control farms and 30% of the standard control farms remained on Movement Control then 
1 0 9  
a sample size of 294 farms per group would have been required in order to obtain a power of 80%. 
With the current sample size a power of 1 6% was achieved. 
1 .0 
0 .8  
0 .6  
'-Q) 3: Cl 0 .4  CL 
0 .2  
0 .0  
0 
Power as  a Functi on of Effect Size and N 
Two sample proporti ons 
1 00 200 300  
Number  of  cases  per  g roup  
Prop(1 )=0 3 Alpha= . 05 Tai ls=2 
400  
/ P2= 0 .20  
/ P2= 0 . 1 5 
/ P2= 0 . 1 0 
P2= 0 .05  
Figure 7. Relationship between power and sample size at  four different proportions of 
focused control farms and 30% of standard control farms remaining on Movement Control. 
To detect a difference in the two-year cumulative incidence of TB using 35 focused control and 3 5  
standard control farms and a difference of 0.0 1 3  in cumulative incidence, this study had a power of 
22%, using a common standard deviation of 0.04 5 .  With this sample size a difference in cumulative 
incidence of 0 .03 1 would have been necessary to obtain a power of 80%. To detect, with 80% 
power, the difference of 0.0 1 3  observed in this study, a sample size of 190 farms in each group 
would have been necessary. Figure 8 shows the relationship between sample size and power to 
detect a difference in cumulative incidence between the two groups of magnitude 0.0 1 and 0.02. 
1 1 0 
1 .0 
0 .8  
I- 0 .6  (I,) 
� 0 
a.. 0.4 
0.2 
0 .0 
0 
Power as a Function of Effect Size and N 
Two sample t-test 
1 00 200 300 
Number of cases per group 
SD=O .045 Alpha= .05 Tails=2 
/ Di1f= 0.01 
/ Di1f= 0.02 
400 
Figure 8. Relationship between sample size and power to detect a difference in cumulative 
incidence of TB of 0.01 and 0.02 between focused and standard control farms with a common 
standard deviation of 0.045. 
Vector control conducted on focused control farms by the research team 
Due to time constraints by the research team and factors relating to the individual TB situation, not 
all the farms were visited an equal number of times, and control was not conducted on all farms at 
the same time due to time constraints. However, all the farms went through the sequence 
identifying hot-spots - possum control/grazing control - maintaining bait stations/conducting own 
control. 
In total 28 of the 34 focused control farms underwent possum control by the research team. Of the 
other six farms, one (BB, IDlO) did no possum control, but implemented only grazing control by 
keeping cattle out of the suspected TB hot-spot area. It subsequently achieved and maintained 
Clear TB status. Another relatively small 30 ha farm (BD, ID6) conducted its own possum control 
with shooting and bait stations. The farm came off Movement Control in the second year of the 
project, but because of drought conditions the farmer grazed the cattle on a different farm, where 
one of his animals became infected, thus causing the herd to revert to Infected status. The other 
four farms not receiving intensive possum control by the research team (2 BD, IDH, 1 BB) had 
received intensive possum control by the Regional Council in the past and had no suspected high­
risk areas for TB on their farms. Three of these four farms came off Movement Control during the 
first year and the fourth farm (BD) came off during the last year of the project. Approximately half 
the farmers in any one year incorporated the recommended grazing management, depending on 
management constraints and other considerations in management decisions. 
1 1 1  
In the third year of the project, 33 farms remained in the study. One farm had ceased farming cattle. 
On 25 of these farms bait stations using Talon®, a dicoumarin-poison, were set up. The remaining 
nine farms either had their own possum programme in place (denoted with 'Farmer control' in 
Table 3), were at that time involved in an intensive RC vector control programme (denoted with 
'RC' in Table 3), or had a TB Clear status and no suspect high-risk TB areas (denoted with 'None' 
in Table 3). The number of bait stations set up depended on the size of the farm, the number and 
size of suspectJknown hot-spot areas and the willingness of the farmer to keep them going. 
Relatively more bait stations were set up on farms where the farmer had shown active control and 
interest during the previous two years. Due to the action mechanism of the poison in the bait 
stations, no possum carcasses could be recovered. These control efforts are therefore not included 
in Table 3. 
On nine of the focused control farms, tuberculous feral animals were found during project 
activities. Because tuberculous possums were not found on the other farms , this did not mean, that 
they did not have any TB hot-spots on the farm. Therefore any suspect areas continued to be 
controlled. On seven of the farms where tuberculous animals were found, one or more tuberculous 
possums were found, on one farm tuberculous ferrets and on one farm a tuberculous possum, ferret 
and hedgehog were found within the same hot spot area. On average 89 possums were trapped on 
each of the farms (range from 3 to 405). On beef breeding farms, on average 144 (range 4 - 405) 
possums were caught, on beef fmishing farms on average 1 1  (range 3 - 21 )  and on dairy farms on 
average 30 (range 3 - 71 )  possums were caught. 
Table 3 gives the details of possum control by the research team per farm. The cyanide-bait stations 
used had eight poison pellets in the station, therefore these stations had the potential of killing eight 
possums. For the calculation of trap nights on each farm, the bait stations were arbitrarily judged 
the equivalent of five traps, as it was difficult to assess how many poison pellets were left amongst 
the feed pellets. Farms that came off MC early in the project (denoted by b in Table 3) generally 
received less control in the second year of the study. 
1 1 2 
Table 3. Details of possum trapping and poisoning on the focused control farms, together 
with their effective farmed area. 
Farm Total Total number of Total number of Tuberculous Number of 
ID effective trap and poison possums caught feral animals bait stations 
farm size nights (during 1st (during 1st and (found during set up in third 
(ha) and 2nd yr) 2nd yr) 1st and 2nd yr) year3 
Beef breeding farms 
2 1 600 1 282 405 1 2  
3 1 880 249 1 14 1 0  
4 800 400 43 8 
5 729 205 1 26 1 6  
8 b  1 22 342 53 5 
1 0  1 200 Only grazing control Grazing control 
1 3 b  1 270 207 7 5 
1 7  56 200 28 4 
23 2900 703 1 32 1 possum 20 by RC 
25 1 21 5  382 43 8 
26 b 1 500 Possum control by RC only None 
27 1 075 732 1 14 1 possum 5 
1 ferret 
1 hedgehog 
28 4500 41 3 plus 1 600 by 360 2 possums RC 
farmer control 4 ferrets 
31 1 1 45 385 67 6 
32 1 1 85 590 267 1 possum 1 0  
34 800 51 3 252 1 possum 1 8  
1 ferret 
Beef dry stock farms 
1 323 361 9 1 0  
6 30 Possum control by farmer only 2 possums Farmer control 
7 b  300 96 21 7 
1 2 b  1 1 0 Possum control by RC only None 
1 5 b  1 8  57 1 0  4 
24 1 33 Possum control by farmer and RC only None 
29 b 590 1 68 3 5 
33 36 Possum control by farmer in first year 1 possum No cattle 
Dairy farms 
9 1 045 596 21 None 
1 1  b 1 53 45 1 8  3 
1 4 b  60 1 57 21 5 
1 6 b  76 68 6 4 
1 8  1 54 283 3 6 
1 9  1 95 1 885 34 1 possum 1 0  
20 221 285 55 5 
21 200 450 32 8 ferrets 6 
22 b 1 1 0  294 71 5 
30 b 1 04 Possum control b� RC onl� None 
a 'Farmer control' if farmers conducted their own vector control already and were not in need of bait stations; 
'No cattle' this farmed ceased farming cattle in the third year; 'None' if the farm had a TB clear status and no 
suspect/known hot-spots. 
b Farm came off MC before 12/1997 
1 13 
Vector control by Regional Councils 
In total the focused control farms had 42 and the standard control farms 40 'initial' vector control 
operations by RC, as recorded in the NLDB. The percentage of focused and standard control farms 
being under RC vector control programmes from 1988 to 1999 is shown in Figure 9. Eight focused 
control farms and eight standard control farms had received a second round of vector control 
operations, several years after the conclusion of the fIrst one. For the analysis only the date of the 
fIrst initial vector operation was used. There was no significant difference in the dates of initial 
vector control operations between focused and standard control farms (Mann-Whitney U test, Z=-
0.935, p=O.350). 
() 
ex: 25 ... Cl) "CJ c:: 20 :.::s 0 -... 0 1 5  Cl) ... e -... r::: ca 0 1 0  - (,) -0 - 5 c:: Cl) (.) ... 0 Cl) 0. 
• • I I 
2>'l.- 2>t>< 2>'0 2>CO ,,<15 ,,<15 ,,<15 ,,<15 
Year 
- o Focused control 
• Standard control 
Figure 9. Percentage of farms under RC vector control, assuming that a control operation 
lasts for four years. 
Three standard control farms received initial vector control prior to 1 992, the first year six of the 
focused control farms had received vector control. Vector control was equally irregular for both 
farm groups. 
Setting the duration of any control operation to a maximum of 4 years, then the focused control 
farms had on average 2.55 years of control prior to July 1996, the commencement of the project, 
whereas the standard control farms had 2.92 years control (Mann-Whitney U test, Z= 1 .22, 
p=O.222). 
1 14 
Analysis of tuberculosis testing records 
Time spent on Movement Control and Herd TB status at the end of the project 
All 67 focused and standard control farms were on Movement Control (MC) at the beginning of the 
project in July 1 996. Seven of the focused control farms and eight of the standard control farms had 
one clear test just prior to the commencement of the project. 
The median time spent on MC within the pre-project period ( 1994 until 1996) was 28.5 months for 
the focused control farms (n=34), and 29 months for the standard control farms (n=33). For the 
study period ( 1997 until 1999) the median time on MC was 1 5  months for the focused control 
farms (n=33, and 1 1  for the standard control farms (n=28). 
During the three years of the project 30 (90.9 %) of the 33 remaining focused control farms came 
off MC, whereas 22 (78 .6 %) of the remaining 28 standard control farms came off MC (Fisher' s 
exact test p=O.28). Five of the focused control farms and three of the standard control farms 
subsequently became infected again and were put back on MC restrictions. On average these five 
focused control farms had 8.8 TB-free months in between the MC periods; the three standard 
control farms had on average 7.7 months in between. Three (9. 1 %) of the 33 focused control farms 
and six (2 1 .4 %) of the 28 standard control farms did not come off MC at all during the time of the 
project (Fisher exact test, p=0.279). 
One of the five focused control farms that came off MC and subsequently became re-infected, was 
conducting its own intensive possum control on the study property. However, due to drought the 
farmer had to graze his cattle n a neighbouring property that was known to be infected with TB . 
The cattle were all tested clear before they left the home farm. When they returned after five 
months, they were tested again, and one animal reacted to the skin-test, and subsequently was 
found at slaughter to contain lesions. All remaining animals grazed thereafter on the home property 
and subsequently tested clear. As the main purpose of the study was to evaluate the effectiveness of 
on-farm control measures, this farm was therefore classified as 'Off MC' in the analysis. The 
farmers of the nine standard control farms that were under MC at the end of 1 999, all stated that 
they suspected their home-farm, but not run-offs or other grazing options, as being the source for 
the TB infection in their cattle. 
Table 4 gives the TB status at the end of the project in December 1 999 for both groups of farms. 
1 15 
Table 4. Number of focused and standard control farms by TB status at the end of the 
intervention programme (December 1999). 
TB status Focused control farms Standard control farms p-value 
n % of % of n % of % of (Chi-squared 
remaining total remaining total test) 
herds herds herds herds 
Clear Status (Off 26 78.8 76.5 1 9  67.9 57.6 0 .33 (remaining 
MC) herds) 
0 . 10  (all herds) 
Infected Status 6 1 8.2 1 7.6 6 21 .4 1 8.2 
with reactors in 
last test 
Infected Status 1 3.0 2.9 3 1 0.7 9 . 1  
with one clear 
whole herd test 
Ceased farming 1 2.9 5 1 5.2 0.1 1 
cattlea 
Total 34 33 
aduring the project period 1 996-1 999 
Table 4 shows that 15% of the standard control farms and 3% of the focused control farms ceased 
farming cattle during the study period (Fisher exact test p=O. 105). The reasons given by the six 
farmers for ceasing to farm cattle were a mix of personal and enterprise business reasons. Two 
farmers retired, one concentrated more on off-farm employment and the other three farmers stated 
reduced profitability of cattle as a result of general economics and TB in their herd. Table 4 shows 
also that of the farms still farming cattle, 79% of the focused control farms and 68% (Clear status 
plus Works monitored) of the standard control farms were off MC at the end of the project. 
Comparing herd TB status within each of the herd types (DH, BB, BD), the project achieved a 
higher differential impact relative to standard control farms in the beef breeding farms than in the 
other two herd types, which were cleared of infection in most cases in both groups (Table 5). All 
focused control dairy and focused control beef dry stock farms came off MC during the time of the 
project. However, subsequently one dairy farm was re-infected. In the beef breeding farm group 
8 1  % of focused control BB farms and 57% of standard control BB farms came off MC during the 
time of the project (Fisher's exact test p=O.236). However, three beef breeding farms in the focused 
control and three in the standard control farm group subsequently became re-infected, leaving 63% 
of focused control and 36% of standard control beef breeding farms off MC at the end of the 
1 16 
project in December 1 999 (X2 = 2. 14, p=O. 143), considering only beef breeding farms which were 
still in cattle production at the end of the project. 
Table 5. Transitions of TB status of focused and standard control farms during the project 
period 1997-1999. (1= infected, C= clear status, D= disbanded). 
Farm Group Herd type 1 -> C 1 -> C -> I I -> I I -> D I ->C ->D Total 
Focused All farms 26 4 3 34 
control BB 1 0  3 3 1 6  
BD 7 8 
OH 9 1 1 0  
Standard All farms 1 9  3 6 3 2 33 
control BB 5 3 6 2 1 7  
BD 4 2 6 
OH 1 0  1 0  
Number of TB cattle 
As the annual pattern of TB cattle numbers is variable and dependent on test dates, any small 
changes can result in misleading patterns. Thus the present analysis used two-year blocks, 
representing the periods prior to the study, initial study period and late study period. Figure 10 and 
Table 6 show the percentage of focused and standard control farms that had at least one, two or 
three reactors in any one of the pairs of 1 995/96; 1 997/98 and 1998/99. Although by the final year 
( 1 999) proportionately fewer focused control f� still had reactors, when two-year blocks were 
considered the difference in this particular indicator was not yet apparent for the 98/99 block. Of 
the 33 focused control farms, 1 3  had reactors in 1998 and/or in 1999, while 1 1  of 28 standard 
control farms had reactors in 1 998 and/or 1 999. However, in the last two years of the project a 
lower percentage of focused control farms than standard control farms had two or more reactors 
(Figure 1 0) .  The chi-squared test result for the difference between the focused control and the 
standard control farms for having two or more reactors was X2=3.45 (p=O.063). The difference for 
having three or more reactors was X2=5 .09 (p=O.024), indicating that the control measures achieved 
a significant lower likelihood of having three or more reactors in any one year. 
1 1 7 
Focused control farms Standard control farms 
1 00 
III 80 E 
"-� 60 No. of .... Cl TB cattle -c 40 ell 0 >=1 (.) 
"-ell 20 • >=2 Q. 
0 >=3 
0 
95196 97198 98199 95196 97198 98199 
years years 
Figure 10. Percent of focused and standard control farms that had equal or more than one, 
two, or three reactors in any one of the two years. 
Table 6. Number of focused and standard control farms with one, two or more reactors in 
any one of the two years. 
Focused control farms Standard control farms 
Years � 1 TB � 2 TB � 3  TB n farm � 1 TB � 2 TB � 3 TB n farm 
animal animals animals cattle animal animals animals cattle 
95/96 34 27 21 34 33 25 1 9  33 
97/98 20 1 1  7 34 1 5  9 9 28 
98/99 1 3  5 3 33 1 1  1 0  9 28 
Cumulative incidence and its reduction over three years 
For focused and standard control farms the two-year cumulative incidence at the end of the project 
was compared with the one of the two years prior to the project (Table 7). 
1 1 8 
Table 7. Average two-year cumulative incidence (cum inc.) of TB animals in focused and 
standard control farms and the reduction versus the 1995/96 cumulative incidence. 
Group Cum inc. Cum inc. Reduction Cum inc. Reduction 
1995/96 1 997/98 vs. 95/96 1998/99 vs. 95/96 
All farms included in the study 
Focused control farms 0.0629 0.01 38 78. 1  % 0.0034 94.6 % 
Standard control farms 0.0354 0.01 65 53.4 % 0.0 161 54.5 % 
Only farms included that existed at the end of the project: 
Focused control farms 0.0486 0.0090 81 .5 % 0.0034 93.0 % 
Standard control farms 0.0350 0.01 65 52.9 % 0.01 61 54.0 % 
The Mann-Whitney U-test for detecting a difference in the cumulative incidence yielded a z-value 
of 0.0 125 (p=0.990) for the 1 995/96 period, z=0.4056 (p=0.685) for 1 997/98, and z=0. 1 553  
(p=0.877) for the 1998/99 period. 
Figure 1 1  and Figure 12 present frequency histograms of the two-year cumulative incidence for 
focused and standard control farms for the period prior to commencement of the study (Figure 1 1 ) 
and for the last two years of the project (Figure 1 2) .  
20 
VJ '1 5 E "-:! 
'0 1 0  "-11) ,Q E :l 5 = 
0 o 
Focused control farms 
I I I I I 
CD q o 
� N 
ci 
I 
Cum incidence 1995196 
CD M o 
I 
N "': o 
I 20 1 5  
1 0  
5 
o o 
Standard control farms 
.1. I I 
I.D q o 
Cum incidence 1 995196 
I.D � o 
Figure 1 1 .  Frequency histogram of pre-study cumulative incidence of TB animals in 1995/96 
for focused and standard control farms. 
1 19 
Focused control farms Standard control farms 
25 25 
tIl 20 20 e .. 
.:! 1 5  1 5  '5 .. III 1 0  1 0  .c e := = 5 5 
0 • 0 • 
0 (£) N (0 v l"l � N 0 (£) N co v (') (D N q .... ... N 0 v 0 .... .... N c:i l"l "'= 0 c:i c:i c:i 0 c:i c:i 0 c:i c:i 0 0 
Cum incidence 1998199 Cum incidence 1998199 
one farm disbanded five farms disbanded 
Figure 12. Frequency histogram of cumulative incidence of TB animals 1998/99 for focused 
and standard control farms. 
Table 8, Figure 13 and Figure 14 present the two-year cumulative incidences of TB animals (pre­
project and end of project), stratified by herd enterprise type and farm group. In Table 8 the 
reduction in cumulative incidence achieved over the three years is also given. 
Table 8. Within group average of two-year cumulative incidence of TB animals in focused 
and standard control farms, stratified for herd type; and reduction in cumulative incidence 
achieved. (The number of farms in each category is shown in brackets.) 
Group Herd type Cum Inc. 1995/96 Cum Inc 1998/99 Reduction 
Focused control farms BB 0.0321 (n=16) 0.0043 (n=1 6) 86.6 % 
BD 0.1714 (n=8) 0.0030 (n=7) 98.2 % 
DH 0.0254 (n=1 0) 0.0024 (n= 10) 90.6 % 
Standard control farms BB 0.0497 (n=17) 0.0308 (n=14) 38.0 % 
BD 0.0254 (n=6) 0.0000 (n=4) 1 00 % 
DH 0.01 69 (n=10) 0.0019 (n= 10) 88.8% 
Dairy and beef finishing farms in both groups had a similar reduction in cumulative incidence 
(Table 8). However, the beef breeding farms in the focused control farm group achieved a higher 
reduction than the ones in the standard control farm group. 
The Mann-Whitney V-test for detecting a difference in the cumulative incidence between focused 
and standard control farms yielded a z-value of 1 . 1026 (p=O.270) for the beef breeding farms, 
z=O.8956 (p=O.370) for the beef finishing farms, and z=0.4323 (p=O.665) for the dairy herds. 
120 
BB Focused control 
12  
e 10 .. 8 .:! 
li 6 .. Cl ,&j 4 E 
i 2 
0 
11 
11 I II  
BB standard control 
1 2  
• 10  
e 
! 8 
� 6 .. Cl ,&j 4 E :::J Z 2 
0 
11 I 
11111 1 1 
Cum incidence 1995/96 
1 2  
10  
8 
6 
4 
2 
o 
1 2  
1 0  
B 
6 
4 
2 
o 
BD Focused control 
I 
11 1 
BD standard control 
I 
11 I 
CUm incidence 1995.e5 
I 
1 2  
1 0  
B 
6 
4 
2 
o 
1 2  
1 0  
B 
6 
4 
2 
o 
OH Focused control 
I .. 
OH standard control 
11 
CUm incidence 1MS.S 
Figure 13. Distribution of 2yr cumulative incidence (skin test reactors plus lesioned culls) for 
1995/96, stratified by farm group and herd type. 
12  
� 10 .. 8 :! 
� .. 6 Cl 
i 4 :::J 2 z 
0 
12  
• 10 E .. :! 8 
li 6 .. Cl &l 4 
i 2 
0 o 
BB Focused control 
J 
BB standard control 
three farms disbanded 
I I 
CUm Incidence 199..", 
12 
10 
8 
6 
4 
2 
o 
1 2  
1 0  
B 
6 
4 
2 
0 0 
BD Focused control 
one farm disbanded 
1 
BD standard control 
two farms disbanded 
8 CD .". � � ..- N 0 ci ci ci C) 
Cum tnddence 1998111 
1 2  
1 0  
8 
6 
4 
2 
o 
1 2  
1 0  
8 
6 
4 
2 
0 
OH Focused control 
I 
I 
OH standard control 
I 
o CD ID � N � q � � (P! 0 o 0 C) 0 
Cum incidence 1a/99 
Figure 14. Distribution of 2yr cumulative incidence (skin test reactors plus lesioned culls) for 
1998/99, stratified by farm group and herd type. 
1 2 1 
The control measures employed in this study achieved a higher reduction in two-year cumulative 
incidence in non-RC control areas than in control areas (Table 9). 
Table 9. Two-year cumulative TB incidence for 1995/96 and 1998/99 for focused and 
standard control farms, stratified by Regional Council control received prior to June 1996. 
Group RC Control Cum inc. 1995/96 Cum inc. 1998/99 Reduction 
Focused control farms Yes 0.0166 (n=18) 0.0034 (n=1 8) 79.5 % 
No 0.071 9 (n=16) 0.0054 (n=15) 92.5 % 
Standard control farms Yes 0.01 79 (n=17) 0.0004 (n=15) 97.8 % 
No 0.0346 (n=16) 0.0231 (n=13) 33.2 % 
Discussion 
This study was the first intensive intervention programme that has evaluated the effect of specific 
targeted control methods on TB outcomes in livestock. These control methods consisted of targeted 
hot-spot possum control in autumn and spring, and grazing management during winter and 
summer. While the differences found between the focused and standard control farms were mostly 
not statistically significant because of the limited number of herds which could be included, the 
study nevertheless provides valuable information on trends of such low-cost, practical on-farm 
control measures, consisting of targeted possum control and grazing management. 
The number of farms in the study, 35 focused control and 35 standard control farms, was set by 
capacity for conducting the field work and the number of farmers available, not to meet statistical 
power requirements. A sample size of 200 to 300 farms per group would have been needed to 
obtain statistical significance with the results obtained in this study. It was recognised that an 
intensive study of this nature involving 35 focused control farms would be unlikely to achieve 
statistical significance in most variables and the primary aim was to evaluate the technology and 
determine whether evidence of a pattern of response could be detected. Consistency of data with 
that of other studies (Chapter 4) would provide strength to the interpretation of data even if the 
results are not statistically significant. While most of the differences found in this study were not 
significant by conventional standards it is proposed that the overall pattern of differences, though 
individually not significant, when combined provide valuable evidence for the effectiveness of the 
control measures. 
1 22 
Focused control farms were significantly less likely to have multiple TB animals in the last two 
years of the project than the standard control farms. Other differences, which did not reach 
significance at p=O.05, were: A higher proportion of focused control than standard control farms 
came off MC during the three years of the project, a higher proportion of focused control than 
standard control farms were off MC at the end of the project, a lower proportion of focused control 
farms than standard control farms remained continuously infected, and focused control farms 
achieved a higher reduction in cumulative TB incidence. All these results point in the same 
direction - that focused control farms achieved more effective TB control than standard control 
farms. These results therefore support the effectiveness of the implemented control measures. The 
only outcome that differed from this direction was the median time spent under movement control 
by focused and standard control farms, which might be attributable to . the higher pre-study 
cumulative TB incidence on focused control farms. 
Although the difference was not significant, the focused control farms had a higher pre-study 
cumulative incidence than the standard control farms. Thus it might be argued that it was easier for 
them to reduce the cumulative incidence, as it might be easier to reduce the reactor numbers from 
five to two, than from one to zero. However, the result that more focused control than standard 
control farms came off MC during the study period cannot be explained by the higher initial 
incidence. Therefore, a more realistic interpretation is that the focused control farms were if 
anything the more difficult farms of the total group, even though allocation was randomly. 
It was expected that a number of standard control and focused control farms would come off MC, 
even without any additional targeted control implemented, and the percentage of standard control 
farms which achieved this is similar to the percentage of all farms over the W airarapaIW ellington 
region (63.5%) coming off MC between 1995 and 1999 (Animal Health Board, 1 999). 
There were no differences in the achievements of dairy and beef dry stock farms between the 
focused control and the standard control farm group. Between 90% and 100% of both groups 
achieved a Clear TB status by the end of the project and a similar reduction in two-year cumulative 
TB incidence. In contrast, a higher proportion of beef breeding farms in the focused control group 
than in the standard control group obtained Clear TB status (63% versus 36%). Also the focused 
control beef breeding farms achieved on average a higher reduction in cumulative incidence than 
the standard control beef breeding farms (85% versus 38%) and in the standard control beef 
breeding group one farm had a high TB cumulative incidence in 1998/99, whereas there were no 
longer any high-incidence focused control farms. Although these differences were not significant at 
the alpha level of 0.5, the number of farms in this enterprise category only comprised 30 farms in 
total. It is likely that the differences in success between the herd types are due to farm 
characteristics. Dairy and beef dry stock farms are generally on more productive land with less 
123 
potential possum habitat, thus less likelihood of having persistent TB hot-spots on fann. Targeted 
vector control cannot offer any additional benefits on these fanns over standard RC vector control, 
since standard control was able to virtually eliminate TB from these farms without the need for 
extra effort. In contrast, the beef breeding farms have generally larger farm sizes and more possum 
habitat, therefore a higher likelihood of more hot-spots and of greater contact between livestock 
and diseased possums. Standard control did not seem to be able to reliably reduce the likelihood of 
these contacts to a low level, thus additional targeted control was beneficial. The differences found 
in the study between the herd types might therefore suggest that the implication of the suggested 
control measures is especially effective on the beef breeding farms or possibly on fanns with these 
habitat characteristics. It is therefore proposed that a similar study is conducted using only beef 
breeding farms.  If a group of 35 beef breeding farms had achieved the same proportional changes 
against 35 standard control farms as in the present study, this would not only support the present 
study, but the results would also be significant at the conventional statistical level of p=O.05. 
The differences found in this study cannot be attributed to the official vector control provided by 
the RC, as focused and standard control farms received a similar amount of official control by this 
organisation. On fanns that did not receive RC control effort prior to the study period, the TB 
incidence was reduced much more on focused control farms than on standard control fanns (92% 
versus 33%), whereas the standard control farms achieved a marginally higher reduction on farms 
that had previously received RC control (98% versus 80%). This would suggest that the 
employment of targeted on-farm control measures can be effective, and may match the results of 
the much more expensive blanket control undertaken by RCs. Often farmers believe that they 
cannot have any worthwhile effect on TB control at fann level, and that they need regional possum 
control to make progress; the results in this study show that this might not be correct. Further 
studies should address this in more detail. 
When evaluating this study, one has to take two additional factors into account: the study was 
evaluated after only three years, and it was done under realistic conditions of commercial farming. 
It was recognised that the control measures would take some time to have an effect on the TB 
outcomes measured, mainly as the animals are mostly tested only annually or biannually and 
therefore there is a considerable lag before improvements in TB control show up in TB testing 
results. Due to the sensitivity of the caudal fold skin tuberculin test, about 20% of infected cattle 
may be incorrectly classified as 'negative' (Ryan et aI. , 199 1 ). Because of these factors, it was 
expected that reactor rates would decline gradually. The one significant result was the lower 
number of focused control farms having multiple reactors in any one year. It is therefore likely that 
if a fourth year had been included in the study, the difference in the number of farms coming off 
1 24 
MC may have been greater, as it is more likely that farms with only one reactor come off MC 
within one year than farms with more than two reactors. 
Within this context also the likelihood of cattle-to-cattle transmission has to be considered. In the 
present study there were no means to differentiate between vector-to-cattle and cattle-to-cattle 
transmission. It is known that cattle-to-cattle transmission can occur in the absence of wildlife 
infection (Menzies and Neill, 2000), but also in New Zealand with infected feral populations 
(Barlow et aI. , 1997a; Barlow et aI. , 1 997b). The results of a simulation model using no within­
herd transmission, consistently underestimated the observed number of reactors after control 
strategies (Barlow et aI. , 1997a). In a model fitted to the Hohotaka area in NZ, it was deduced that 
cattle-to-cattle transmission contributed between 20% and 32% of the infections prior to the control 
of wildlife, which is below the threshold needed for maintaining the disease without input from 
infectious wildlife. It was concluded that the control of wildlife vectors was the most effective 
method of reducing cattle TB. The model suggested that the reduction in transmission of TB from 
wildlife will result in an equal proportional reduction in cattle reactor rates (Kean and Barlow, 
1 999). It cannot be excluded that cattle-to-cattle transmission occurred within some of the study 
herds, which might have influenced the rate of decline of reactor numbers on these farms. 
The second point to consider in evaluating the present study, the fact that the study was done under 
realistic commercial farming conditions, means these conditions could not be perfectly controlled. 
We know of two farmers in the focused control group who have not followed-up the vector control 
work in the third year. Other farms followed the vector control recommendations, but were not able 
to implement the grazing recommendations. If it had been possible to implement all control 
measures on all farms the TB outcomes on focused control farms might have been more different 
from the outcomes in the standard control farms. 
To illustrate the effectiveness of employing control measures in reducing TB on farms, two focused 
control beef breeding farms are discussed in more detail here: The first farm never received any 
externally provided possum control, such as by the RC. The farmer himself employed a possum 
trapper for several years. The trapper, who was skinning all possums for commercial purposes, 
suspected some possums were tuberculous on the basis of typical lesions. The farm had a 
continuous problem with TB and the farmer made the decision to graze only sheep in the areas 
where the trapper found the suspected tuberculous possums. By the time the project started, this 
farm had already used this practice for one year, without employing the trapper any more. The farm 
came off MC within another year and has stayed clear ever since. This farm shows that grazing 
management on its own may be highly effective in reducing the risk of spread of TB from vectors 
to cattle, if the population of tuberculous possums is small and localised. The influence of earlier 
on-farm control is difficult to assess in this case. 
125 
On the second focused control farm, a tuberculous possum was found during the wildlife survey. 
The farm had been infected for at least 13 years. By setting up grazing plans that avoided this high 
risk area and by conducting possum control in this area, the farm managed to come off MC for the 
fIrst time. This farmer kept strictly to the recommendations given to him, he followed the grazing 
management over all three years and he was very dedicated to the possum control effort. At the end 
of 2000 (one year after the study concluded) he had another clear TB test and achieved the status of 
C3, something he had never believed to be possible. This outcome shows that by following the 
specifIc grazing and possum control recommendations, TB can be eliminated from cattle on farms 
with TB hot spots. 
One crucial component of this study was the knowledge about TB hot-spots, which formed the 
basis for targeted possum control, rather than relying entirely on whole-farm approaches, as used in 
the standard control system. It would have been desirable to conduct a detailed wildlife survey on 
each of the farms to obtain more exact data on suspect TB hot-spot areas. However, the time and 
resources required were not available. On nine of the focused control farms tuberculous feral 
animals were found. This does not mean that on the other farms no tuberculous animals were 
present, since with low prevalence of TB in feral animals the sampling intensity was not sufficient 
to have high detection sensitivity. The aim was to reduce possum populations in all parts of the 
farm considered to be potential hot-spots, since the goal was to achieve TB freedom. 
The second crucial components of this study was the close working relationship between the 
research team and farmers, which was partly achieved by using only two persons in the research 
team. One was a research veterinarian, and the other one a fIeld worker, who had been managing a 
farm for many years. This combination proved to be highly effective, as various farmers could 
relate more to one than to the other. Often, farmers would tell one team member more than the 
other, or accept advice more readily from one person than from the other. The fIeld worker was 
considered by many farmers as their equal, as one of their own. The close working relationship 
between the veterinarian and the fIeld worker resulted in practical solutions, often a compromise 
between what would be desirable and what was possible. So the study could be argued to have 
achieved what Syme called for in intervention studies with humans relating to smoking and 
coronary diseases: "experts must learn to be creative and inventive enough to become experts in the 
role of not being an expert" in order to achieve familiarity with the real community (Syme, 1 998). 
Often, farmers were more readily prepared to try out certain control methods once the fIeld worker 
told them about his own experience on the farms, therefore the advice was seen less as theoretical 
science. By spending several days or even weeks on the farms, farmers could observe the work 
being put into vector control on their property and the autopsies of feral animals. Often, the 
enthusiasm of the research team was conveyed to the farmers. However, this enthusiasm and 
1 26 
motivation required boosters from time to time, either by frequent phone calls or farm visits. 
Essentially the 'hands-on' practice of the research team empowered the farmers to assist TB control 
on their own farms. This empowerment was found to be a crucial part of smoking related 
intervention studies (Syme, 1998). Involving farmers in possum control and the decision process of 
setting up specific grazing routines, was a part of assisting farmers in .their own learning process 
(Walker and Bell, 1994). 
For the analysis of reactor numbers it was decided not merely to use the number of lesioned 
animals, the criterion AgriQuality is using to determine the TB status of the herd. If only animals 
with lesions typical of TB at slaughter were included in the analysis, animals with very small 
lesions would have been classified as negative although they were truly infected. For the analysis it 
was assumed that all skin test positive animals were true TB infected animals, unless they tested 
negative in an ancillary test. Using this criterion was judged to give results closest to the true status 
of the herd, than the less accurate system currently used by AgriQuality. In the Wairarapa generally 
no herds under MC are eligible for ancillary testing (Animal Health Board, 1998b). 
Interestingly, a greater proportion of standard control than focused control farms ceased farming 
cattle during the three years ( 15% versus 3%). Although this difference was not statistically 
significant at the 0.05 level, it was marginally above 0. 10. While TB was not the only reason to 
cease farming cattle on those farms, TB was always . mentioned among other reasons. Financial 
margins in farming are very narrow - and having a herd infected with TB, these farmers felt, 
reduced these margins. As the farms were allocated randomly to focused control and standard 
control groups, but five standard control farms stopped farming cattle while only one focused 
control farm did so, it is possible that some focused control farms saw in the project new hope for 
controlling TB on their farms, and thus the hope of being able to sell their cattle without discount 
and therefore increasing the profit margins. If more TB infected herds than non-infected herds go 
out of farming cattle, this will lead to a reduction in the number of herds infected, without 
necessarily indicating that TB is becoming less prevalent in the Wairarapa. In contrast, the 
indicator for wildlife infection is removed and if no vector control is conducted any more on these 
farms, tuberculous possum populations might increase and dispersing juveniles might spread into 
neighbouring areas, therefore posing a higher risk to farms in the same area in a few years time. 
The on-farm control measures tested in this study can be adapted to each individual farm, and can 
incorporate new control methods, such as vaccination of possums. Research into this field is 
making good progress (S.Norton, pers. comm., 2000) and it provides an effective measure which is 
complementary to poisoning possums. The devices for vaccination are intended to be less work­
intensive than bait stations to operate, and by shifting the vaccination devices around the farm, the 
whole of the farmed area can be covered. 
127 
In order to achieve effective control and eradication of TB in livestock, it is necessary that national 
organisations and farmers work together on the control programme. Therefore a greater awareness 
for effective on-farm control methods has to be created. Farmers have to be convinced that it is 
possible for them to help in the control and to achieve control of TB on their farms.  Using targeted 
vector control on farms as well as on the regional level could result in a cost-effective use of the 
available resources. With the availability of computer expert systems, such as EpiMAN(TB) 
(McKenzie, 1999) identification of TB hot-spots is made easier and provides the basis for targeted 
vector control, therefore facilitating the use of such on-farm control measures. According to Morris 
et al. ( 1995), farmers are more likely to adopt innovations or new ideas if they had been tested on 
other farms in the area. As the present intervention study was conducted not on research farms, but 
on commercial, randomly selected farms, the first step in disseminating the knowledge has been 
undertaken, and should be reinforced by follow-up extension programmes, such as discussion 
groups, media and for example using livestock officers who come regularly on to the farm to TB 
test cattle (for their importance see Chapter 2). However, the adoption of such on-farm control 
measures not only depends on their effectiveness, but also on the availability of resources and the 
prospect of production gains (Morris et aI. , 1995) (see also Chapter 6). 
Conclusions 
The study has provided evidence that specifically targeted possum control, combined with stock 
management practices that minimise or prevent contact with infected vectors are effective methods 
of reducing the herd incidence of TB and the rate at which herds are declared TB-free. These 
strategies are low cost, and are not technologically intensive or complex. They should be easily 
adopted and should contribute significantly to the AHB strategy to reduce the herd prevalence of 
TB. 
Bibliography 
Animal Health Board ( 1995). National TB Strategy: proposed pest management strategy for bovine 
tuberculosis. Animal Health Board, Wellington, New Zealand. 
Animal Health Board ( 1998a). Annual report for the year ending 30 June 1998. Animal Health 
Board, Wellington, New Zealand. 
Animal Health Board ( 1998b). Five Year National Bovine Tuberculosis Pest Management Strategy 
to 30 June 2001 .  July 1998 to June 2001 .  Animal Health Board, Wellington, New Zealand. 
128 
Animal Health Board ( 1999). Regional TB Operational Plan, Wellington Region, 1 July 1999 to 30 
June 2004. Animal Health Board, Wellington, New Zealand. 
Animal Health Board (2000). Bovine Tuberculosis National Pest Management Strategy 2001 -201 1 ;  
Towards a TB free NZ; A discussion paper on the future options. Animal Health Board, 
Wellington, New Zealand. 
Anon ( 1998). Regional TB Operational Plan, Wellington. July 1997 to June 2002. Regional Animal 
Health Committee, Wellington. 
Barlow, N.D., Kean, J.M., Hidding, G., Livingstone, P.G. and Robson, A.B. ( 1997a). A simulation 
model for the spread of bovine tuberculosis within New Zealand cattle herds. Preventive 
Veterinary Medicine, 32: 57-75. 
Barlow, N.D., Kean, J.M., Hidding, G.J., Livingstone, P.G. and Robson, A.B. ( 1997b). A 
simulation model for the spread of bovine tuberculosis within New Zealand cattle herds. 
Preventive Veterinary Medicine, 32: 57-75. 
Bortz, 1. ( 1993). Statistik fur Sozialwissenschaftler. Springer Verlag, Berlin. 
Caley, P., Hickling, G.J., Cowan, P.E. and Pfeiffer, D.U. ( 1999). Effects of sustained control of 
brushtail possums on levels of Mycobacterium bovis infection in cattle and brushtail possum 
populations from Hohotaka, New Zealand. New Zealand Veterinary Journal, 47 (4): 133-142. 
Coleman J., Livingstone P. (2000). Chapter Twenty: Fewer Possums: Less Bovine TB. Montague 
T.L. (editor). The Brushtail Possum. Lincoln, New Zealand Manaaki Whenua Press, pp.:220-
231 .  
Cullen R., Bicknell K. (2000). Chapter Eighteen: Economic Analysis of Possum Management. 
Montague T.L. (editor). The Brushtail Possum. Lincoln, New Zealand Manaaki Whenua 
Press, pp.: 198-207. 
Eason C., Warburton B., Henderson R. (2000) .  Chapter Fourteen: Toxicants Used for Possum 
Control. Montague T.L. (editor). The Brushtail Possum. Lincoln, New Zealand Manaaki 
Whenua Press, pp.: 154- 163. 
129 
Hickling, G. ( 1995). Clustering of tuberculosis infection in brushtail possum populations: 
implications for epidemiological simulation models. In: Griffin F, de Lisle G (Tuberculosis in 
Wildlife and Domestic Animals - Otago Conference Series No. 3. University of Otago Press, 
Dunedin, pp. 174-7. 
Hickling, G. J. and Efford, M. G. (1996). Assessing the risk of bovine tuberculosis infection for 
New Zealand cattle herds in TB vector risk areas (VRAs). In: Fletcher, D. J., Kavalieris, L. 
and Manly, B. F. J. (eds.), Statistics in Ecology and Environmental Monitoring 2. Otago 
Conference Series No. 6. University of Otago Press, Dunedin, pp. 79-90. 
Ibrahim, M.S. and Trpis, M. ( 1986). BASIC programme for calculating Jolly-Seber popUlation 
parameters. Bulletin of the Entomological Society of America, 32 (4): 253-255. 
Jackson, R. (1995). Transmission of tuberculosis (Mycobacterium bovis) by possums. Unpublished 
PhD thesis, Massey University, Palmerston North, New Zealand. 
Jackson, R., Cooke, M.M., Coleman, J.D. and Morris, R.S. ( l995a). Naturally occurring 
tuberculosis caused by Mycobacterium bovis in brushtail possums (Trichosurus vulpecula): I. 
An epidemiological analysis of lesion distribution. New Zealand Veterinary Journal, 43 (7): 
306-3 14. 
Kean, J.M. and Barlow, N.D. ( 1999). Evaluating potential sources of bovine tuberculosis infection 
in a New Zealand cattle herd. New Zealand Journal of Agricultural Research, 42 ( 1 ): 101- 106. 
Livingstone, P.G. ( 1997). Update on the New Zealand TB situation. In: Proceedings of a Seminar 
on Possum and Mustelid Control Research. National Possum Control Agencies, Wellington, 
pp. 17-30. 
Lugton, I. W. ( 1997). The Contribution of Wild Animals to the Epidemiology of Tuberculosis in 
New Zealand. Unpublished PhD Thesis, Massey University, Palmerston North. 
McKenzie, J. S. ( 1999). The use of habitat analysis in the control of wildlife tuberculosis in New 
Zealand. Unpublished PhD thesis, Massey University, Palmerston North, New Zealand. 
McKenzie, J.S., Morris, R.S. and Pfeiffer, D.U. ( 1997). Identification of environmental predictors 
of disease in a wildlife population. In: Proceedings of the 8th International Symposium on 
Veterinary Epidemiology and Economics, 0 1 . 1 3 . 1 -0 1 . 1 3.3 .  
130 
Menzies, F.D. and Neill, S .D. (2000). Cattle-to-Cattle Transmission of Bovine Tuberculosis. The 
Veterinary Journal, 160 (2): 92- 106. 
Montague T., Warburton B. (2000). Chapter Fifteen: Non-toxic Techniques for Possum Control. 
Montague T.L. (editor). The Brushtail Possum. Lincoln, New Zealand Manaaki Whenua 
Press, pp.: 164- 174. 
Morgan D., Riclding G. (2000). Chapter Thirteen: Techniques Used for Poisoning Possums. 
Montague T.L. (editor). The Brushtail Possum. Lincoln, New Zealand Manaaki Whenua 
Press, pp. : 143- 153. 
Morris, c., Loveridge, A. and Fairweather, J.R. ( 1995). Understanding why farmers change their 
farming practices: the role of orienting principles in technology transfer. Research Report 
No.232. Agribusiness and Economics Research Unit, Lincoln University, Canterbury, New 
Zealand. 
Morris, R.S.,  Pfeiffer, D.U. and Jackson, R. (1994). The epidemiology of Mycobacterium bovis 
infections. Veterinary Microbiology, 40 (1) :  1 53- 177. 
Pannett, G. ( 1995). Possum control in the Wellington region: how successful has it been. In: Griffin 
F, de Lisle G (Tuberculosis in Wildlife and Domestic Animals - Otago Conference Series No. 
3. University of Otago Press, Dunedin, pp. 294-6. 
Pfeiffer, D. U. ( 1994). The Role of a Wildlife Reservoir in the Epidemiology of Bovine 
Tuberculosis. Unpublished PhD thesis, Massey University, Palmerston North, New Zealand. 
Pfeiffer, D.U. and Morris, R. S. ( 1991) .  A Longitudinal Study of Bovine Tuberculosis in Possums 
and Cattle. In: R.Jackson (Convenor), Symposium on Tuberculosis. N.Z. Veterinary 
Association Foundation for Continuing Education, Palmerston North, pp. 17-39. 
Ryan, T.J. ( 1997). The New Zealand National Livestock Database. In: Epidemiology Program. 
Proceedings of the 1 0th Federation of Asian Veterinary Association (FAVA) Congress. 
Australia, pp. 167-70. 
1 3 1  
Ryan, TJ., de Lisle, G. W. and Wood, P. R. ( 1 99 1 ). The performance of the skin and gamma 
interferon tests for the diagnosis of tuberculosis infection in cattle in New Zealand. In : Dr. R. 
Jackson (Convenor), Symposium on Tuberculosis. New Zealand Veterinary Association 
Foundation for Continuing Education, Palmerston North, pp .  1 43-50. 
Sauter, C.M. and Morris, R.S. ( 1 995). Behavioural studies on the potential for direct transmission 
of tuberculosis to farmed livestock from feral ferrets (Mustela jura) and possums (Trichosurus 
vulpecula) .  New Zealand Veterinary Journal, 43 (7): 294-300. 
Shortridge, E. H. ( 198 1 ) . Tuberculosis in cattle and oppossums. In: Proceedings of a Veterinary 
Public Health Seminar. Palmerston North, New Zealand. 
Syme, L. ( 1 998). Syme presents Keynote Talk at ACE Meeting in San Francisco. EpiMonitor The 
Epidemiology Monitor, 1 9  (9): 1 -6. 
Walker, A. and Bell, B .  ( 1994). Aspects of New Zealand's Experience in Agricultural Reform since 
1 984. MAF Policy Technical Paper 94/5, Wellington, New Zealand. 
1 32  
Bush area on farm in Wairarapa 
Cyanide pellets (in between feed pellets) used for localised possum control 
Chapter 4 
Effectiveness of on-farm tuberculosis control 
programmes: Comparison of Wairarapa Study with a 
Contemporary National Study 
133 
1 34 
1 35 
Abstract 
In New Zealand bovine tuberculosis is transmitted from vector species to livestock, thus 
complicating the control and eradication of the disease. The main wildlife reservoir and vector is 
the Australian brushtail possum. Control of the disease involves livestock and vector species, 
resulting in expensive control programmes. Therefore other control programmes have to be 
investigated, including measures that can be applied at the farm level. Farmers are increasingly 
expected to take greater responsibility for the TB problem on their own properties. 
In four areas with endemic wildlife infection, an intervention study was conducted by a national 
organisation. The study involved 35 focused control farms (receiving interventions) and 68 
standard control farms (receiving no interventions). A multi-disciplinary team of researchers, pest 
control staff and veterinary staff, together with the farmer developed a customised, farm-specific 
plan for controlling and/or eradicating TB on each of the focused control farms. The team was 
active for two years, reviewing and adapting the plans frequently. Implementation of the control 
work was the sole responsibility of the farmers. 
The evaluation of the effectiveness of the control measures indicated that focused control farms had 
an advantage over standard control farms. The two-year cumulative TB incidence was reduced 
more on focused control than standard control farms, especially on farms with persistent TB 
problems. 
The effectiveness of this intervention study was also compared with that of the Wairarapa project 
(Chapter 3). In the Wairarapa project a higher proportion of focused control farms achieved clear 
TB herd status and the reduction in cumulative TB incidence was greater for the Wairarapa than the 
national project focused control farms.  It was concluded that the hands-on operational approach 
taken in the Wairarapa project was more successful than the team advisory approach used in the 
national project. 
Introduction 
In New Zealand the Australian brushtail possum (Trichosurus vulpecula) is a reservoir host for 
bovine tuberculosis and a vector of the disease to livestock (Morris et aI. , 1994). In some areas of 
New Zealand also other feral animals, such as ferrets or stoats, are suspected of transmitting 
tuberculosis to livestock (Ragg et al. , 1995). The existence of feral reservoir species necessitates 
136 
not only control o f  TB i n  cattle and deer, but also in feral animals. However, these feral control 
operations are very expensive (Animal Health Board, 1995) .  Therefore other control options have 
to be investigated that require the assistance of farmers, not only in a financial way, but also in a 
direct way of controlling his/her own property. In the national pest management strategy paper of 
the Animal Health Board ( 1995; 2000) one of the objectives was to 'encourage farmers to take 
greater responsibility for the Bovine TB problem on their individual farms and in their area and 
region' .  Out of this objective the present study was developed as a pilot programme, in which 
farmers of herds infected with TB were advised with a customised plan for controlling and/or 
eradicating TB on their farm (Rhodes, 1997). As a result of previous submissions on the draft pest 
strategy the project adopted a team approach in providing advice and support for farmers. 
The present chapter describes this national project, evaluates its effectiveness and compares it with 
the Wairarapa project (Chapter 3). Four main objectives of the national project were identified, as 
stated by the organisers who conducted this project (Rhodes, 1997): 
• 'to assist in changing farmer attitude toward TB and self-motivated on-farm control' 
• 'to reduce the length of time the herd remained on Movement Control' 
• 'to reduce the within-herd TB prevalence' and 
• 'to maintain a high health status for the herd over time' 
It was considered important that farmers have ownership of an effective programme for the 
management of TB on their farm (Rhodes, 1997). 
Materials and Methods 
Study areas 
The study was conducted in four areas of New Zealand: Taumarunui, Marlborough, West Coast 
and Otago. Figure 15  presents the location of focused control and standard control farms. 
• Focused control farms 
� Standard control farms 
Figure 15. Study farm locations within four areas of New Zealand 
Farm selection process 
1 37 
The project was targeted towards cattle and deer herds which either had a high TB prevalence or a 
long history of persistent TB occurrence (Rhodes, 1 997). Prospective participants were also 
selected for their willingness to take pro-active action. Farmers were screened for their positive 
outlook and their willingness to take control and accept responsibility. 
The focused control farms were divided into two groups: 'high' and 'low' incidence, whereby 
herds with more than five TB animals in any infection period (which might have extended over 
several years) were classified as 'high' incidence herds. 
Under the original project plan standard control farms were to be selected prospectively at the start 
of the project. However, this was not achieved. Therefore, the standard control farms were selected 
retrospectively at the end of the project, in early 2000. The criteria to match the farms were area, 
herd type and a similar herd history for the time period 1993 to 1 996. These criteria were given to a 
1 38 
MAF officer, who had access to all national TB testing data. MAF then selected two farms for each 
of the focused control farms, achieving as good a match as possible. In a second step, the herd 
histories were checked to determine if they would fit in the respective categories 'high' and ' low' . 
Four farms were excluded for not having enough TB animals. In the third step the veterinary 
officers in the various study regions were approached with the list of farms to attain their opinion 
on the suitability of the farms. On this basis some farms were additionally excluded, as they were 
disbanded or had new owners or were otherwise not a good match to the focused control farm. In 
the fourth step the veterinary officers were asked to select replacement suitable farms for each 
excluded standard control farm originally selected. It was hoped that the veterinary officers could 
select one additional farm for each focused control farm, to allow for drop-outs. However, this was 
not possible as not enough TB farms were available in most regions. Therefore only 68 standard 
control farms could be found, instead of the desired 70. 
After all standard control farms had been selected, contact was made with all the farmers, both 
focused control and standard control. By phone the nature of the study, the nature of the 
comparison between the two projects (national and Wairarapa) and the nature of the questionnaire 
was explained to the farmers . The questionnaire (see Appendix IV, p. 339) was developed in 
conjunction with the leaders of the national one-on-one project and also used with the farmers in 
the Wairarapa project. It was left to the farmers to decide if they wanted to answer the questions on 
the phone or by mail. In both cases the questionnaire was mailed out to the farmers .  If no reply had 
been obtained within three weeks, another phone call was made as a reminder, often resulting in 
sending out the questionnaire a second time. Intensive phone follow-up was undertaken to achieve 
maximum compliance. Once the questionnaires were returned, the answers were entered into the 
database and if questions had been left unanswered, the farmer was phoned to clarify these. 
Methods employed 
A multi-skilled team was chosen for each farm, including the farmer, a farmer mentor, the 
AgriQuality veterinarian, the private veterinarian servicing the farm, a pest control specialist and a 
farm management consultant (Rhodes, 1 997). In total five farmer mentors were selected, one for 
each region, except Otago with two. These mentors were mainly facilitators between the 
representatives of the different organisations. 
For each farm this team set up a management plan. The team planned to have an annual review of 
each property, plus additional three follow-up visits by the consultant to review the plan and 
progress in  completing it (Rhodes, 1 997). In the first management meeting the aims and goals of 
the farmer were elucidated and the possibility of TB management and control assessed. The 
meeting resulted in a detailed management plan, which the farmer tried to implement. The plan was 
1 39 
reviewed within the group several times during the project, sometimes up to six times, in order to 
revise and adapt it to changed circumstances. 
TB control involved pest management strategies, such as possum and ferret control using traps and 
bait stations and TB management strategies, such as frequent testing of livestock. On some farms 
feral animal surveys were conducted, on others the farmers were trained to recognise tuberculous 
lesions when conducting an autopsy of feral animals .  In some areas, especially in the West Coast 
region, the Regional Council conducted major vector control. 
Cattle and deer TB data and confirmation of TB status 
Cattle and deer TB testing data, dates on vector control and the confirmation of TB status in cattle 
and deer were as described in Chapter 3 ,  for the Wairarapa project. 
All cattle and deer are subject to annual TB testing. Any cattle/deer that reacted positive to this test 
was considered TB positive (Animal Health Board, 1 998) unless the animal was serial tested 
negative thereafter with an ancillary test. If the animal did not get tested with an ancillary test, but 
went to slaughter, it was considered TB positive, irrespectively of whether visible lesions at 
slaughter were found or not. The TB status of cull animals, that went to s laughter without being 
tested first and in which lesions, indicative of TB, were found at slaughter, was determined 
according to the results of the histological examination. Any positive animal was termed 'TB 
animal' . 
A cattle/deer herd was kept under Movement Control restrictions ( Infected TB status), until it had 
two clear whole herd tests with a minimum of six months in between (when the herd achieved 
Clear TB status). 
Analysis of data 
The project was actively conducted from mid 1995 to mid 1997 on the farms. For the analysis a full 
three-year period was required in order to compare the study with the contemporary Wairarapa 
study. Therefore tuberculosis data was used from 1993 until the end of 1 998. The years 1993 and 
1994 were used as a pre-project period and the years 1 997 and 1 998 as the final project period. 
Information on cattle and deer TB herd testing data and dates for possum control operations 
conducted by the Regional Council were obtained from the National Livestock Database (NLDB), 
which is held by AgriQuality New Zealand (the state veterinary service) (Ryan, 1 997). 
140 
The database management software Microsoft Access 97 (Microsoft Corporation, Redmond, W A) 
and spreadsheet software Microsoft Excel 97 (Microsoft Corporation, Redmond, W A) were used to 
store and manipulate the data. Statistical analyses were conducted using NCSS 2000 (Number 
Cruncher Statistical Systems, Kaysville, Utah, U.S .A.), and SPSS for Windows version 9.0. 1 
(SPSS Inc. Headquarters, Chicago, Illinois, U.S.A.). 
Results 
Vector control conducted on the project farms by the Regional Council 
In total 38  initial vector control programmes were conducted on focused control farms and 76 on 
standard control farms. The percentage of focused and standard control farms being under RC 
vector control programmes from 1984 to 1998 is shown in Figure 16 .  The graph assumes that each 
control operation lasted four years. 
U 
It: ... Q) 'C t: � 
rJ) '0 E ... .. ... t: .!!! 0 .... u 0 .. t: Q) u ... Q) 
Q. 
60 
50 
40 
30 
20 
1 0  
0 fji\ nil rn rra � a-<.0 co 0> co co 0> o 0'> 0'> 
1- -
- - -
- -
- - - -
Years 
I--
<.0 0> 0> 
-
-
I--
-
-
co 0'> 0'> 
-
I--
-
-
o Focused control 
El Standard control 
Figure 16. Percent of farms under RC vector control, assuming that a control operation lasts 
for four years. 
Prior to the start of the project 30 initial vector control programmes were performed on focused 
control fanus and 52 on standard control farms. From the beginning of the project to the end of 
1 998 seven focused control fanus and 1 8  standard control farms received vector control by the 
Regional Councils. Overall the distribution of RC vector control operations was similar for focused 
and standard control farms. 
141  
Analysis of tuberculosis testing records 
For all available farms (35 focused control and 68 standard control farms) TB data was obtained. 
Time spent on Movement Control and herd TB status at the end of the project 
All 35 focused control and 68 standard control farms were on Movement Control at the beginning 
of the project in mid 1995 .  For the three years project period ( 1 996 to 1998) a focused control farm 
spent on average 22.44 months under Movement Control restrictions, while a standard control farm 
spent 20.66 months under Movement Control (median focused control farms 23.5 months, median 
standard control farms 2 1 .0 months; Mann-Whitney U test 2=0.66, p=0.51) .  Figure 17 presents the 
frequency histogram for the time spent on MC for focused and standard control farms. 
35 
30 
(I) E 25 "-ca -- 20 0 -c 1 5  G) (,) .. 1 0  G) a.. 
5 
0 
5 1 0  1 5  20 25 
Months 
30 35 40 
[]Focused control 
• Standard control 
Figure 17. Time (in months) spent on Movement control by focused and standard control 
farms between 1996 and 1998. 
By the end of the project, in December 1998, a higher percentage of focused than standard control 
farms had infected herds (Table 10), however this difference was not statistically significant 
Cl=1 .41,  p=0.236). 
142 
Table 10. Number of focused control and standard control farms by TB status at the end of 
1998 - in brackets the percentage of total farms 
TB status Focused control farms Standard control farms 
Clear Status (Off MC) 1 5  (42.9 %) 37 (54.4 %) 
Infected Status 20 (57. 1 %) 30 (44.1 %) 
Ceased farming cattle 1 (1 .5%) 
Total 35 68 
From 11111996 until the end of 1998 23 (65.7 %) focused control farms and 44 (64.7 %) standard 
control farms came off MC (x2=O.OI,  p=O.919). Eight focused control farms and six standard 
control farms subsequently became infected again and were put back on MC restrictions. One of 
the standard control farms that came off MC ceased to farm cattle in the final year of the study. 
Table 1 1  presents the number of farms on and off Movement Control at the end of 1998 for each 
herd type, comparing focused control versus standard control farms. For deer and beef finishing 
farms the focused control group achieved the same results as the standard control, whereas the 
results in the focused control beef breeding and dairy herd were less successful than in the standard 
control group. However, the number of farms in the beef finishing and the dairy group were smaller 
than for the other two herd types. 
Table 1 1. comparing TB status at end of 1998 within herd type, in brackets percentage of 
farms in that herd type group. 
Herd type Focused control farms Standard control farms 
Deer 3 Clear (50 %) 5 Clear (50 %) 
3 infected (50 %) 5 Infected (50 %) 
BB 1 1  Clear (45.8 %) 29 Clear (59.2 %) 
1 3  Infected (54.2 %) 20 Infected (40.8 %) 
BD 1 Infected (1 00 %) 1 Infected ( 1 00 %) 
OH 1 Clear (25 %) 4 Clear (50 %) 
3 Infected (75 %) 4 Infected (50 %) 
Number  of TB animals 
Due to the strong variation in the annual pattern of TB cattle and deer numbers, depending on 
animal numbers present on farm and on test dates, small changes in these factors can result in 
143 
misleading patterns. Thus the present analysis used two-year blocks, representing the periods prior 
to the study, initial study period and late study period. 
Figure 1 8  and Table 1 2  present the proportion of focused and standard control farms that had at 
least one, two and three or more TB animals in any one year of the pairs of 1 993/94, 1996/97 and 
1 997/98. Although a similar percentage of farms still had TB animals in the final year ( 1 998), the 
graphs show that the focused control farms had a more rapid decline in TB animal numbers than 
the standard control farms. There were no significant differences in the number of TB animals 
between focused and standard control farms (two or more TB animals: l=0.22, p=O .64 ; three or 
more TB animals: X2=O.O I ,  p=O. 91 ). 
Focused control farms Standard control farms 
100 
III 80 
E ... Number of .::! 60 TB animals .... 
0 
- Cl >=1 t: 40 GI 
u >=2 ... 
GI 
0.. 20 0 >=3 
0 
93/94 96/97 97/98 93/94 96/97 97/98 
year year 
Figure 18. Proportion of farms with one, two, three or more TB animals in the years 1993/94; 
95/96 and 98/99 for focused control and standard control farms. 
Table 12. Number of focused and standard control farms with one, two or more TB animals 
in any one of the two years. 
Focused control farms Standard control farms 
Years � 1 TB � 2 TB � 3  TB n farm � 1 TB � 2 TB � 3 TB n farm 
animal animals animals cattle animal animals animals cattle 
93/94 32 30 29 35 59 50 43 66 
96/97 26 17  15  35 51 42 35 68 
97/98 23 1 5  14  35 49 32 26 67 
Cumulative TB i ncidence and its reduction over three years 
The cumulative incidence was calculated by dividing the total number of TB animals by the 
average number of animals tested in whole-herd tests during the interval. Table 1 3  shows that the 
144 
average pre-project cumulative incidence was higher for the focused control than the standard 
control farms. This difference was significant at the 0.05 level (Mann-Whitney U test, Z=2.44, 
p=0.0 146). The focused control farms achieved on average a reduction of 62% in the two-year 
cumulative TB incidence, whereas the standard control farms only achieved a reduction of 29%.  A 
higher proportion of focused than standard control farms achieved 1 00% reduction in the two-year 
cumulative incidence (n=17,  49% of focused control farms; n=22, 33% of standard control farms, 
X2=2.4 1 ,  p=0. 1205) (Figure 1 9) .  Overall the difference in the reduction between focused and 
standard control farms was not statistically significant (Mann-Whitney U test, Z=1 .70, p=0.090). 
able 13. Average two-year cumulative TB incidence (Cum inc.) for national focused and 
standard control farms 
Focused control farms 
Standard control farms 
50 
III 40 E .... ca - 30 -0 -c::: 20 Cl) u .... Cl) Q.. 1 0  
0 
Cum inc. 93/94 Cum ine. 97/98 
0.0457 
0.031 8 
0.01 75 
0.0224 
o 1 0 20 30 40 50 60 70 80 90 1 00 
% reduction in cum inc. 
Reduction 
61 .8 % 
28.6 % 
[J Focused control 
• Standard control 
Figure 19. Reduction in cumulative TB incidence between 1993/94 and 1997/98 for focused 
and standard control farms. 
The two-year cumulative incidence was also compared between the different herd types of the two 
farm groups (Figure 20) . As there was only one beef finishing farm in each farm group, these farms 
were combined with the beef breeding farms to form a combined group called 'beef farms' .  Table 
14 gives the cumulative incidence for the period 1993/94 and for 1 997/98, stratified by herd type 
and farm group and the reduction achieved in each group. The focused control farm group had 
greater reduction in all herd types than the standard control group. However, the focused control 
145 
group farms had a greater cumulative TB incidence to start with than the standard control farms. 
Especially in the deer farms it is noticeable that the focused control farms had a very high reduction 
in cumulative incidence whereas the standard control fanns had a five-times higher cumulative 
incidence in 1 997/98 than in 1993/94. 
0.08 
G.I 0.07 
u i 0.06 
"'0 
u 0.05 .5 
� 0.04 
E 0 .03 
::s u 
:r... 0.02 
>,0 
C'IoI 0.01 
o 
!-
-
!-
!- - -
!- - -
:�r:-
-
'-
Deer Beef Dairy 
93194 93194 93194 
- -
r-
r-- !-- r-
'- '-- '-
Deer Beef Dairy 
97198 97198 97198 
[] Focused control 
• Standard control 
Figure 20. Cumulative TB incidence in different herd types of focused and standard control 
farms for 1993/94 and 1997198. 
Table 14. Average two-year cumulative incidence (Cum inc.) of TB in focused control and 
standard control farms, stratified for herd type and reduction in cumulative incidence 
achieved. 
Herd type (number Cum inc. 93/94 Cum inc. 97/98 Reduction 
of farms) 
Focused control farms Deer (6) 0.07004 0.02698 61 .5 % 
Beef (25) 0.041 0.01 372 66.5 % 
DH (4) 0.04478 0.0265 40.8 % 
Standard control farms Deer (1 0) 0.00991 0.04953 Increase 
Beef (49) 0.03565 0.01 671 53. 1 % 
DH (8) 0.03581 0.02309 35.5 % 
The two-year cumulative TB incidence was also compared between the four regions where the 
project was performed. Table 1 5  presents the reduction in two-year cumulative incidence between 
focused and standard control fanns and regions. All farm groups in the four regions, except the 
146 
standard control farms in Otago, achieved a reduction in cumulative incidence from 1993/94 to 
1997/98. In Marlborough and West Coast the focused control farms achieved a slightly higher 
reduction than the standard control farms. In Taumarunui the standard control farms achieved a 
better result than the focused control farms. 
Table 15. Comparison of two-year cumulative TB incidence (Cum inc.) and reduction 
between focused and standard control farms stratified by regions (the number of farms is 
shown in brackets). 
Region Cum inc. 93/94 Cum inc. 97/98 Reduction 
Focused control Marlborough 0.0372 (n=8) 0.0075 (n=8) 79.84 % 
Otago 0.0475 (n=9) 0.01 67 (n=1 0) 64.84 % 
Taumarunui 0.0486 (n=9) 0 .0 130 (n=9) 73.25 % 
West Coast 0.0490 (n=8) 0.0335 (n=8) 31 .63 % 
Standard control Marlborough 0.0249 (n=15) 0.0062 (n=16) 75. 1 0  % 
Otago 0.01 31 (n=17) 0.0287 (n=1 8) Increase 
Taumarunui 0.0286 (n=18) 0.0063 (n=1 8) 77.97 % 
West Coast 0.0616 (n=1 6) 0 .0481 (n=1 6) 21 .92 % 
Table 16 gives the two-year cumulative TB incidence for the years 1993/94 and 1997/98 for 
focused and standard control farms stratified on whether they had received vector control 
programmes by the Regional Council before the start of the project. The focused control farms 
achieved better results than the standard control farms, both farms that received control prior to the 
start of the project and farms that did not receive official vector control prior to the project. 
Table 16. Two-year cumulative TB incidence (Cum inc.) on focused and standard control 
farms stratified on whether they had received vector control prior to the start of the project 
mid 1995 (the number of farms is given in brackets). 
RC Control Cum inc. 1993/94 Cum inc. 1997/98 Reduction 
Focused control farms Yes 0.0467 (n=27) 0.0 197 (n=28) 57.8 % 
No 0.0420 (n=7) 0.0085 (n=7) 79.8 % 
Standard control farms Yes 0.0379 (n=51 )  0.0231 (n=51 )  39.1 % 
No 0.01 1 0  (n= 15) 0.01 20 (n=16) Increase 
147 
Comparison of the national project with the Wairarapa project 
Comparing on ly cattle farms 
In a fIrst step only the cattle farms which took part in the national project were compared with the 
cattle farms studied in the Wairarapa project (see Chapter 3) .  Table 17 presents the number of 
Wairarapa and national project cattle farms and their TB status at the end of the proj ects. A 
statistically signifIcantly greater percentage of Wairarapa focused control farms than national 
focused control farms were off Movement Control (X2=9. 1O, p=O.OO26), indicating that the 
Wairarapa intervention study resulted in more cattle farms coming off Movement control 
restrictions than the national project. A slightly greater percentage of Wairarapa standard control 
farms than National standard control farms were off MC, however, this difference was not 
statistically signifIcant (X2= 1 .07, p=0.3000) .  
Table 17. TB status of Wairarapa and national cattle study farms at the end of the projects. 
Wairarapa National Wairarapa National 
focused control focused control standard control standard control 
Clear Status 26 (78.8%) 1 2 (41 .4%) 1 9  (67.9%) 32 (56. 1%) 
Infected Status 7 (21 .2%) 1 7  (58.6%) 9 (32.1%) 25 (43.9%) 
Total number 33 29 28 57 
In the fInal year of the projects a lower percentage of Wairarapa than national focused control cattle 
farms had two or more TB animals (l=4.76, p=O.029 1) .  The Wairarapa standard control farms and 
the national standard control farms were similar (X2= l .05, p=0.3049 for two or more TB animals). 
Table 18 presents the number of cattle farms having at least one, two or three TB animals in the 
fInal study year. 
Table 18. Number of Wairarapa and national study cattle farms with one, two, three or more 
TB animals in the final study year. 
TB animal Wairarapa National Wairarapa National 
number in final focused control focused control standard control standard control 
project year 
� 1  7 (21 .2%) 1 2  (41 .4%) 8 (28.6%) 24 (42.1 %) 
� 2  3 (9.1%) 9 (31 .0%) 5 (1 7.9%) 1 6  (28. 1 %) 
� 3  1 (3.0%) 7 (24.1 %) 3 (1 0.7%) 14 (24.6%) 
Total no. of farms 33 29 28 57 
148 
In order to allow comparison between the two projects, the time periods 1993/94 and 1997/98 were 
chosen for the national study and 1995/96 and 1998/99 for the Wairarapa project. This way, the 
two years prior to the commencement of any work on the farms were taken as a reference point for 
the last two years in each project. 
Table 19 presents the two-year cumulative TB incidence for focused and standard control farms in 
the national and the Wairarapa project and the reduction in incidence achieved by the four farm 
groups. The Wairarapa focused control farms achieved a much higher reduction in cumulative TB 
incidence than the national focused control farms, whereas the standard control farms in both 
projects achieved a similar reduction in incidence of about 50%. 
Table 19. Comparison of two-year cumulative TB incidences in the years prior to the 
commencement of the intervention studies and the last two years of the projects of Wairarapa 
and national study farms (cattle farms only). 
Cum inc. 1993/94 Cum inc. 1997/98 Reduction 
National Focused control 0 .0425 (n=28) 0.0160 (n=28) 62.4 % 
National Standard control 0.0351 (n=53) 0.0173 (n=54) 50.7% 
Cum inc. 1995/96 Cum inc. 1998/99 Reduction 
Wairarapa Focused control 0.0629 (n=34) 0.0034 (n=33) 94.6 % 
Wairarapa Standard control 0.0354 (n=33) 0.01 61 (n=28) 54.5 % 
The reduction in cumulative incidence on each individual cattle farm was also compared between 
farms of the national and the Wairarapa project. If a herd had an increase in cumulative incidence, 
the reduction was set to zero. In Figure 21 the reduction in cumulative TB incidence is presented 
for the two Wairarapa and the two national programme farm groups. 
149 
70 
60 I-
VI 
E 50 11... -
J! 
.... 40 0 
.... c 30 " Co) 11... 20 " 
D.. 
1 0  
0 n r. n rl n rI rI • .  I . n 1 1 � 1 1 
o 20 40 60 80 100 o 20 40 60 80 1 00 
Reduction in cum. TB inc. Reduction in cum. TB inc. 
ID Waif focused _ Nat focused I ID Waif standard _ Nat standard I 
Figure 21. Comparison of reduction in cumulative TB incidence achieved on individual cattle 
farms of Wairarapa and national focused and standard control farms. 
Figure 2 1  shows that none of the Wairarapa focused control farms had a reduction of less than 
17%, whereas five of the national focused control farms had an increase in cumulative incidence 
(reduction=O in the graph). Furthermore, a greater percentage of Wairarapa than national focused 
control farms had 100% reduction. However, this difference was not statistically significant (Mann­
Whitney V-test z=-1 ,637, p::O. 102, mean rank National 27.76, n=29; mean rank Wairarapa 34.79, 
n=33).  
Figure 21  also shows that a higher percentage of  national standard control farms than Wairarapa 
standard control farms had an increase in cumulative incidence, and a lower percentage of national 
standard control farms had 100% reduction. This difference was significant (Mann-Whitney U test 
z=-2.496, p=O.013 ,  mean rank. Wairarapa 52. 1 6, n=28; mean rank National 38.50, n=57) ,  indicating 
that the Wairarapa standard control farms achieved a higher reduction in cumulative incidence than 
the national standard control farms. 
Comparing all farms 
Table 20 presents the TB status of Wairarapa and national study farms at the end of the projects. A 
statistically significant higher percentage of Wairarapa focused control farms had a Clear TB status 
than national focused control farms (X2=9. 16,  p=0.0025), indicating that the Wairarapa intervention 
study got more farms off Movement Control restrictions than the national focused control study. 
No statistically significant difference was found between the Wairarapa and national standard 
control farms (X2= 1 . 30, p=0.2538). 
1 50 
Table 20. TB status of Wairarapa and national study farms at the end of the projects. 
Wairarapa National Wairarapa National 
focused control focused control standard control standard control 
Clear Status 26 (78.8%) 1 5  (42.9%) 1 9  (67.9%) 37 (55.2%) 
Infected Status 7 (21 .2%) 20 (57. 1 %) 9 (32. 1 %) 30 (44.8%) 
Total number 33 35 28 67 
In the fmal year of the projects a lower percentage of Wairarapa than national focused control cattle 
farms had two or more TB animals ci=5. 1 8, p=D.0228). No statistically significant difference 
between the Wairarapa and national standard control farms was found (X2= 1 .8 1 ,  p=0. 1 789 for two 
or more TB animals). Table 2 1  presents the number of farms having at least one, two or three TB 
animals in the final study year. 
Table 21. Number of Wairarapa and national study farms with at least one, two or three TB 
animals in the fmal study year. 
TB animal Wairarapa National Wairarapa National 
number in fmal focused control focused control standard control standard control 
project year 
� 1  7 (21 .2%) 1 5  (42.9%) 8 (28.6%) 30 (44.8%) 
� 2  3 (9. 1 %) 1 1  (31 .4%) 5 (1 7.9%) 21 (31 .3%) 
� 3  1 (3.0%) 9 (25.7%) 3 (1 0.7%) 1 9  (28.4%) 
Total no. of farms 33 35 28 67 
The reduction in cumulative incidence on each individual farms was also compared between farms 
of the national focused control and the Wairarapa focused control farm group. If a herd had an 
increase in cumulative incidence, the reduction was set to zero. The Mann-Whitney V-test 
indicated a significant difference in reduction between the two project farms (z=-2.096, p=0.036), 
indicating that the Wairarapa focused control farms (n=33) achieved a higher reduction than the 
national focused control farms (n=35) (mean rank Wairarapa: 39.39, mean rank National: 29.89). 
Also the Wairarapa standard control farms (n=28) achieved higher reductions than the national 
standard control farms (n=67) (Mann-Whitney V test z=-3.088, p=D.OO2, mean rank Wairarapa 
6 1 .05, mean rank National 42.54). Table 22 presents the overall cumulative incidence prior to the 
projects and at the end of the projects for the national and Wairarapa focused and standard control 
farms. 
1 5 1  
Table 22. Comparison of two-year cumulative TB incidences (Cum inc.) in the years prior to 
the commencement of the intervention studies and the last two years of the projects of 
Wairarapa and national study farms (all farms). 
Cum inc. 1993/94 Cum inc. 1997/98 Reduction 
National Focused control 0.0457 (n=35) 0.0175 (n=35) 61 .8 % 
National Standard control 0.0318 (n=66) 0.0224 (n=67) 28.6% 
Cum inc. 1995/96 Cum inc. 1998/99 Reduction 
Wairarapa Focused control 0.0629 (n=34) 0.0034 (n=33) 94. 6 % 
Wairarapa Standard control 0.0354 (n=33) 0.0 161 (n=28) 54.5 % 
Discussion 
Evaluation of the national one-on-one project 
This project was, together with the Wairarapa project (Chapter 3), the fIrst intensive intervention 
study, evaluating on farm advice activities on vector and livestock control to reduce TB in 
livestock. No statistically signifIcant results were expected with the limited number of farms used 
in the project. Nevertheless, the evaluation provided valuable evidence that on-farm advice, 
followed by on-farm control measures in the form of vector control and grazing management, can 
reduce TB incidence in livestock. 
The farms selected in the national one-on-one study were not a representative sample of the farms 
in New Zealand; the farmers were selected specifIcally for their motivation and outlook towards 
more effective TB control. This might make interpretation of the results in relation to the wider 
population of infected herds problematical, but the project was only intended as a pilot study and 
no statistically signifIcant results could be expected with only 35 focused control farms, It was 
considered more important to have pro-active farmers in the group, rather than spending time and 
money on farms where the farmer was not interested in applying the management plan (Rhodes, 
1997). The study was totally reliant on farmers complying with the management plan set up within 
the team. Two of the objectives of the study were to show if on-farm control methods could reduce 
the within-herd TB prevalence or reduce the time spent on MC. Therefore it was considered 
important by those responsible for managing the project to have participants who would comply 
with the management plan and apply it, to show if such on-farm control methods can have an effect 
on TB. 
152 
Although a random sample of fanners could have indicated the percentage of fanners complying 
with such on-fann control programmes, the costs involved would have been too high for a fann not 
applying the methods. Including more than 35 farms in the study was not possible due to time 
constraints of the individual team members. All members had regular occupations and time was 
limited. 
It would have been more advantageous if the selection of the standard control farms had taken 
place at the time when the focused control farms had been selected, as originally planned. In 
selecting standard control farms only three years later, potential selection biases could not be 
excluded. Such biases are well recognised problems of retrospective selection of controls and are 
difficult to prevent although they can be minimised. All possible effort was made to avoid such 
biases when selecting the standard control farms. The use of two standard control farms per 
focused control farm applied to this situation a recognised method of reducing uncontrollable 
selection bias in case control studies (Kleinbaum et aI. , 1982), because the standard control farms 
had to be selected retrospectively, much as would be done for controls in a case control study. 
However, it was very difficult to find standard control farms with high TB incidence. 
The multi-disciplinary approach taken in this study is increasingly used in agriCUltural extension 
(paine, 1993; McRae et al. , 1993), providing each team member with insights from different fields. 
Furthermore, farmers were only advised on management changes and practices, but the 
implementation of these control measures was left to fanners themselves, thus resulting in a very 
high participatory approach. The participation of fanners in planning and implementing changes is 
advocated by many authors, such as Chambers et al. ( 1989); Roling and Engel ( 1991) ;  Pyke and 
lohnstone (2000);  Verkerk et al. (2000). 
The evaluation of the project found differences in the cumulative incidence and its reduction 
between focused and standard control farms. The pre-project cumulative TB incidence ( 1993/94) 
was significantly higher for focused control farms than for standard control fanns, indicating that 
bias existed between focused control and standard control farms. This bias could therefore mean 
that the focused control farms had a higher potential to show reduction in cumulative TB incidence 
than the standard control farms, as it might be easier for farms with several TB animals to reduce 
this number than farms with only one TB animal. However, on the other hand, it is also possible 
that the focused control farms were the more difficult farms than the standard control farms. 
Over the three years of the project evaluation, the focused control farms achieved a reduction in 
cumulative incidence of 62%, while the standard control farms achieved 29% reduction. This result 
could be in part a result of the bias. However, although the difference was not significant, a higher 
proportion of focused control farms than standard control farms achieved 100% reduction in 
153 
cumulative incidence (49% versus 33%). This result cannot be explained by the higher pre-project 
cumulative incidence. 
It was noticeable that the overall high reduction in cumulative incidence in the focused control 
farms was mainly due to the deer herds. The focused control deer herd achieved a reduction of over 
60% ,  whereas the standard control deer herds experienced an increase in cumulative incidence. The 
beef and dairy focused control farms only achieved a slightly higher reduction than the standard 
control farms (67% versus 53% reduction for the beef farms, and 4 1  % versus 36% for the dairy 
farms). Therefore it is concluded that the project was particularly successful for the deer farms 
involved in the project. 
In three of the four study areas the focused control farms achieved a higher reduction in cumulative 
incidence than the standard control farms, whereas in one area (Taumarunui) the standard control 
farms were slightly more successful. 
The project achieved a higher reduction in cumulative TB incidence especially in areas, where no 
RC vector control had been conducted. In these areas the focused control farms achieved a 
reduction of 80%, while the standard control farms had no reduction at all, but an increase in 
cumulative incidence. However, the project also achieved a higher reduction in cumulative 
incidence in areas where vector control work had been conducted by the RC (58% in focused 
control farms versus 39% in standard control farms). 
There were no differences in the time and the frequency of vector control programmes conducted 
by the Regional Councils between focused and standard control farms . Therefore, the differences 
found relating to TB in livestock, cannot be attributed to these official control programmes, but to 
the interventions conducted on the focused control farms . 
TB parameters measured in focused and standard control farms that were similar in the two farm 
groups were: the time spent on MC, the number of farms coming off MC during the project period, 
and the number of TB animals per farm at the end of the project. However, the statistically 
significant difference in the pre-project cumulative TB incidence between focused control and 
standard control farms could also have had an influence on these parameters. Fifty percent of deer 
farms (focused control as well as standard control farms) were still on MC at the end of the project, 
whereas the standard control beef and dairy farms achieved a higher percentage of farms off MC 
than the focused control farms. If the standard control farms were the less severely affected farms, 
then it  would have been easier for them to clear their infection entirely than for the focused control 
farms . 
1 54 
In judging the effect of the on-farm control measures employed in this project, one also has to take 
into account that the project was evaluated after only three and a half years. It was expected that the 
control measures would take some time to have an effect on TB outcomes measures, especially as 
animals are mostly tested only annually or biannually and therefore there is a considerable lag 
before improvements in TB control show up in TB testing results. The sensitivity of the skin test is 
not perfect (Ryan et al. , 1 99 1 ), resulting in incorrectly classified 'test negative' animals, causing a 
gradual decline in TB in livestock. 
Overall the results indicate an advantage over the standard control farms, indicating that the 
interventions implemented provided additional benefits in reducing TB incidence in livestock on 
farms with persistent problems, independent of official vector control programmes conducted by 
the Regional Councils. 
Comparison between national and Wairarapa project 
In order to ensure direct comparison between the national and the Wairarapa project, the 
effectiveness of the two projects was first evaluated using only cattle farms. The Wairarapa project 
achieved a significantly higher proportion of cattle farms coming off Movement Control than the 
national project, indicating a higher success in the Wairarapa project. This is also supported by a 
lower percentage of farms in the Wairarapa having multiple TB animals, and a higher reduction in 
cumulative incidence achieved on Wairarapa focused control farms. Using all farms, including the 
deer farms of the national project, these differences were even more pronounced. 
As there were no statistically significant differences between the performances on standard control 
cattle farms in the two projects, it can be concluded that the significant differences in the two 
focused control farm groups was attributable to the interventions conducted. Thus it can be 
concluded that the interventions used in the Wairarapa project, using a small team with hands-on 
practice, was more successful than using a large team with an advisory role only. Furthermore, the 
results obtained in the Wairarapa project are more reliable, as the farms were selected randomly, 
therefore providing valid representativeness. In the national project only the focused control farms 
were chosen at the start of the project, and the standard control farms had to be selected 
retrospectively some years later, thereby possibly reducing the interpretability of the results. 
Bibliography 
Animal Health Board ( 1995). National TB Strategy: proposed pest management strategy for bovine 
tuberculosis. Animal Health Board, Wellington, New Zealand. 
155 
Animal Health Board ( 1 998). Five Year National Bovine Tuberculosis Pest Management Strategy 
to 30 June 2001 .  July 1998 to June 2001 .  Animal Health Board, Wellington, New Zealand. 
Animal Health Board (2000). Bovine Tuberculosis National Pest Management Strategy 2001-20 1 1 ;  
Towards a TB free NZ; A discussion paper on the future options. Animal Health Board, 
Wellington, New Zealand. 
Chambers, R, Pacey, A. and Thrupp, L.A. (eds.) ( 1989). Farmer First: Farmer Innovation and 
Agricultural Research. The Bootstrap Press, New York, NY. 
Kleinbaum, D. G., Kupper, L.L. and Morgenstern, H. ( 1982). Epidemiologic Research. John Wiley 
& Sons, Inc., New York. 
McRae, A.F., Anderson, F.M. and Brazendale, R ( 1993) .  Applying farmer first research in New 
Zealand. Proceedings of the XVII International Grasslands Congress, I: 637-639. 
Morris, RS., Pfeiffer, D.U. and Jackson, R ( 1994). The epidemiology of Mycobacterium bovis 
infections. Veterinary Microbiology, 40 ( 1 ): 153-177. 
Paine, M.S. ( 1993). Extension agents can perform more effectively through an appreciation of 
individual learning styles. Proceedings of the New Zealand Society of Animal Production, 53: 
1 15- 1 19 .  
Pyke, N.  and Johnstone, J. (2000). Knowledge in the arable industry - responding to need. In: 
Stantiall ID, McDiarmid CM (eds.), Achieving Change Through Improved Knowledge 
Systems. Centre for Agricultural and Veterinary Continuing Education, Massey University, 
Palmerston North, New Zealand, pp. 35-37. 
Ragg, J., Moller, H. and Waldrup, K. ( 1995). University of Otago Wildlife Management Report 
Number 63, a contract report to the Animal Health Board (Project 306/92). University of 
Otago, Duniden, New Zeland. 
Rhodes, T. ( 1997). Report to the Animal Health Board, March 1 997. Agriculture New Zealand, 
Dannevirke, New Zealand. 
Raling, N. and Engel, P. G. H. ( 1991) .  (Chapter12:)  The Development of the Concept of 
Agricultural Knowledge Information Systems (AKIS): Implications for Extension. In: Rivera, 
156 
W. M.  and Gustafson, D. J. (eds.), Agricultural Extension: Worldwide Institutional Evolution 
and Forces for Change. Elsevier, Amsterdam, pp. 125-37. 
Ryan, T.J. ( 1997). The New Zealand National Livestock Database. In: Epidemiology Program. 
Proceedings of the 10th Federation of Asian Veterinary Association (FAVA) Congress. 
Australia, pp. 167-70. 
Ryan, TJ., de Lisle, G. W. and Wood, P. R. ( 1991). The perfonnance of the skin and gamma 
interferon tests for the diagnosis of tuberculosis infection in cattle in New Zealand. In: Or. R. 
Jackson (Convenor), Symposium on Tuberculosis. Foundation for Veterinary Continuing 
Education, Palmerston North, pp. 143-50. 
Verkerk, G., Burke, C., Paine, M. and Jolly, P. (2000). Factors influencing the uptake of controlled 
breeding technologies on dairy fanns. In: Stantiall ID, McDiannid CM (eds.), Achieving 
Change Through Improved Knowledge Systems. Centre for Agricultural and Veterinary 
Continuing Education, Massey University, Palmerston North, New Zealand, pp. 1 83-5 . 
1 57 
Beef farm in the Wairarapa 
Dairy farm in the Wairarapa 
1 58 
Chapter 5 
Attitudes of farmers to bovine tuberculosis control i n  
New Zealand 
1 59 
1 60 
161  
Abstract 
The Australian brushtail possum is the main wildlife vector for bovine tuberculosis in cattle and 
deer in New Zealand. Thus, control and eradication is complicated and cost intensive. There is an 
increased call for farmers to have more responsibility for control measures themselves. For 
promoting the adoption of effective on-farm control measures, it is necessary to know the beliefs 
and attitudes of farmers towards TB and its control. A questionnaire survey was conducted with 
farmers involved in two intervention studies, evaluating the effectiveness of such on-farm control 
measures (Chapters 3 and 4), plus a group of 42 farmers whose herds were classified as TB free. In 
total 205 farmers responded to the questionnaire. The questionnaire sought demographic 
information of the farm, information about farm and stock management, TB and vector control, and 
attitudes towards TB and its control. Farmers were surveyed by personal interviews, phone or mail. 
There were only slight differences in the farm demographic variables between the two farm groups 
within each of the intervention studies, indicating that the random allocation of farms into focused 
and standard control farm group was effective in the Wairarapa study and to a lesser degree in the 
national study. Overall farmers regarded TB eradication as being important, however most farmers 
interviewed in this study were not in favour of stricter Movement Control regulations, removal of 
compensation or having to pay TB testing costs directly. Many farmers saw Government and 
Regional Council as being responsible for eradicating TB and saw no need to conduct control 
programmes themselves. Thus, future eradication strategies should include farmer motivation, 
possible on-farm control strategies and financial incentive packages for farms adopting these 
control measures. 
Introduction 
Control and eradication of bovine tuberculosis (TB) in New Zealand is complicated by the 
existence of wildlife species that are both reservoirs of TB and vectors for the disease to livestock. 
The Australian brushtail possum (Trichosurus vulpecula) is the main vector for the disease in 
livestock (Morris and Pfeiffer, 1995). Current TB eradication strategies have been successful in 
reducing the number of infected cattle and deer herds, as well as reducing the incidence of TB 
within herds (Animal Health Board, 2000). However, these programmes are cost intensive and with 
the decreasing prevalence of TB it may become more difficult to justify spending much of 
taxpayers' money on continuous vector control (Animal Health Board, 2000). It is important that 
continuing effort is made to improve both the effectiveness and the efficiency of TB control in New 
162 
Zealand. This is more likely to be achieved if farmers take greater responsibility for the control of 
TB on their own farms, rather than leaving control as the responsibility of a national body. The 
proposed new strategy for TB management places increasing emphasis on farmers being 
responsible for funding control rather than Government. 
On-farm control programmes provide an effective way for farmers to assist the control of TB (see 
Chapters 3 and 4), however, they have to be promoted effectively. Effective promotion is 
dependent on understanding the existing attitudes and beliefs of farmers so that extension and 
education programmes can be tailored accordingly. The present study involved farmers 
participating in the Wairarapa project (Chapter 3) and the National one-on-one project (Chapter 4), 
each project consisting of a group of 'focused control' and ' standard control' farms. The study had 
three main objectives: (i) to understand the characteristics of the farms on which on-farm TB 
control interventions were conducted within the two projects; (ii) to identify differences between 
the focused and standard control farms in the two projects; and (iii) to identify current attitudes of 
farmers towards TB control. It collected information on general farm management, TB 
management and attitudes towards TB and its control. 
Materials and Methods 
Surveys 
A questionnaire survey was conducted with farmers that had been involved in two projects which 
were designed to evaluate the impact of on-farm control measures on the tuberculosis situation in 
their herds (Wairarapa - Chapter 3 and National project - Chapter 4). The questionnaire was 
developed in conjunction with Chris Carter (AgriQuality) and Tony Rhodes (Agriculture New 
Zealand), the main co-ordinators of the National project. It was designed to explore farmers' 
attitudes and to obtain general information on the two farm groups in each project. The 
questionnaire was extensive, with questions seeking demographic information of the farm, 
information about farm and stock management, TB and vector control, and attitudes towards TB 
and its control (see Appendix IV, p. 339). Details of the groups of farmers involved with each 
project are described below. 
Wairarapa project farms 
For the Wairarapa project, 70 TB-infected farms were selected randomly and half were allocated to 
each of the 'focused control' and the 'standard control' groups (for more details see Chapters 3 and 
163 
4). An on-farm programme, involving targeted possum control and targeted grazing management 
was employed on focused control farms for three years from 1997 to 1999. Standard control farms 
received the standard official TB control. All farmers in the focused and standard control groups 
were surveyed using the questionnaire described above. As the focused control farmers were 
visited on a regular basis, the questionnaire was conducted in two parts with these farmers. The 
first part, asking about farm and stock management, was filled in with the farmers/managers at an 
interview during the third farm visit (1997). The second part of the questionnaire, asking about 
farmer attitudes and effort put into TB control, was mostly conducted during the fourth farm visit 
in early 1998. This was before focused control farmers were responsible for maintaining vector 
control themselves as a part of the intervention project. This way the vector control efforts of 
focused control farmers could be compared with those on standard control farms. The standard 
control farms were surveyed by a combination of personal interviews, phone and mail towards the 
end of the project in early 1999, when the whole questionnaire was completed at one time. Only 
minimal information was obtained for two standard control farms as the farmers had passed away 
during the last year of the project. Information for these two farms was obtained from family 
members, who could not answer all the questions, especially the ones regarding attitudes. 
A group of 70 farmers, whose herds had been clear of TB for at least five years was randomly 
selected in the Wairarapa in 1999 to act as a comparative group for the TB-infected farms. These 
herds are described as 'non-TB' farms. This group was surveyed by mailing a shorter modified 
version of the questionnaire. The questionnaire omitted all questions regarding TB history and its 
control on farms, but included all questions relating to attitudes towards TB. A second letter and 
questionnaire was sent out to forty non-respondents four weeks later. In total 42 of the 70 farmers 
in the non-TB group replied with completed questionnaires. 
Details on the number of farms involved and responding to the questionnaire are given in Table 23. 
National project farms 
For the National project, 35 farms in four TB endemic areas were selected by AgriQuality (then 
Ministry of Agriculture and Fisheries) to implement an on-farm TB management programme. 
These farmers were chosen mainly on the basis of their past TB herd history and partly on positive 
farmer motivation/outlook, which was ascertained in screening interviews held with potential 
participating farmers. For two years focused vector control and management interventions were 
conducted on these 35 'focused control' farms. Retrospectively seventy 'standard control' farms 
were randomly selected to match the area and enterprise type of the focused control farms (for 
more details see Chapter 4). 
1 64  
Both the focused and standard control fanners involved in the National project were surveyed in 
2000, after the project had concluded. Fanners were initially contacted by telephone to explain the 
nature of the study and the nature of the comparison between the Wairarapa and the National 
projects. It was then left to the fanners to decide if they wanted to answer the questions on the 
telephone or by mail. In both cases the questionnaire was mailed out to the fanners. If no reply had 
been obtained within three weeks, another telephone call was made as a reminder, often resulting in 
a second questionnaire being sent. A lot of effort was put into contacting fanners to remind them of 
the questionnaires, sometimes telephoning up to eight times. Once the questionnaires were 
returned, the answers were entered into the database and if questions were left unanswered, the 
farmer was phoned to clarify these. Table 23 presents a summary of the number of fanners 
involved in the intervention project, the number that responded to the questionnaire, and the 
method of data collection for each group. The National project (see Chapter 4) involved 35 focused 
control and 68 standard control fanns. It was not possible to obtain survey information from one 
focused control and three standard control farms. Additionally one focused control and two 
standard control fanns counted for two farms. These farms involved different herd types run at 
separate locations, but managed by the same ownerlherd manager. For the evaluation of the 
questionnaires these farms were considered as only one farm each, thus avoiding duplicated 
answers. This resulted in 33 focused control and 63 standard control fanns being involved in the 
questionnaire survey. 
A national group of non-TB fanns was not selected for comparison within this project, but were 
surveyed and described in a concurrent study (Corner et aI. , 2000) .  
Table 23. Number of farmers responding to the questionnaire in  the Wairarapa and National 
project. 
Number of Wairarapa Wairarapa Wairarapa National National 
farms Focused Standard non-TB Focused Standard 
control control control control 
Involved in intervention study 34 33 35 68 
Responded to questionnaire 34 33a 42 33b 63c 
Non·respondents 28 3 
Information Interview 34 1 0  
collected using Phone 5 20 1 8  
Mail 1 8  42 1 3  45 
a Only minimal information could be obtained on two farms 
b one fann counted for two farms (different herd type and separate fann units) 
C two farms counted for two farms each (different herd type and separate fann units) 
165 
Definition of terms used 
Farm enterprise types were referred to as 'Beef breeding' (BB), 'Beef dry stock' (BD), 'Dairy' 
(DH), and 'Deer' . 
'White-tagged' cattle were cattle identified with an official white Movement Control ear tag. They 
originated from a herd with a TB status of 'Infected' .  
'TB reactor cattle/deer' was any animal that was tested positive in any official TB test (mostly 
caudal fold tuberculin skin test). 
'AHB' is the abbreviation for the Animal Health Board, the national organisation responsible for 
co-ordinating and setting up TB control and eradication schemes. 
Analysis of data 
Responses to the survey for the group of Wairarapa farmers and the National farmers were 
analysed separately, as the methods used to select each group were different. 
Descriptive analysis was used to identify patterns and trends in the data and to compare the 
responses for different groups of farmers.  Open-ended questions were handled by developing a 
number of different themes from the responses to each question. They were coded by assigning 
each response to one or more of the themes, irrespective of the tone of any comments made to this 
question. 
In total there were over 80 variables, with up to five categories within each variable. Screening 
these variables for differences between the categories, would have resulted in numerous individual 
statistical tests, thus severely increasing the likelihood of type I and type n errors. Therefore all .the 
variables were initially screened visually and only variables with a difference between the groups 
were tested statistically. Three main families of variables were created: general farm 
characteristics, TB related issues and attitudes. Within each of the variable families a Bonferroni 
correction term of 20 was set, resulting in application of an alpha of 0.0025 (Ott, 1988). This was a 
conservative method, resulting in a small Type I error probability, but a large Type n error 
probability . 
As many continuous variables showed a skewed distribution, non-parametric tests such as Mann­
Whitney U tests and Kruskal-Wallis tests were used to compare the answer categories between the 
farm groups (Bortz, 1993). For post-hoc tests the critical Z-value was set at 3.29 1 5  where the p­
value was 0.05 divided by 3 times the correction term, accounting for a correction factor of 60. 
166 
Chi-squared tests were used to compare proportions and counts. If in a contingency table any cell 
had an expected cell count of less than five, the exact p-values were calculated (Agresti, 1990). 
Data was stored and manipulated in the database management software Microsoft Access 97 
(Microsoft Corporation, Redmond, W A) and spreadsheet software Microsoft Excel 97 (Microsoft 
Corporation, Redmond, W A). NCSS 2000 (Number Cruncher Statistical Systems, Kaysville, Utah, 
U.S.A.), and SPSS for Windows version 9.0. 1 (SPSS Inc. Headquarters, Chicago, lllinois, U.S.A.) 
were used for statistical analyses. 
Results 
Results for the surveys of the Wairarapa group of farmers and the National group of farmers are 
presented separately. 
Wairarapa farms 
The information for the Focused control farms was collected over a period of two years. For the 
analysis of farm management and characteristics, stock management, TB risk assessment, costs and 
stock management the information collected in 1997 on focused control farms (first part of the 
questionnaire) was used, whereas the information regarding attitudes was collected in 1998 on 
focused control farms. The only information collected twice or three times on focused control 
farms, were stock numbers and amount of control effort, farmers put into vector control. For the 
analysis the information regarding these issues collected on the focused control farms in 1998 (at 
the time of the second part of the questionnaire) was compared with the information collected on 
the standard control farms early 1999. This way, the third year of the project, where the farmers 
were responsible for maintaining the vector control, did not influence the comparison. 
General farm characteristics 
General farm characteristics were compared for focused and standard control farms . Stock numbers 
for the focused, the standard control, and the non-TB farms were taken as of June 1998. 
Farmers were asked for the size of their home farm (both total and effective hectares), the size of 
other owned land, locally leased land and other leased land. Farmers also provided information on 
income and labour units. Table 24 presents some of the characteristics for the three farm groups. 
167 
Table 24. General characteristics of focused control, standard control, and non-TB farms. 
Focused control Standard control Non-TB 
(n=34)a (n=33) (n=42) 
Farmed area (ha) b 759.9 ha 695.3 ha 239.07 
Number of farms using off-farm grazing 1 1  (32.4 %) 1 2  (36.4 %) 27 (64.3%) 
Average self-concept value c 3.1 9  (n=33) 3 . 16 (n=31 )  3.15 (n=37) 
I ncome from cattle (% of total income) 59.79 64.94 71 .56 
I ncome from sheep (% of total income) 36. 15  29.66 26.0 
Income from deer (% of total income) 1 .1 8  4.38 2.44 
Full time labour units on farm 2.1  (n=34) 2.2 (n=32) 1 .5 (n=41 ) 
a gives the sample size for each farm group, unless otherwise stated 
b is the total number of effective hectares on the home property, plus other owned land if less than lOOkm 
away, plus leased land if less than lOOkm away 
C from Seabrook (1984) 
Between 60% and 72% of the total farm income was generated through cattle in all three groups. 
Non-TB farms had the lowest number of labour units, which is consistent with their smaller 
average farm size. 
The distribution of farm size across all three groups is shown in Figure 22. About a quarter of the 
farms (25.6%) were smaller than 100 effective hectares. 
70 
60 
III 
E 50 lo.. 
� 40 .... 0 lo.. 30 41 -'I: 
E 20 ::3 
Z 
1 0  
0 
Effective farm ed area (h a) 
Figure 22. Effective farm size distribution of all farms included in the Wairarapa study. 
On average the non-TB farms were smaller than the focused and standard control farms (239 
versus 728 hectares), however this was not statistically significant when the Bonferroni correction 
168 
was applied (Kruskal-Wallis test 'l=8.20, 2 df, p=0.0 1 7, which is according to the multiple test 
correction applied in this study greater than 0.0025 and therefore not statistically significant). 
Figure 23 gives Violin plots for the effective farmed area on focused control, standard control and 
non-TB farms. 
co :S 
� 0;;; 
E 
� 
(I) > 
U 
� w 
5000 
3333 
1667 
O+-----�-+-L----�+_�--���----� 
Focused 
control 
Standard 
control 
Non-TB 
Figure 23. Violin plots for effective farmed area of Wairarapa focused control, standard 
control and non-TB farms. 
Table 25 presents the average effective farm size for the different enterprise types in the three farm 
groups. Of the non-TB farms a higher percentage of farms were dairy enterprises (54.8% versus 
30% in focused and standard control farms). 
Table 25. Average farmed area stratified by herd type and Wairarapa farm group (with ' 
range in brackets). 
BB 
BD 
DH 
Focused control farms Standard control farms Non-TB farms 
1 373.6 (56 - 4500) n= 1 6  1 1 65.4 (25 - 3644) 
1 92.5 (1 8 - 590) n=8 21 3.0 (78 - 374) 
231 .8 {60 - 1 045) n=1 0 1 85.5 (68 -308) 
n=1 7  392.2 (35 - 1 41 9) n=1 5 
n=6 306.5 (46 - 51 0) n=4 
n=1 0 1 27.5 (31 - 226) n=23 
The farms in the Wairarapa study comprised on average 485 1 livestock units (s.e.=606.58 ; n= 108 ;  
median 2455.9 SU). On average the herds (n=108) comprised 1 990 cattle SU (s.e.= 1 88 . 1 5 ;  median 
1 545.7 SU; see Figure 24), 2794 sheep SU (s.e.=508.5 1 ;  median 80 SU) and 42 deer SU 
(s.e.=2 1 .06; median 0 SU). Only 64 of the 108 farms had sheep and only nine had deer on their 
1 69 
properties. Cattle comprised an average proportion of 0.64 (s.e.=0.035 ; median 0.70) of total 
livestock units on the farms. Figure 25 presents the violin plots of the proportion of cattle for the 
three farm groups. The non-TB farms had the highest proportion of cattle, with a median of 1 .0, 
indicating that no livestock other than cattle were kept on the farm. The differences seen in the 
plots were not statistically significant (Kruskal-Wallis test i= 1 .49, 2 df, p=0.476). The average 
cattle density on the Wairarapa farms was 7.7 SU per effective hectare of land (s.e.=O.59; 
median=5.76; see Figure 26). There was no difference in the cattle density between the three farm 
groups (Kruskal-Wallis test X2=3.43, 2 df, p=0. 1 79). 
30 
25 
1/1 
E 20 ... 4! .... � 1 5  GI � 
; 1 0  
z 
5 
0 0 0 0 0 Cl 0 Cl 0 l.O L,(') l.O l.O N cry 
Cl Cl 0 Cl Cl 0 0 0 Cl Cl 0 Cl Cl 0 0 0 l.O l.O l.O l.O l.O l.O l.O l.O � l.O co r- CD 0) Cl 
Cattle l ivestock units 
Figure 24. Cattle herd size distribution of farms included in the study in livestock units. 
170 
1 .0 
.!!! 'E ::J 
� 0 0 -I/) Q) 
� 0.7 
m -0 --0 
c: 0 :e 
8. 0.3 0 ... 0.. 
Q) 
E en 
U 
0 
Focused Standard Non-TB control control 
Figure 25. Violin plots for the cattle proportion of total livestock units for focused control, 
standard control, and non-TB farms. 
!iJ .J::. 
Q) > 
:u 
� 
� ::::l 
Cl) 
U -
� 'ii) c: Q) "0 
Q) 
E en 
U 
25 
1 6.7 
8.3 
O +---------�--------_,r_--------._--------_, 
Focused 
control 
Standard 
control Non-TB 
Figure 26. Violin plots for cattle density on focused control, standard control, and non-TB 
farms. 
1 7 1  
Herd manager 
The median age of herd managers across all farm groups was in the range 40-50 years (Figure 27). 
The age distributions were bell curve shaped with smaller numbers of farmers being in the 
extremities and the majority of farmers being in the central age groups. 
I/) E ... 
:! .... 0 ... Cl) .Q E ::::I 
Z 
20 
1 5  
1 0  
5 
0 
50-60 
Age group 
>60 
Non-TB 
Standard control 
F ocused control 
Figure 27. Distribution of age groups of herd managers on focused control, standard control, 
and non-TB farms. 
Ninety-eight percent (n=106) of farmers were male, 93 .6% (n=1 02) lived on the farm, 2 1 . 1% 
(n=23) had additional employment, and 83 .5% (n=9 1 )  of fanners stated that their farm income was 
covering their living expenses. Sixty-six percent (n=72) of interviewees owned the fann, 14% 
(n= 1 5) were share milkers and 1 3%  (n=14) were managers. Fifty-six percent of farmers (n=6 1 )  
considered themselves to be the main decision makers for the farm. On average farmers had 
worked 1 5 .9 years (s.e. 1 . 13, n=1 06, median 12 .5) on their current fann. Overall they had worked 
an average 25. 1 years (s.e. 1 .27, n=107, median 24) on a fann. Nineteen percent of herd managers 
(n=20 of 1 07 farmers with information) started their farming job without a farming background, 
whereas 75 .7% (n=8 1 )  were brought up on a fann. Sixty-six percent of herd managers (n=7 1 of 
1 07) had no formal farming-specific qualification. 
Interviewees were also asked to fill in a self-concept form, which was adapted from Seabrook 
( 1984). This part of the questionnaire was filled in by a total of 1 0  1 farmers. Figure 28 shows the 
arithmetic means of personality traits for focused control, standard control, and non-TB farmers. 
1 72 
Not easy going 
Meek 
Patient 
Unsociable 
Not modest 
Persevering 
A worrier 
Cheerful 
Talkative 
Speaks one's m ind 
Difficult to get on with 
Lacking confidence 
Liking change 
Forceful 
Prefers machinery 
Prefers bying new machine 
Dislike to learn 
Still learning 
Likes to avoid hard work 
Dislikes using records 
Values traditional ways 
Dislikes to set targets 
Likes to look after favourite 
animals a bit better than the 
I 
r"" 
Focused 
control 
Standard 
• control - ., Non-TB 
, ''--
.-:';�� 
<:.. / �-.. � ,  �t"" 
..... � ':,...1.', 
.. , .. 
.� -
� .. :!... 
� '> .. " �''''''''' �--l "' '''' 
,�- JJ 
..\,P- -" - .. � .. .... . � ... . � 
.,,,,,,,�,:/-, " 
� .-'�- --... � =----cl �;,.:..,. . 
-.� i"" 
�"r " 
",� 
'4� .. \ -..... !.--' , .v 
Easy going 
Not meek 
Impatient 
Sociable 
Modest 
Giving up easily 
Not a worrier 
Grumpy 
Not talkative 
Keeps quiet 
Easy to get on with 
Confident 
Suspicious of change 
Giving in easily 
Prefers animals 
Prefers choosing new animal 
Very keen to learn 
Very knowledgeable 
Values hard work 
Likes using records 
Likes adopting new ideas 
Likes setting targets 
Likes to strictly monitor 
performance of the herd 
Figure 28. Personality trait means for focused control, standard control, and non-TB farmers 
(adapted from Seabrook, 1984) 
Figure 2 8  suggests that the focused control farmers were more likely to ' speak their mind' and 
'preferred to buy new machinery' than the other two farm groups. The graph also suggests that 
non-TB farmers were less inclined to ' set targets', were more 'impatient' and considered 
themselves more knowledgeable than the focused and standard control farmers ( comparing non-TB 
farms with focused control/standard control farms, Mann-Whitney U test: ' knowledgeabl e '  Z=-
2 . 5808, p=0.0 1 0 ;  set targets : Z=-2 .729, p=O.O 1 9). Comparing all three farm groups with each other, 
the one variable which was most different was ' still l earninglvery knowledgeable' (Kruskal-Wallis 
X2=6 . 69, 2 df, p=O . 0 3 5 ) ,  A higher proportion of non-TB farmers believed that they were very 
knowledgeable and were less likely to consider themselves as ' still learning' compared to focused 
and standard control farmers. 
Stock management 
Table 26 presents the results stock movements on and off the farm for the three farm groups. 
Information collected during the first part of the questionnaire in 1 997 was used for the focused 
control farms. This information could not be obtained for one of the standard control farms. 
1 7 3  
Table 26. Information on stock movements on and off farms for focused control, standard 
control, and non-TB farms. 
Focused control Standard control Non-TB farms 
farms (n=34) farms (n=32) (n=42) 
White Tagged Yes 7 (20.6%) 9 (28. 1%) 1 (2.4%) 
cattle bought No 27 (79.4%) 23 (71 .9%) 41 (97.6%) 
Buying Every year 1 5  (44.1%) 1 1  (34.4%) 9 (21 .4%) 
frequency Now and again 3 (8.8%) 10  (31 .2%) 1 3  (31 .0%) 
Other/nevera 1 6  (47. 1%) 1 1  (34.4%) 20 (47.6%) 
Trading Crucial 1 0  (29.4%) 5 (1 5.6%) 4 (9.5%) 
importance Important 2 (5.9%) 2 (6.2%) 9 (21 .4%) 
Moderately 5 (14.7%) 3 (9.4%) 1 1  (26.2%) 
Minor/unimport. 17  (50.0%) 22 (68.8%) 1 6  (38. 1 %) 
Not stated 2 (4.8%) 
a farms that only bought in single bulls for breeding purposes were classified under 'other' 
There was a difference in the frequency of buying white-tagged cattle amongst the three farm 
groups (x2=9.95, 2 df, p=O.OO7), however after applying the Bonferroni correction term, this was no 
longer significant. The non-TB farms had the lowest frequency of buying in white-tagged cattle. 
There was no significant difference in the buying frequency between the three farm groups 
(X2=8.7 1 ,  4 df, p=O.067). Of the farms that bought in stock, only four focused control and six 
standard control farms took TB and the herd history of the animals into account, whereas 1 3  
focused control and eight standard control farms took production requirements into account. Of the 
non-TB farmers, 1 2  took TB and 1 3  took production requirements into account. 
Ten (29.4%) focused control and eight (25 .8% of 3 1  farms with relevant information) standard 
control farmers indicated that they had changed their stock selling policies due to having TB in 
their herds. Five ( 1 4.7%) focused control and four (of 3 1  farms with relevant information, 1 2.9%) 
standard control farmers also indicated that they had changed their mix of stock classes. Most 
farmers changed from selling weaners to finishing more stock themselves. 
TB risk assessment 
Farmers were asked about the TB situation in their own herd. As part of this they were asked what 
reason they suspected for the TB infection in their herd, and if they knew or suspected areas on 
their farm properties that could be classified as TB hot-spots and if so, whether they grazed cattle in 
174 
these areas or not. To compare the data between focused and standard control farms, the 
information collected at the fIrst part of the questionnaire was used for the TB high risk areas. This 
ensured, that wildlife surveys, which were conducted as a part of the project, did not severely 
influence the outcome, as only a few properties had been subject to wildlife surveys by that time. 
Table 27 presents information on these subjects. 
Most farmers regarded infected wildlife as the source of TB infection in their herds. Neighbouring 
herds and stock movements on and off farms were only mentioned by a minority of farmers. 
A higher proportion of focused control than standard control or non-TB farmers knew or suspected 
high risk areas for TB on their farms. However, this difference was not statistically signifIcant 
(i=3 .96, 1 df, p=O.046). A lower percentage of focused control than standard control farmers 
grazed their cattle in TB high-risk areas (68% vs. 83%). 
Focused control farms had the greatest awareness of the TB situation of their neighbouring herds, 
whereas 2 1  % of the standard control and 41  % of the non-TB farmers did not know whether their 
neighbours were infected or not. 
175 
Table 27. Information on the TB situation and perception of focused control, standard 
control, and non-TB farms. 
Focused control Standard control Non-TB 
(n=34) (n=33) (n=42) 
TB problem Continuous 8 (23.5%) 1 1  (33.3%) 
classification On and off 1 3  (38.2%) 7 (21 .2%) 
Seldom 2 (5.9%) 4 (1 2.1 %) 23 (54.8%) 
No problem anymore 1 1  (32.4%) 1 1  (33.3%) 
Never 18 (42.9%) 
Not stated 1 (2.4%) 
Reason for TB TB feral animals 25 (73.5%) 26 (78.8%) 
i nfectiona Neighbouring herds 5 (1 4.7%) 6 (1 8.2%) 
Stock movements 7 (20.6%) 8 (24.2%) 
High risk areas for Yes 25 (73.5%) 1 8  (54.5%) 22 (52.4%) 
TB suspected on No 8 (23.5%) 1 1  (33.3%) 1 5  (35.7%) farm 
Unknown/not stated 1 (3.0%) 4 ( 1 2. 1%) 5 (1 1 .9%) 
Graze cattle i n  high Yes 1 7  (68.0%) 1 5  (83.3%) 17 (77.3%) 
risk areas No 8 (32.0%) 3 ( 1 6.7%) 5 (22.7%) 
Any neighbouring Yes 20 (58.8%) 22 (66.7%) 1 6  (38. 1 %) 
farms infected with No 10 (29.4%) 4 (12.1%) 9 (21 .4%) TB 
Unknown 4 (1 1 .8%) 7 (21 .2%) 1 7  (40.5%) 
Reactor numbers Increased 2 (5.9%) 0 
over last three Decreased 21 (61 .8%) 23 (69.7%) years 
No change/not stated 1 1  (32.3%) 10 (30.3%) 
a multiple answers were possible 
Vectors and vector control on farms 
Farmers were asked how likely they thought contact was between their livestock and feral animals. 
Table 28 presents the data for possums and ferrets stratified for focused control, standard control 
and non-TB farms. More focused control farmers than standard control and non-TB farmers 
believed there was a high likelihood that livestock had contact with possums or ferrets. However 
this was not statistically significant (X2=3.38, 1 df, p=O.066 for contact with possums, and X2=3.71 ,  
Idf, p=O.054 for contact with ferrets between Wairarapa focused and standard control farms). 
1 76 
Table 28. Assumed likelihood of contact between possums/ferrets and livestock, as indicated 
by focused control, standard control, and non-TB farmers. 
Focused control Standard control Non-TB 
Likelihood of Very l ikely 24 (70.6%) 1 5  (45.5%) 1 4  (33.3%) 
contact with Possible 6 ( 1 7.6%) 1 0  (30.3%) 1 8  (42.9%) possums 
Unlikely 4 ( 1 1 .8) 2 (6. 1 %) 2 (4.8%) 
Unknown / not stated 0 6 ( 1 8.2%) 8 (19.0%) 
Likelihood of Very l ikely 1 7  (50.0%) 8 (24.2%) 1 0  (23.8%) 
contact with Possible 10 (29.4%) 1 3  (39.4%) 1 8  (42.9%) ferrets 
Unlikely 4 (1 1 .8%) 3 (9. 1%) 0 
Unknown / not stated 3 (8.8%) 9 (27.3%) 1 4  (33.3%) 
Farmers were also asked about vector control on their properties, what percentage of land was 
controlled by themselves and/or by the Regional Council and how much time was spent on these 
controls (see Table 29 and Figure 29). For this comparison the data collected in 1998, when the 
second part of the questionnaire was conducted, was used for the focused control farms. During 
this time vector control was performed by the research team, not by the farmers, therefore it did not 
influence the time spent on vector control by farmers themselves. For two standard control farms 
no relevant information could be obtained. 
1 77 
Table 29. Vector control by Regional Council and fanners as stated by focused control, 
standard control, and non-TB farmers. 
Focused control Standard control Non-TB 
(n=34) (n=31) (n=42) 
Avg % of farm 76.5% (n=29) 79.3% (n=22) 67.0% (n=23) 
controlled by Rca 
Avg % of farm 40.4% (n=29) 65.5% (n=20) 83.0% (n=30) 
controlled by farmera 
Regional Council I ncreased 22 (64.7%) 7 (21 .2%) 4 (9.5%) 
control Decreased 2 (8.7%) 1 1  (33.3%) 1 0  (23.8%) 
Farmer Control Increased 1 8  (52.9%) 6 ( 1 8.2%) 2 (4.8%) 
Decreased 3 (1 1 .8%) 8 (24.2%) 1 7  (40.5%) 
Farmer Control = 0  5 (14.7%) 1 0  (32.3%) 1 4  (33.3%) 
(days/year) > 0  29 (85.3%) 20 (64.5%) 26 (61 .9%) 
(median 20, (median 1 7, (median 8.5, 
avg 36.4, avg 22.9, avg 1 5.9, 
range 3 - 1 92) range 1 - 72) range 1 - 1 1 0) 
Not stated 0 1 (3.2%) 2 (4.8%) 
Regional Council = 0  8 (23.5%) 7 (22.6%) 1 9  (45.2%) 
(days/year) > 0  26 (76.5%) 23 (74.2%) 21 (50.0%) 
(median 1 0, (median 1 0, (median 5 ,  
avg 1 3.8, avg 1 4.4, avg 5.4, 
range 3 - 42) range 2 - 50) range 1 - 1 0) 
Not stated 0 1 (3.2%) 2 (4.8%) 
Farmer Control Regularly 7 (20.6%) 5 (1 6. 1%) 9 (21 .4%) 
a according to the farmer 
1 7 8  
0- 70 :l 
0 60 .... 
Cl 
.r:. 50 (,) ca Cl> 
c:: 40 
t/) .... 30 Cl> 
E .... 20 ca --
0 .... 1 0  c:: Cl> 0 (,) ... Cl> 
Q. 
0 0 0 0 0 C\J '<:t <.0 co 0 0 0 C\J 
o Focused control 
11 Standard control 
o Non-TB 
o <.0 o co o o N 
Days spent per year on vector control by farmers 
Figure 29. Time spent per year on vector control by focused control, standard control, and 
non-TB farmers. 
Of the farms that conducted vector control on their properties the focused control farmers had the 
lowest average percentage of their land area being controlled by themselves. Standard control 
farmers indicated that they controlled on average 65% of their farms, and non-TB farmers indicated 
that they on average controlled 83% of their farms, whereas on average the focused control farmers 
estimated that 40% of their farms was controlled by themselves (Kruskal-Wallis test i= 1 4.27, 2 df, 
p=0.OOO8). This might reflect different perceptions of control effort. Figure 30 presents a box and 
whisker plot for the area controlled by focused control, standard control and non-TB farmers. 
c o w 
Focused 
control 
Standard 
control 
Cl 
179 
Figure 30. Box and Whisker plot of the proportion of farm area controlled by focused 
control, standard control, and non-TB farmers. 
The proportion of farmers controlling vectors on their properties themselves was greatest for 
focused control farms with 8 5 %  of farmers controlling possums, yet this difference was not 
statistically significant (i=5 .56, 2 df, p=O.062). Also the time spent on vector control by farmers 
themselves was greater for focused control farmers than for standard control farmers.  However, the 
average was influenced by a few focused control farms which spent a high number of days on 
vector control, whereas the medians for the focused and standard control farms were similar. The 
difference seen in the average time spent across all farms in the three farm groups was not 
statistically significant (Kruskal-Wallis test X2=10.88, 2 df, p=O.OO4). 
Attitudes towards TB and its control 
The relevant information on focused control farms was collected in the second part of the 
questionnaire, which was conducted during the fourth farm visit, early 1 998, about half way 
through the project. For two of the standard control farms it was not possible to obtain information 
on attitudes, as the herd managers had deceased. 
Figure 3 1  presents the importance that was placed on TB eradication by focused control, standard 
control, and non-TB farmers. Thirty (88.2%) of the 34 focused control farmers considered the 
eradication of TB as crucial, two as important, and one each as moderately important or of minor 
importance. Of the 3 1  standard control farmers with relevant information, 23 (74.2%) considered 
TB eradication as crucial, six as important and two as minor. Of the 39 non-TB farms with relevant 
information, 35 (89.7%) considered TB eradication as crucial, three as important, and one as 
1 80 
moderately important. This difference was not significant (Chi-squared test for crucial, combining 
all other categories due to small numbers: X2=3.72, 2 df, p=0. 156). 
Percent of farmers 
0% 20% 40% 60% 80% 1 00% 
Focused control 
Standard control 
Non-TB 
o crucial • important 0 moderately 0 minor 
Figure 31. Importance of TB eradication as considered by Wairarapa focused control, 
standard control and non-TB farmers. 
Most farmers thought TB eradication was possible on their farms (73.5% (n=25) focused control 
farmers, 74.2% (n=23) standard control farmers). Only six focused control farmers and five 
standard control farmers considered TB eradication on their properties impossible due to vegetation 
and habitat factors. Three farmers in each group answered 'unknown' to this question. This 
question was not asked in the questionnaire with the non-TB farmers, as they did not have TB on 
their property. 
When farmers were asked if Movement Control should be made stricter, what the effect of having 
to pay their own TB testing costs directly would be, and what the effect of removal of 
compensation would be, there were slight differences between the three farm groups (Table 30). 
While farmers were also asked for the reasons for their responses, not all farmers gave specific 
reasons to this question. 
1 8 1  
Table 30. Attitudes towards TB control by Wairarapa focused control, standard control, and 
non-TB farmers (percentages of farm group in brackets). 
Focused control Standard control Non-TB 
(n=34) (n=31) (n=42) 
Movement stricter 1 0  (29.4%) 9 (29.0%) 20 (47.6%) 
Control Less strict 23 (67.6%) 1 8  (58. 1 %) 1 5  (35.7%) 
Unknowna 1 (3.0%) 4 (1 2.9%) 7 ( 1 6.7%) 
Effect of positive 6 (1 7.6%) 5 (1 6. 1%) 6 ( 14.3%) 
directly paying negative 25 (73.5%) 23 (74.2%) 25 (59.5%) TB testing 
Unknown 3 (8.8%) 3 (9.7%) 1 1  (26.2%) 
Effect of Positive 5 ( 14.7%) 7 (22.6%) 1 3  (31 .0%) 
removal of Negative 21 (61 .8%) 1 5  (48.4%) 1 8  (42.9%) compensation 
NoChangeb 8 (23.5%) 9 (29.0%) 1 1  (26.2%) 
a includes 1 Wairarapa non-TB with no response to this question 
b includes 4 Wairarapa non-TB with no response to this question 
Nearly half the farmers with TB free herds were in favour of stricter Movement Control, whereas 
less than a third of the focused and standard control farms shared this opinion. However, across the 
three farm groups there was no statistically significant difference (X2=9.57, 4 df, p=O.046). Figure 
32 presents the percentage of answers given by the farmers to this question. 
Percent of farmers 
0% 40"/0 60% 80"/" 1 00% 
Focused control 
Standa:-d control 
Non-TB 
Cl more strici iI less strict D unknown 
Figure 32. Attitudes of farmers towards Movement Control restrictions, if they should be 
stricter or less strict, stratified by Wairarapa focused control, standard control, and non-TB 
farmers. 
1 82 
Twenty-four of 39 farmers (61 .5%) who indicated that MC should be stricter, argued that the risk 
of spreading TB could be reduced by this measure. Of the 56 farmers that were in favour of less 
strict MC regulations, 17  (33.9%) believed that the penalty was already high enough and that any 
further restrictions would impede the financial situation of the farm and therefore the likelihood of 
vector control being performed on the farm. 
There were no differences in the proportion of farmers in each group associated with the effects 
which they believed paying TB testing costs directly would have. Between 14% to 17% of the 
farmers believed in a positive effect of paying TB testing costs directly. Of a total of 17  farmers 
with this opinion 15 argued that this would give farmers financial incentive to pursue TB control 
more effectively. Of the 73 farmers who believed in a negative effect of having to pay TB testing 
costs directly, the main arguments were lost co-operation by farmers (n=1 9), the reasoning that TB 
is a national problem not a problem of individual farmers (n=6), and that this would result in less 
available resources for TB control on farms (n=6). 
With regards to the effect of removing compensation for TB reactor cattle, there were no 
differences between the three farm groups (Figure 33). The non-TB farmers had the highest 
percentage of farmers believing in a positive effect of removing compensation. Of 25 farmers who 
believed in a positive effect of removal of compensation, 21  used the argument of additional 
financial incentive for this effect. Lost co-operation, e.g. farmers testing less frequently, or not 
presenting all animals for testing, was the main argument for the negative effect of removal of 
compensation (n=42 of 54 farmers). 
Percent of farmers 
0% 20% 40% 60% 80% 1 00% 
Focused control 
Standard control 
Non-TB 
Figure 33. Fanners' belief regarding the effect of removing compensation for TB reactor 
cattle, stratified by Wairarapa focused control, standard control, and non-TB fanners. 
When farmers were asked whether they could see any value in vector control or other forms of TB 
control being conducted by farmers themselves, four main answer categories were found. Some 
1 83 
farmers believed that farmers should/could help the Regional Council and continue their work once 
a main vector control programme had been conducted. Other farmers thought it reasonable that 
farmers look after their own properties, whereas others did not see any importance or responsibility 
for vector control by farmers or argued that this would not be effective as not all farmers would 
participate. Table 3 1  gives the distribution of these four categories amongst the answers of the three 
farm groups.  
Table 31.  Importance of farmer conducted vector control as seen by the three farm groups -
focused control, standard control, and non-TB farmers. 
Focused control Standard control Non-TB 
farmers (n=34) farmers (n=31) farmers (n-42) 
Help RCI continue 8 (23.5%) 1 3  (41 .9%) 1 8  (42.9%) 
Responsible for own farm 8 (23.5%) 6 (19 .4%) 3 (7. 1 %) 
Not effective I responsibility of RC 3 (8.8%) 4 (1 2.9%) 5 (1 1 .9%) 
No importance seen in conducting 9 (26.5%) 8 (25.8%) 1 5  (35.7%) 
own vector control on farm 
Other answers or no answer 6 (8.7%) 0 1 (2.4%) 
The focused control farmers had the largest proportion of farmers believing that farmers are 
'responsible for their own farm' ,  while the non-TB farmers had the highest proportion of farmers 
that did not see any importance in conducting their own vector control on their farms. Statistically 
there was no difference between the farm groups with respect to the answer categories after 
applying the correction term (l=16.49, 8 df, p=0.030). 
Farmers were asked whom they thought responsible for the eradication of TB , whereby they could 
nominate more than one institution/organisation. They were also asked to give priorities to these 
organisations. Table 32 gives the seven most nominated organisations and the number of farms that 
ranked the organisations with highest priority. Figure 34 presents the percentage of all farmers 
considering the different organisations as being responsible for TB eradication and the number of 
times, the different organisations were given highest priority in being responsible. 
The nominations were similar across the three farm groups. A wide range of organisations/ 
institutions was mentioned by the farmers, with Government and AHB/ AgriQuality receiving most 
frequently receiving highest priority. A slightly higher percentage of non-TB farmers voted 'only 
farmers with the problem' ( 1 9.0 %) as the highest priority compared with focused and standard 
control farmers (8.8 %, respectively 6.S %). However, there was no statistical significant difference 
between the answer groups across the three farm groups (X2=14.40, 1 4  df, p=0.43 1 ). 
1 84 
Table 32. Organisations considered by Wairarapa herd managers to be responsible for 
eradicating TB (in brackets the number of farms that ranked these organisations with the 
highest priority). The last column gives the percentages of all farmers that nominated this 
organisation. 
Focused Standard 
control control 
farmers farmers 
Government 23 (19) 26 (1 1 )  
Regional Council 23 (5) 24 (4) 
All farmers 25 (10) 23 (7) 
AgriQuality and Animal Health 17 (7) 20 (4) 
Board 
All landowners 1 9  (9) 20 (7) 
Only farmers with the problem 3 (3) 6 (2) 
Other organisations/groups 7 (2) 1 0  (0) 
Local farming group 6 ( 1 )  3 (0) 
a on two standard control farms this information could not be obtained 
I/) ... Cl) E ... t'G -
� -0 -c Cl) u "-Cl) c. 
80 
60 
40 
20 
0 
. _. .�. -.n � 
c:=:J % of respondents --+-Highest priority 
Non-TB % of all 
farmers farmers 
(n=107)8 
27 ( 12) 71 .0 % 
28 (9) 70. 1 % 
26 (6) 69.2 % 
34 (29) 66.4 % 
22 (7) 57.0 % 
1 4  (8) 21 .5 % 
5 ( 1 )  20.6 % 
3 (0) 1 1 .2 % 
-I/) 
50 Cl) .s:: 
.21 
40 .s:: 0) c 
30 � 
0) >-
20 1/) :::: ... .... Cl) 0 
1 0  E .;: ... C. t'G -
0 -0 
... Cl) ..0 
E ::::I 
Z 
Figure 34. Organisations/institutions perceived as being responsible for TB eradication. 
Table 33 gives a similar presentation of organisations that were held responsible by the fanners for 
actually doing the work required to eradicate TB. 
185 
Table 33. Organisations considered by Wairarapa herd managers to be responsible for doing 
the actual work to eradicate TB (in brackets the number of farms that ranked these 
organisations with the highest priority). The last column gives the percentages of all farmers 
that nominated this organisation. 
Focused Standard Non-TB % of all 
control control farmers farmers 
farmers farmers (n=107)8 
Regional Council 31 ( 1 5) 26 ( 12) 28 ( 17) 79.4 % 
All farmers 1 7  (5) 17 (3) 20 (6) 50.5 % 
AgriQuality and Animal Health 1 3  (4) 17 (2) 24 ( 1 0) 50.5 % 
Board 
All landowners 1 5  (5) 1 2  (1 ) 20 ( 10) 43. 9 % 
Government 12 (5) 1 6  (5) 1 8  (8) 43.0 % 
Only farmers with the problem 5 (4) 1 4  (7) 1 4  (7) 30.8 % 
Other organisations! groups 5 (2) 8 (0) 3(1 )  1 4.9 % 
Local farming group 5 (2) 2 (0) 3 (0) 9.3 % 
a on two standard control farms this information could not be obtained 
The Regional Council was the most frequently nominated organisation, followed by AgriQualityl 
Animal Health Board and 'all farmers' .  Around one third of farmers nominated 'only farmers with 
the problem' as being responsible for actually doing the work required to eradicate TB. 
Farmers were asked where they would expect help from if they had to do their own vector control 
on their farms, and what expectations of help they would have from these organisations. The most 
sought-after organisation was the Regional Council. Twenty-eight (82.4%) focused control, 24 
(77.4%) standard control and 3 1  (73.8%) non-TB farmers expected help from the Regional 
Council. The main help expected from Regional Council was subsidised poison, traps and bait 
stations (n=33), advice (n=28), actual work (n=18) and funding (n=15). 
The second most frequently mentioned organisation from which farmers expected help if they had 
to conduct their own vector control was AgriQuality/Animal Health Board. Nineteen (55.9%) 
focused control, 22 (71 .0%) standard control and 16 (38. 1%) non-TB farmers would expect help 
from AgriQuality/AHB. Advice (n=20), funding (n=7) and co-ordination (n=4) was the main help 
expected from these organisations. 
Eleven (32.4%) focused control, 1 1  (35.5%) standard control farmers and 12 (28.6%) non-TB 
farmers expected help from Government, whereby financial help was mostly expected (n=23). 
Thirteen (38.2%) focused control, six ( 19.4%) standard control and four (9.5%) non-TB farmers 
1 86 
would also have expected co-operation of their neighbours if they had to control vectors 
themselves. 
Perceived cost of TB and its control by farmers 
Farmers were also asked what their perception was of the costs of TB and its control to their farms. 
Information for the focused control farms regarding this was collected in the flrst part of the 
questionnaire in 1997. For two of the standard control farms this information could not be collected 
and this question was not asked in the questionnaires for the non-TB farmers. Table 34 gives the 
distribution for the focused and standard control farms. 
Table 34. Estimated costs of TB and its control to Wairarapa focused and standard control 
farms. 
Focused control farmers Standard control farmers 
Cost of TS = 0  1 2  14  
to  the farm > 0  17  1 7  
(avg $1 1 ,21 4, range $650 - $50,000) (avg $9,759, range $1 50 - $75,000) 
Unknown 5 0 
Cost of = 0  1 1  1 5  
poison > 0  21 1 6  
(avg $562, range $45 - $2,500) (avg $952, range $30 - $4,200) 
Unknown 2 0 
Cost of = 0  29 21 
labour > 0  4 7 
(avg $1 ,288, range $50 - $2,000) (avg $792, range $ 150 - $2,000) 
Unknown 3 
Only about half the farmers put a cost greater than zero on the effect of TB on their farms, with a 
wide range of estimates. The reasons given for the costs focused on lost opportunities (14 focused 
control farmers, 7 standard control farmers), lost reactor value (2 focused control farmers, 7 
standard control farmers), and inconvenience and time involved in TB testing (7 focused control 
and two standard control farmers). Other reasons given were costs due to levies and rates (n=2), 
lost production (n=3) and cost due to potential export threats (n=I).  
Most farmers (n=50; 76.9%) did not put any cost on their own labour, they argued that they do it  as 
part of their normal management. 
187 
Multivariate analysis between Wairarapa focused and standard control farms 
In order to assist the interpretation of the Bonferroni correction, an uncorrected univariate analysis 
was followed by a logistic regression comparing focused control with standard control farms. Five 
variables had p-values of less than 0.10 in the univariate analysis: buying frequency, the likelihood 
of contact between livestock and possums, the likelihood of contact between livestock and ferrets, 
the total number of days spent by farmers on vector control and the variable sociable in the self­
concept. A logistic regression was conducted with these five variables. Only one variable was 
significant (buying frequency with Wald statistic 7.08, p=O.029), which was less than expected by 
chance. 
National study farms 
General farm characteristics 
The questionnaires were used to compare general farm characteristics for the two farm groups. 
Although the farmers were interviewed in 2000, they were asked to give stock numbers as they had 
been in June 1998. This way the stock numbers related more closely to the time period when the 
project was actively conducted. One farmer was unable to provide stock numbers as he had left the 
farm and could not remember. 
Table 35 shows some of the general farm characteristics on income, farm size and labour units for 
the two farm groups. In total 16% (n=15) of the national farms used off-farm grazing. Around 50% 
of the farm income is generated through sheep and less than 40% through cattle. 
188 
Table 35. General farm characteristics of National focused and standard control farms. 
Focused control Standard control 
(n=33)8 (n=63) 
Farmed area (ha) b 1 570.7 1 91 5.3 
Number of farms using off-farm grazing 7 (2 1 .2 %) 8 ( 12.7 %) 
Self-concept average value c 3.22 (n=30) 3.24 (n=58) 
Income from cattle (% of total income) 37.55 37.95 
Income from sheep (% of total income) 46.53 50.48 
Income from deer (% of total income) 1 4.27 9.84 
Full time labour units on farm 2.1 8 1 .99 
a gives the sample size for each farm group, unless otherwise stated 
b is the total number of effective hectares on the home property, plus other owned land if less than lookm 
away, plus leased land if less than lookm away 
C from Seabrook ( 1984) 
The average size of all farms in the National project was 1796.9 ha of effective land (s.e.=3 1 3.3, 
n=96, median 650 ha). A frequency distribution of farm size for all farms is shown in Figure 35. 
50 
IJI 
E 40 � to -
- 30 0 � Q) 20 
� o 
J:J 
E ::s 1 0  Z 
0 0 0 Cl = .. 0 = = . CI Cl I::l 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 If) L!') L!') l!) L!') l!) L!') L!') L!') 0 N (") v l!) <.C I'- co (j) /\ 
Effective farm area (ha) 
Figure 35. Effective farm size distribution of all farms included in the national study. 
The size of focused control farms did not differ significantly from standard control farms, as shown 
in (Figure 36). 
189  
20000 
,--.. tu 1 5000 <S-w N (j) 
E 
� 1 0000 (l) .;:: 
(3 
� w 
5000 
o +-________ �_+-L--------�+_�--------_, 
F ocusecl control Standard control 
Figure 36. Violin plots for effective farmed area of national focused and standard control 
farms. 
Table 36 presents the average effective farm size for the different enterprise types. As there were 
only three beef finishing farms within the national farm group, these were combined with the beef 
breeding farms to make 'beef farms. 
Table 36. Average farmed area stratified by herd type and national farm group (with range 
in brackets). 
Focused control farms Standard control farms 
Beef 1 994.5 (62 - 7500) n=24 2422.0 (1 26 - 1 4560) n=46 
OH 338.8 (342 - 1 2000) n=4 1 32.9 (80 -1 6121 ) n=7 
Deer 525.2 ( 197 - 500) n=5 832.2 ( 1 00 - 1 85) n=1 0 
On average the farms comprised a total of 6924 livestock units (n=95 herds ;  s .e.::::990. 19 ;  median 
441 1 , n=95) .  This was made up of 2807 cattle SU (s.e.=86 1 . 3 1 ;  median 1 365.5), 3733 sheep SU 
(s.e.=4 13 .35 ;  median 2630) and 328 deer SU (s.e.=84. 16 ;  median 0). Only 90  farms had cattle, 78 
had sheep and 29 had deer on the property. Cattle and deer comprised an average proportion of 
0.49 (s.e.=0.032;  median 0.38) of the total number of livestock units kept on the farms. The 
average cattle/deer density on the national farms was 3 . 8  SU per effective hectare of land 
(s .e.=0.36; median 2.70). Table 37 presents this information for the focused and standard control 
farms. 
1 90 
Table 37. Livestock units and stock densities of cattle and deer on national focused and 
standard control farms. 
Average livestock units on farm 
Cattle and deer proportion of total SUa 
Cattle and deer density (CSU + DSU)/ha 
a SU = livestock units 
Herd manager 
Focused control farms 
(n==33) 
6944 
(s.e.=1 238, median 5 14 1 )  
0.53 
(s.e .=0.055; median 0.41 )  
3 .80 
(s.e.=0.523; median 3. 1 9) 
Standard control farms 
(n=63) 
691 3  
(s.e. = 1 361 ; median 4256) 
0.47 
(s.e.=0.039; median 0.37) 
3.73 
(s.e .=0.484; median 2.36) 
One of the standard control farmers did not respond to this part of the questionnaire, resulting in a 
total number of95 farms. 
The median age of herd managers of focused and standard control farms was 40-50 years (Figure 
37).  The age distributions were bell curve shaped with smaller numbers of farmers in the 
extremities and the majority of farmers in the central age groups. 
35 
30 1/1 ... GI 25 E ... 
of! .... 0 .... 1 5  c: GI " 1 0  ... GI D.. 5 
0 Nat Standard 
>60 
Age group 
Figure 37. Distribution of age groups for herd managers of national focused and standard 
control farms. 
Table 38 presents some of the information on herd managers for focused and standard control 
farms. There were no obvious differences in these characteristics between the focused and standard 
control farms. The majority of herd managers, 96. 8% (n=92), lived on the farm; 2 1 . 1% (n=20) had 
1 9 1  
additional employment; and 80.0% (n=76) farmers covered their living expenses through the farm 
income. Only one of the 95 fanners was female. 
Eighty percent (n=76) of interviewees owned the farm, one farmer was a share milker and 16 .8% 
(n=1 6) were managers of the farm, two were associated in other ways with the farm they worked 
on. Fifty-nine percent (n=56) of all farmers considered themselves to be the main decision makers 
for the farm. Fanners had worked an average of 1 8.6 years (s.e.=1 .23,  median=18 )  on their current 
farms, while they had worked in farming an average of 29.6 years ( s .e.= 1 .27, median=30). 
Nineteen percent of herd managers (n=1 8) started their farming job without a farming background, 
whereas 8 1 . 1  % Cn=77) were brought up on a farm. Sixty-six percent (n=63) of herd managers had 
no formal farming-specific qualification. A higher percentage of herd managers of focused control 
farms than of standard control farms had a formal qualification (42.4% (n= 14) herd managers of 
focused control farms versus 30.6% (n=19) herd managers of standard control farms, however, this 
difference was not significant, l=1 . 30, 1 df, p=O.253). 
Table 38. Information on herd managers for national focused and standard control farms. 
Focused control farmers Standard control farmers 
(n=33) (n=62) 
Live on farm 30 (90.9%) 61  (98.4%) 
Cover l iving expenses from farm income 27 (81 .8%) 49 (79.0%) 
Owner of farm 26 (78.8%) 50 (80.6%) 
Major decision maker 22 (66.7%) 34 (54.8%) 
Average working time on current farm 1 8.79 1 8.23 
(5.e.=2.20; median 20) (s.e.=1 .51 , median 1 6, n=6 1 )  
Average working time on farms 29.94 29 .23 
(s.8.=2.1 9; median 34) (5.8.=1 .58, median 30) 
Farming background 26 (78.8%) 50 (80.6%) 
Formal qualification 14 (42.4%) 1 9  (30.6%) 
Interviewees were also asked to fill in a self-concept form, which was adapted from Seabrook 
( 1984). In total 88 farmers responded to this part of the questionnaire. Figure 38 shows the 
arithmetic means of personality traits for the focused and standard control farmers. 
192 
-- Nat . Focused control - - Nat. Standard control 
Not easy going 
Meek 
Patient 
Unsociable 
Not modest 
Persevering 
A worrier 
Cheerful 
Talkative 
Speaks one's mind 
Difficult to get on with 
Lacking confidence 
Liking change 
Forceful 
Prefers machinery 
Prefers bying new machine 
Dislike to learn 
Still learning 
Likes to avoid hard work 
Dislikes using records 
Values traditional ways 
Dislikes to set targets 
Likes to look after favourite 
animals a bit better than the 
-<: 
-:% 
� ") / 
LL: 
�>-. �  
-
I",;:" 
1/ 
....-?' 
� 
� 
� 
--
.-J 
---n 
f.....� 
'\ -� 
...... '-
� 
� 
, 
, 1 
Ir I I l 
1 1 
1 
� ! 
� 
i 1 
> 
I 
Easy going 
Not meek 
Impatient 
Sociable 
Modest 
Giving up easily 
Not a worrier 
Grumpy 
Not talkative 
Keeps quiet 
Easy to get on with 
Confident 
Suspicious of change 
Giving in easily 
Prefers animals 
Prefers choosing new animal 
Very keen to learn 
Very knowledgeable 
Values hard work 
Likes using records 
Likes adopting new ideas 
Likes setting targets 
Likes to strictly monitor 
performance of the herd 
Figure 38. Personality trait means for national focused and standard control farms (adapted 
from Seabrook, 1984). 
Differences were found between herd managers in the focused and standard control farm groups. 
Herd managers of focused control farms tended to be less worrying (Mann-Whitney U test, Z= 1 .90, 
p=0.057), more cheerful (Z=-2 .24, p=0.0 1 5), more talkative (Z=- 1 .62, p=0. 1 06), more open to 
change (Z=-2 . 5 1 ,  p=O.0 1 2) and new ideas (Z=1 . 80, p=0.07 1 ), considered themselves much more as 
still learning (Z=-2 .50, p=0.0 12) and preferred using records (Z=1 .68, p=O.092) more than herd 
managers of standard control farms. However, after applying the correction term as numerous tests 
were conducted, these differences were not statistically significant. 
Stock management 
Farmers were also asked about their stock movements on and off the farm. Table 39 presents the 
results for the two national farm groups. There were no significant differences in the frequency of 
buying white-tagged cattle, nor the frequency of buying any cattle, nor the importance of trading 
between the two farm groups (p uncorrected > 0. 10).  
193 
Table 39. Information on stock movements on and off farms for national focused and 
standard control farms. 
White Tagged cattle bought Yes 
No 
Buying frequency Every year 
Now and again 
Other/nevera 
Trading importance Crucial 
Important 
Moderately 
Minor/unimport. 
Not stated 
Focused Control 
(n=33) 
5 (1 5.2%) 
28 (84.8%) 
9 (27.3%) 
9 (27.3%) 
1 5  (45.4%) 
5 (1 5.2%) 
5 (15.2%) 
9 (27.3%) 
1 4  (42.4%) 
Standard Control 
(n=63) 
4 (6.3%) 
59 (93.7%) 
25 (39.7%) 
1 4  (22.2%) 
24 (38.1%) 
1 1  (1 7.5%) 
1 8  (28.6%) 
6 (9.5%) 
28 (44.4%) 
a farms that only bought in single bulls for breeding purposes were classified under 'other' 
Eighteen (54.5%) national focused control and 34 (54.0%) national standard control farmers 
indicated that they had changed their stock selling policies due to having TB in their herds. Eight 
(24.2%) national focused control and 1 6  (25.4%) national standard control farmers also indicated 
that they had changed their mix of stock classes as a result of TB. Most farmers changed from 
selling weaners to finishing more stock themselves. 
TB risk assessment 
Table 40 presents information on the reasons farmers suspect to be the cause of TB infection in 
their herd, about high TB risk areas, neighbouring herds and TB reactor numbers, stratified for 
national focused and standard control farms. 
194 
Table 40. Information on the TB situation and perception of national focused and standard 
control farms. 
Focused control Standard control 
(n=33) (n=63) 
TB problem classification Continuous 14 (42.4%) 1 5  (23.8%) 
On and off 1 2  (36.4%) 25 (39.7%) 
Seldom 4 (1 2. 1%) 1 4  (22.2%) 
No problem anymore 3 (9. 1%) 9 (14.3%) 
Reason for TB i nfectiona TB feral animals 31 (93.9%) 55 (87.3%) 
Neighbouring herds 1 3  (39.4%) 24 (38. 1%) 
Stock movements 1 (3.0%) 1 2  (1 9.0%) 
High risk areas for TB Yes 22 (66.7%) 25 (39.7%) 
suspected on farm No 8 (24.2%) 1 9  (30.2%) 
Unknown/not stated 3 (9. 1%) 1 9  (30.2%) 
Graze cattle in high risk Yes 1 8  (81 .8%) 20 (80%) 
areas No 4 (1 8.2%) 5 (20%) 
Any neighbouring farms Yes 32 (97.0%) 55 (87.3%) 
infected with TB No 0 2 (3.2%) 
Unknown 1 (3.0%) 6 (9.5%) 
Reactor numbers over last Increased 4 ( 12. 1%) 6 (9.5%) 
three years Decreased 24 (72.7%) 43 (68.3%) 
No change/not stated 5 (15.2%) 1 4  (22.2%) 
• mUltiple answers were possible 
The national focused control farms had a higher proportion of farmers who believed that the TB 
problem in their herd was continuous, compared with the standard control farmers (i=3.56, Idf, 
p=0.059). However, after applying the correction term for numerous tests, this difference was not 
statistically significant. 
Most farmers regarded infected wildlife as the source of TB infection in their herds, around 40% of 
farmers believed that the neighbouring herds could have been a source, whereas stock movements 
on and off farms were only mentioned by a minority of farmers. 
A higher proportion of focused control farmers knew or suspected high risk areas for TB on their 
farms, whereas a higher proportion of the standard control farmers were unsure about such risk 
areas (Chi-squared test over the three categories and the two farm groups -l=7.68, 2 df, p=O.021 ;  
which was not significant after the correction term was applied). 
195 
Vectors and vector control on  farms 
Farmers provided infonnation on how likely they thought it was that there was contact between 
their livestock and feral animals. Table 41 presents the data for possums and ferrets stratified for 
focused and standard control farms. Slightly more than half the focused and standard control 
farmers believed that contact with possums was very likely, while between 35% and 42% believed 
that contact between livestock and ferrets was very likely. 
Table 41. Assumed likelihood of contact between possums/ferrets and livestock, as indicated 
by national focused and standard control farmers. 
Focused control Standard control 
farms farms 
Likelihood of contact with Very l ikely 18 (54.5%) 34 (54.0%) 
possums Possible 1 1  (33.3%) 25 (39.7%) 
Unlikely 4 (1 2 .2%) 0 
Unknown / not stated 0 4 (6.3%) 
likelihood of contact with Very l ikely 14 (42.4%) 22 (34.9%) 
ferrets Possible 10 (30.3%) 22 (34.9%) 
Unlikely 3 (9. 1 %) 3 (4.8%) 
Unknown I not stated 6 ( 18 .2%) 1 6  (25.4%) 
Farmers were also asked about vector control on their properties, what percentage of land was 
controlled by themselves and/or by the Regional Council and how much time was spent on these 
controls (see Table 42 and Figure 39). 
196 
Table 42. Vector control by Regional Council and farmers as stated by national focused and 
standard control farmers. 
Focused control Standard control 
Avg % of farm controlled by RCa 82.1 % (n=3 1 )  78.6% (n=49) 
Avg % of farm controlled by 68.6% (n=32) 62.0% (n=45) 
farmera 
Regional Council control Increased 10 (30.3%) 17 (27.0%) 
Decreased 6 (1 8.2%) 21 (33.3%) 
armer Control Increased 13  (39.4%) 1 8  (28.6%) 
Decreased 9 (27.3%) 1 8  (28.6%) 
Farmer Control (days/year) = 0  1 (3.0%) 14 (22.2%) 
> 0  32 (97.0%) 49 (77.8%)b 
(median 20, avg 34.3, (median 1 3, avg 28.4, 
range 1 - 275) range 2.5 - 365) 
Regional Council (days/year) = 0  5 (1 5.2%) 25 (39.7%) 
> 0  28 (84.8%) 38 (60.3%) 
(median 10 ,  avg 1 2.7, (median 1 0, avg 14.2, 
range 1 - 40) range 2 - 80) 
Farmer Control Regularly 6 (1 8.2%) 1 8  (28.6%) 
a according to the farmer 
b four farmers conducted vector control, but considered the area controlled by themselves as being zero 
g. 60 
0 "-
Cl 50 � r-CJ ca Cl,) 40 r--
c 
Cl) 30 E r-"-ca - 20 - fo-0 -c 1 0  Cl,) I-CJ "-Cl,) 0 Il. Ir 
o 
-
I rl n.n 
o co 
, 
o C\I .... 
• 
o co .... 
o o C\I 
o � C\I 
o Focused control 
• Standard control 
n 
o Cl) C\I 
o C\I (I') 
Days spent per year on vector control by farmers 
f---
-
Figure 39. Days spent per year on vector control by national focused control and standard 
control farmers. 
197 
While a higher proportion of focused control farmers than standard control farmers conducted their 
own vector control, this difference was not significant after applying the correction term (Mann­
Whitney U test 2=2.64, p=O.OO8). The time spent on vector control by farmers themselves was 
greater for focused control than for standard control farmers, but this difference was also not 
statistically significant after correction for numerous tests (Mann-Whitney U test 2=2.33, p=0.020). 
Attitudes towards TB and its control 
Farmers were asked about their attitudes towards TB control and possible changes in the future. For 
most the farmers eradication of TB was crucial (Figure 40). Twenty-seven (8 1 .8%) of the 33 
national focused control farmers and 34 (54.0%) of the 62 national standard control farmers with 
relevant information, considered TB eradication as crucial. Although this difference was 
statistically significant in the test (Chi-squared test for crucial, all other categories were combined 
X2=6.75, 1 df, p=0.OO9) it was not considered significant after the application of the correction term. 
Percent of farmers 
0% 20"/0 40"/0 60% 80% 1 00% 
Focused control 
Standard control 
o crucial 1\1 , ,.,-,nrVi',, Ominor 
Figure 40. Importance of TB eradication as considered by national focused and standard 
control farmers. 
Most farmers thought TB eradication was possible on their farms (75.8% (n=25) focused control 
farmers and 7 1 .4% (n=45) standard control farmers). Six focused control and nine standard control 
farmers considered TB eradication on their properties impossible due to vegetation and habitat 
factors. Two focused control farmers and nine standard control farmers answered 'unknown' to the 
question about whether or not it is possible to eradicate TB on their properties. 
Answers obtained by focused and standard control farmers regarding their attitude to stricter 
Movement Control restrictions, having to pay their own TB testing costs directly and about the 
effect of removing compensation for reactor cattle are presented in Table 43. The farmers were also 
198 
asked for their reasons why they chose the responses, but not all farmers gave specific reasons to 
this question. 
Table 43. Attitudes towards TB control by national focused and standard control farmers 
(percentages of fann group in brackets). 
Focused control Standard control 
farmers fanners 
Movement Control stricter 21 (63.6%) 28 (44.4%) 
Less strict 1 2  (36.4%) 24 (38.1%) 
Unknowna 0 1 1  ( 17.5%) 
Effect of directly paying TB testing positive 6 (1 8.2%) 1 1  ( 17.5%) 
negative 25 (75.8%) 41 (65.0%) 
Unknownb 2 (6.0%) 1 1  (17.5%) 
Effect of removal of compensation Positive 6 (1 8.2%) 1 2  ( 1 9.0%) 
Negative 21 (63.6%) 35 (55.5%) 
NoChangeb 6 (1 8.2%) 1 6  (25.4%) 
a includes 3 national standard control fanners with no response to this question 
b includes 1 national standard control fanners with no response to this question 
A higher proportion of focused control farmers were in favour of stricter MC regulations (Figure 
41) .  Thirteen focused control and 19 standard control farmers that where in favour of stricter MC 
regulations gave the reason that it would result in a reduced risk of spreading TB . Of the 36 farmers 
that were in favour of less strict MC regulations, five in each group (15 .2% focused control 
farmers, 7.9% standard control farmers) believed that the penalty is already high enough. 
0% 
Focused control 
Standard control 
20% 
Percent of farmers 
40% 60% 80% 1 00% 
�� � "' ".  ...,....�Y'''''� _ . �' ... < • 
N� • • 
o stricter - less strict 0 unknown 
Figure 41. Percent of national focused and standard control fanners who believed Movement 
Control restrictions should be more or less strict. 
199 
There was no difference in the proportion of farmers per group who judged the effect of having to 
pay TB testing costs directly as being positive or negative (see Table 43). Around 1 8% (n=1 7) of 
farmers believed in a positive effect of paying directly, reasoning that this would increase the 
financial incentive for farmers to control TB (n=1 focused and n=4 standard control farmers).  The 
main arguments for the 66 farmers who believed in a negative effect of having to pay TB testing 
costs directly were lost co-operation of farmers (n=27), TB is a national problem not a problem of 
individual farmers (n=5), and that this would result in less available resources for TB control on 
farms (n=5). 
Most farmers (64% of focused control farmers and 56% of standard control farmers) were not in 
favour of removing compensation for reactor cattle. Of the 56 farmers who believed in a negative 
effect of removing compensation, 46 argued that it would result in lost co-operation, e.g. farmers 
testing less frequently, or not presenting all animals for testing. Eighteen to 1 9% of farmers 
believed in a positive effect, associated with increased financial incentive (n= 1 1 ) .  
The reasons given b y  farmers i n  relation to their response to the question about conducting their 
own vector control or other TB control programmes, are given in Table 44. 
Table 44. Importance of fanner conducted vector control as seen by focused and standard 
control farmers. 
Focused control farmers Standard control farmers 
Help RCI continue 1 7  (S1 .S%) 31 (49.2%) 
Responsible for own farm 4 ( 12. 1 %) 4 (6.3%) 
Not effective I responsibility of RC 8 (24.2%) 5 (7.9%) 
No importance seen in conducting own 0 1 9  (30.2%) 
vector control on farm 
Other answers or no answer 4 ( 1 2. 1%) 4 (6.3%) 
There was a significant difference in the answers given by focused and standard control farmers 
(X2= 1 5.96, 4 df, p=O.003). None of the focused control farmers answered 'no importance seen for 
conducting own vector control' , whereas 30% of the standard control farmers did not see any 
importance in conducting their own vector control on their farms. However, a higher proportion of 
focused control farmers believed that farmer conducted vector control would not be effective or 
that the control should be left to the Regional Council. 
Table 45 gives the seven most frequently nominated organisations and the number of farmers that 
ranked the organisations with highest priority in response to the question about whom they thought 
200 
responsible for TB eradication. Figure 42 presents the percentage of all farmers considering the 
different organisations as being responsible for TB eradication, and the number of times the 
different organisations were given highest priority in being responsible. No individual organisation! 
institution was standing out as being nominated most frequently, however, Government was 
mentioned most frequently with the highest priority . 
Table 45. Organisations considered by national herd managers to be responsible for 
eradicating TB (in brackets the number of farms that ranked these organisations with the 
highest priority). The last column gives the percentages of all farmers that nominated this 
organisation. 
Focused control Standard control % of all farmers 
farmers farmers (n=96) 
All farmers 26 ( 12) 47 ( 1 5) 76.0 % 
Regional Council 24 (5) 46 ( 1 2) 72.9 % 
All landowners 20 (4) 45 (1 5) 67.7 % 
Government 24 ( 14) 41 (1 8) 67.7 % 
AgriQuality and Animal Health Board 21 (7) 42 ( 1 2) 65.6 % 
Other organisations/groups 1 6(4) 20 (4) 37.5 % 
Only farmers with the problem 7 (3) 1 8  (6) 26.0 % 
Local farming group 2 (1 ) 1 8  ( 1 )  20.8 % 
-
c::J % of respond_ents -+- Highest priority 
'" Cl) .!: 
'" ... Cl) 
E ... m --o -c:: Cl) () ... Cl) c. 
80 
60 
40 
20 
o 
.!? 
40 .s:; Cl c:: 
30 > Cl >-
20 ", =:  ... ... Cl) 0 
E 'i:  
1 0  ... � m --
0 0 ... Cl) .0 
E ;:, 
z 
Figure 42. Organisations/institutions perceived as being responsible for TB eradication. 
201 
Table 46 gives a similar presentation of organisations that were considered by the farmers to be 
responsible for actually doing the work required to eradicate TB. 
Table 46. Organisations considered by national herd managers to be responsible for doing 
the actual work to eradicate TB (in brackets the number of farms that ranked these 
organisations with the highest priority). The last column gives the percentages of all farmers 
that nominated this organisation. 
Focused control Standard control % of all farmers 
farmers farmers (n=96) 
Regional Council 24 ( 12) 39 ( 12) 65.6 % 
All landowners 1 5  (4) 38 (8) 55.2 % 
All farmers 15  (6) 37 ( 10) 54.2 % 
AgriQuality and Animal Health Board 1 2  (4) 37 (1 3) 51 .0 % 
Government 1 1  (7) 31 ( 1 0) 43.8 % 
Only farmers with the problem 6 (5) 17  (6) 24.0 % 
Other organisations/groups 1 2  (7) 1 1  (3) 24.0 % 
Local farming group 4 (2) 1 1  (1 ) 1 5.6 % 
The Regional Council was the most frequently nominated organisation, followed by 'all 
landowners' ,  'all farmers' and 'AgriQualityl Animal Health Board' .  Around one quarter of all 
farmers nominated ' only farmers with the problem' as being responsible for actually doing the 
work required to eradicate TB. 
In another question, farmers were asked where they would expect help from, if they had to do their 
own vector control on their farms, and what expectations of help they would have. Nineteen 
(57.5%) focused control and 35 (55.6%) standard control farmers expected help from the Regional 
Council; 18  (54.5%) focused control and 38 (60.3%) standard control farmers from 
AgriQuality/AHB, 1 1  (33.3%) focused control and 30 (47.6%) standard control farmers from 
Government. Thirteen (39.4%) focused control and 22 (34.9%) standard control farmers also 
expected the co-operation of their neighbours. The main help expected from Regional Council was 
advice (n=13), actual work (n= 1 1), funding (n=9) and subsidised poison (n=7). From 
AgriQuality/AHB the main help expected was advice (n=19), funding (n=13) and co-ordination 
(n=4). From Government financial help was mostly expected (n=27). 
202 
Perceived cost of TB and its control by farmers 
Table 47 gives the cost distributions for TB and its control for focused and standard control farms. 
Table 47. Estimated costs of TB and its control to national focused control and standard 
control farms. 
Focused control farmers Standard control farmers 
Cost of TB to the = 0  3 20 
farm > 0  25 37 
(avg $8,650, range $500 - $22,500) (avg $1 6,452, range $300 -
$1 80,000) 
Unknown 5 6 
Cost of poison = 0  9 27 
> 0  24 29 
(avg $745, range $ 100 - $2,500) (avg $552, range $50 - $2,500) 
Unknown 0 7 
Cost of labour = 0  1 8  32 
> 0  1 5  25 
(avg $1 ,963, range $1 00 - $5,000) (avg $2,656, range $1 00 - $25,000) 
Unknown 0 6 
Most of the focused control farmers put a cost greater than zero on the effect of TB on their farms, 
with a wide range of estimates. The difference between the focused and standard control farms was 
not significant after the application of the correction factor (X2=6.2 1 ,  2 df, p=0.045). 
The reasons given for the costs of TB focused on lost opportunities (14 focused control farmers, 27 
standard control farmers), lost reactor value (four focused control farmers, 12 standard control 
farmers), and inconvenience, direct costs and time involved TB testing (seven focused control and 
three standard control farmers). Other reasons given were costs due to levies and rates (n=4), lost 
production (n=5) and capital loss (n=4). 
Slightly more than half the farmers (n=50, 52. 1 %) did not put any cost on their own labour, as 
control work is being done as part of their routine management. 
Multivariate analysis between national focused and standard control farms 
In comparing focused control with standard control farms twelve variables had p-values of less 
than 0.10 in the uncorrected univariate analysis: trading importance, areas with high TB risk, graze 
203 
in high risk areas, TB infection contributed to animals on and off fann, the likelihood of contact 
between livestock and possums, the proportion of farm being controlled by Regional Council, the 
proportion of fann being controlled by farmer, the presumed effect of stricter movement control, 
and four variables within the self-concept (talkative, difficult to get on with, likes changes, 
knowledge). A logistic regression was conducted with these five variables. The only variable 
significant was the knowledge variable with Wald statistic 4 . 1 7, p=O.04 1 ,  which was less than 
expected by chance. 
Discussion 
This study looked in detail at the farm and management characteristics of all farms involved in the 
two intervention projects described in Chapters 3 and 4. Apart from providing the data for 
comparing focused control with standard control farms, the study provided useful insights into the 
way farms operated, which was particularly helpful in the Wairarapa project (see Chapter 3). 
Comparing focused control farms with standard control farms 
In order to identify differences between the farms the projects worked actively with (focused 
control farms), and the farms which were used for comparison (standard control farms), non­
parametric tests were employed, as most variables were not normally distributed. As there were 
many different variables (over 80) it was not possible to compute test statistics for each individual 
variable, without increasing Type I and Type n errors. This would have meant that some variables 
would have shown significance in the test, although there was no real difference between the farm 
groups. Therefore the variables were first screened visually and subjective comparisons were made 
between the groups. In the second step only those variables that showed differences in the visual 
comparison were tested statistically. In the third step the critical p-value for each test was set to 
0.0025, which resulted from a Bonferroni correction factor of 20 (Ott, 1 988). This way it could be 
ensured that differences found to be significant in the study had a high likelihood of being real 
differences. 
In the Wairarapa project none of the general farm variables nor the herd manager variables were 
significantly different between the focused control and the standard control farms. In general the 
focused control and standard control farms were more similar to each other in terms of general 
farm characteristics than the non-TB fanns. The similarity of the Wairarapa focused control and 
standard control farms is an indication that the random allocation of farms to these two groups was 
effective. Therefore any differences found in the intervention study cannot be attributed to 
204 
differences between the two fann groups, but are most likely due to the interventions conducted on 
these farms. 
Although in the National project no significant differences were found after the application of the 
correction term, there were more differences seen in the visual assessment than in the Wairarapa 
fann group. Farmers had to assess their own personality traits and whereas the values in most 
variables for this self-concept (Seabrook, 1984) were very close for Wairarapa focused and 
standard control farmers, these values differed more between national focused and standard control 
farmers. The visual assessment of the national fanners suggested that the focused control farmers 
were more cheerful, more open to change and conside ed themselves more as ' still learning' . This 
is in agreement with the biased selection of focused control farmers, who were selected specifically 
on their positive motivation, positive outlook and their TB herd history (for more details see 
Chapter 4). In studies with dairy and pig farmers it was found that differences in the self-concept 
could also reflect differences in fann management (Seabrook, 1984; Ravel et al. , 1996). These 
differences found between the national focused and standard control fanns would suggest that 
retrospective sampling was not optimal, but there was no scope to influence the selection process. 
Attitudes towards TB and its control 
In the Wairarapa study group the only significant difference between the focused and standard 
control farm groups was the proportion of fann area being controlled for vectors by fanners 
themselves. Of the farmers that performed vector control, the focused control farmers indicated that 
on average only 40% of their fann area was being controlled by themselves, while the standard 
control farmers controlled on average 65% of their farm and the non-TB farms 83%. However, 
these focused control farmers spent on average more days per year on vector control than the 
farmers of the other two groups. Although this difference was no longer significant after the 
application of the correction term, it could still indicate different perceptions of vector control. If 
the focused control farmers spent the same or even more time on less area, several explanations are 
possible. Firstly, the standard control farms needed less time input for control than the focused 
control farms, due to already reduced possum numbers on their farms, or secondly focused control 
farmers focused their effort and intensified their control on specific areas of their farms, while the 
farmers of the other two farm groups tried to cover the whole of their property. In 1 998, when this 
information was collected from focused control farmers, most farmers already had a good idea of 
high risk areas for tuberculosis through interactions with the research team. The 1998 figures were 
judged to be more conservative for the analysis than the 1999 figures, when the farmers themselves 
had the responsibility for maintaining the bait stations set up by the research team. The 1999 
figures might have been an overestimation of the actual time spent on control which can be 
205 
expected of farmers. It might be expected that the control effort by farmers in the third year of the 
project would decrease over time to the level of 1998, as the number of possums decreased, the 
number of reactors decreased or the perceived importance of vector control decreased. 
With respect to the attitudes of farmers towards the eradication of TB, an overall positive attitude 
existed among farmers, with around three quarters of all interviewees considering the eradication of 
TB as crucial and possible. Yet, nearly half of the farmers were in favour of less strict MC 
restrictions .  Around 70% considered the effect of having to pay TB testing costs directly as having 
negative consequences, and over 50% considered the effect of removing compensation for reactor 
cattle as being negative. These results are similar to those found in a study of farmers' attitudes in 
four additional regions of New Zealand, where 20% to 30% of the beef farmers believed in a 
positive effect of removal of compensation (Cowan and Clout, 2000). The negative consequences 
were mainly believed to be the loss of co-operation by farmers, e.g. farmers testing less frequently, 
or not presenting all animals for testing. Farmers believed this would have an overall negative 
effect on the TB eradication scheme. 
At the current stage of the control strategy farmer co-operation is crucial. If it is not possible to 
continue current levels of funding farmers will then have to play an increasing role in TB control 
(Animal Health Board, 2000). Therefore, it has to be evaluated whether the disadvantages outweigh 
the advantages, such as an increased incentive for farmers (Animal Health Board, 1995; Cullen and 
Bicknell, 2000). It is proposed that compensation should be kept at current levels for the next few 
years. With increasing numbers of farmers being aware of and using on-farm control methods, it 
will be possible to reduce compensation, in order to create a greater incentive for farmers currently 
not willing to employ on-farm control methods. 
Another potentially negative consequence of removing compensation could be that more and more 
farmers would cease farming cattle and deer. Although this would mean a lower number of herds 
being on MC, it would also mean that an important indicator of infection in feral animals would be 
removed. 
Many organisations and institutions were mentioned by farmers as being responsible for 
eradicating TB, whereby no single organisation was mentioned most frequently. This indicates that 
farmers do not have a clear idea of who is in fact responsible for the eradication of TB. It is 
noticeable, that on average less than a quarter of all farmers considered that only those farmers with 
the TB problem should be responsible for its control. Within the non-TB farm group in the 
Wairarapa only a third were of this opinion. Therefore the TB problem was considered by most 
farmers as a national problem, not an individual' s  problem. This is supported by the high 
206 
proportion of fanners considering 'all fanners' and Government as being responsible for 
eradicating TB . 
While fanners were considered equally responsible for the eradication of TB as were Regional 
Council, AgriQualityl AHB and Government, farmers were nominated less frequently in answer to 
the question of who is responsible for actually doing the control work. Over 70% of the fanners 
indicated that the RC was responsible for this, while only 50% of the farmers considered 
themselves as being responsible. The same attitude was found when farmers were asked whether 
they could see any importance in themselves conducting vector control or other means of TB 
control. Only 12% stated that farmers should be responsible for vector control on their own 
properties.  Forty-three percent of farmers believed that they could assist the RC, but nearly 40% 
did not see any importance for conducting their own vector control programmes or they believed 
that such programmes would be ineffective. This indicates that much more emphasis should be 
placed on motivating fanners and on fanners' awareness of TB and its control. Any future control 
strategies should include specific extension and education programmes that address these points. 
Another point which should be addressed in any future strategies is the cost of TB to the fanners. 
Only about 60% of the farmers interviewed could put a figure on the cost of TB to the farm. It was 
noticeable that the national focused control fanners had the lowest percentage of farmers (24%) not 
able to identify specific costs of TB to the farm. These were the farms with whom a multi­
disciplinary team was working, including financial advisors. 
Conclusions 
The study showed that the random allocation of farms into focused and standard control farm group 
was effective in the Wairarapa study and to a lesser degree in the national study. 
Overall a positive attitude existed among fanners regarding the importance of TB eradication. 
However, the majority of farmers were not in favour of stricter Movement Control regulations, 
removal of compensation or having to pay TB testing costs directly .  
Many farmers did not see themselves, but other organisations, such as Government and Regional 
Council, as being responsible for eradicating TB and did not see any importance in conducting 
control programmes themselves. 
Therefore, any future eradication strategy should include fanner motivation and some form of 
extension programme, that addresses specific points, such as the possible control strategies and 
fmancial implications of TB. 
207 
Bibl iography 
Agresti, A. ( 1990). Categorical data analysis. John WHey and Sons, New Yark. 
Animal Health Board (1995). TB Beyond 2000. Towards a National Pest Management Strategy. A 
Discussion Paper. Animal Health Board, Wellington, New Zealand 
Animal Health Board (2000) .  Bovine Tuberculosis National Pest Management Strategy 2001-20 1 1 ;  
Towards a TB free NZ; A discussion paper on the future options. Animal Health Board, 
Wellington, New Zealand. 
Bortz, J. ( 1 993) .  S tatistik fur Sozialwissenschaftler. Springer Verlag, Berlin. 
Corner, R. A . ,  Morris, R.S. and Davies, P.R. (2000). Attitudes of New Zealand Farmers to Bovine 
Tuberculosis and the Implications for Control Strategies. Final Report to the Animal Health 
Board - Project R-20452B. EpiCentre, Massey University, Palmerston North, New Zealand. 
Cowan P., Clout M.  (2000). Chapter Three: Possums on the Move: Activity Patterns, Home 
Ranges, and Dispersal. In: Montague T.L. (editor). The Brushtail Possum. Manaaki Whenua 
Press, Lincoln, New Zealand, pp.:24-34. 
Cullen R., Bicknell K. (2000). Chapter Eighteen: Economic Analysis of Possum Management. In: 
Montague T.L. (editor). The Brushtail Possum. Manaaki Whenua Press, Lincoln, New 
Zealand, pp. :  1 98-207. 
Morris, R.S. and Pfeiffer, D.U. ( 1 995) .  Directions and issues in bovine tuberculosis epidemiology 
and control in New Zealand. New Zealand Veterinary Journal, 43 (7): 256-265. 
Ott, L. ( 1 988).  An Introduction to Statistical Methods and Data Analysis. Third Edition. PWS-Kent 
Publishing Company, Boston. 
Ravel, A. ,  Ravel, A . ,  D'Allaire, S . ,  Bigras-Poulin, M. and Ward, R. ( 1 996). Psychodemographic 
Profile of S tockpeople Working on Independent and Integrated Swine Breeding Farms in 
Quebec. Canadian Journal of Veterinary Research, 60: 241 -248 .  
Seabrook, M.F. ( 1984) . The psychological interaction between the stockman and his animals and 
its influence on performance of pigs and dairy cows.  Veterinary Record, 1 15 :  84-87.  
208 
Beef farm in South Wairarapa 
Sheep 
209 
Chapter 6 
Economic evaluation of TB control programmes and 
potential benefits of using incentives or an insurance 
scheme for d ifferent farm types 
210 
2 1 1 
Abstract 
In New Zealand bovine tuberculosis is of major concern with potential consequences in the trade 
environment. The Australian brushtail possum is a wildlife reservoir and the main vector for 
transmitting the disease to livestock. Control and eradication of TB therefore has to involve 
livestock and wildlife, resulting in expensive control programmes. As a result of these programmes 
the herd incidence of TB has been reduced, however, more emphasis will have to be put on farms 
with persistent TB problems. Increasing expectations are put on farmers to take more responsibility 
for the TB control on their own properties. However, the adoption of available on-farm control 
measures will depend on their economic viability, and incentives or compliance auditing. 
Economic analyses were conducted using deterministic, stochastic and decision analyses methods 
to evaluate the economic outcomes for an average dairy, an average beef breeding and an average 
beef finishing farm. From the analyses it was concluded that under the current compensation level 
of 65% for TB test positive animals, the adoption of on-farm control measures generally was 
beneficial to dairy farms. For beef farms it was only economically beneficial if they achieved TB 
free herd status. Reducing the compensation level to zero did not alter the situation significantly. 
The net gain in dairy farms increased, the situation in the beef breeding farms changed hardly at all 
and on beef finishing farms the adoption of control programmes became beneficial if the number of 
TB animals was reduced at least by two. Thus, the reduction of compensation does not create a 
significant incentive for beef farmers to adopt on-farm control measures. Other forms of incentives 
might have to be put in place, such as provision of vouchers for vector control or payments for 
achieving a TB free herd status. 
Defin itions 
AgriQuality: the national veterinary service, contracted by the AHB to manage and conduct TB 
testing of all herds, was until recently part of MAF. 
Animal Health Board (AHB): the national organisation in charge of TB control. 
Herd types were abbreviated using 'BB' for beef breeding farms, 'BD' for beef dry stock farms, 
also termed beef finishing farms, and 'DH' for dairy herds 
MAF: Ministry of Agriculture and Forestry 
2 12 
Movement Control (MC): restrictions put on herds that limit the cattle and deer movements of 
herds that are suspected or known to be infected with TB 
Reactor: an animal that reacted positively to the caudal fold intradermal tuberculin test. 
TB animal: any animal that was found with lesions at slaughter (with or without previous positive 
skin test), whose lesions were found to be typical of TB by histology or culture. 
SD: standard deviation 
TB status: a herd is classified as 'Infected' if there is known TB infection in the herd, and 'Clear' if 
the herd had at least two whole-herd tests, in which none of the animals showed positive skin-test 
results. 
White-tagged animal: an animal identified with an official AHB Movement Control tag, to indicate 
that this animal comes from a herd that is infected with TB 
Whole-herd test: using the intradermal tuberculin test on all animals older than 3 months 
Introduction 
The Australian brushtail possum (Trichosurus vulpecula) was introduced to New Zealand from 
Australia from 1 840 to establish a fur trade. Since then the number has increased to a peak of 60-70 
million. As a consequence possums have become a national pest, impacting on native flora and 
fauna (Howard, 1964), and acting as a reservoir and vector of bovine tuberculosis (Morris et aI., 
1994; lackson, 1995). Consequently TB eradication programmes in many areas have been 
unsuccessful and expensive, necessitating control of infected feral animals as well as tuberculous 
cattle and deer on farms. In 1996/97 $26 million were spent on possum control, with an expected 
$30 million by the year 2000 (O'Neil and Pharo, 1995). The funding for TB management and 
associated vector control programmes are obtained through levies and grants from the farming 
industries, regional and district councils, and Government (Coleman and Livingstone, 2000). As a 
result of these programmes the number of cattle and deer herds under movement control for TB has 
decreased (Pannett, 1995; Livingstone, 1997). However, in many areas tuberculosis is endemic in 
feral possum populations and requires continuous control efforts to keep tuberculosis incidence in 
livestock animals at low levels. These areas where TB has been identified in feral vector 
populations, are termed Vector Control areas, for which stringent TB control regulations are in 
place, such as annual TB testing of all animals. All animals suspect of being infected with 
tuberculosis have to be removed to slaughter. While a herd is classified as being Infected with TB, 
213  
i t  remains under Movement control (MC), and all animals sold to market from these herds have to 
be identified with an official AgriQuality Movement Control Tag ( 'white-tagged' animals). These 
white-tagged animals can often only be sold at a discount, as their market is limited and farmers 
from Vector Free Areas are reluctant to take the risk. For more detailed infonnation on MC 
restrictions and TB related regulations see Appendix ID (p. 334). 
Need for Economic Analysis 
As more herds come off MC, emphasis has to be put on the remaining infected fanns. Control of 
TB on these farms will most likely involve a combination of available measures. The marginal 
revenue for each dollar spent in large-scale possum control operations will decrease with 
decreasing number of herds on MC in that area. Therefore, it will be more difficult to expect 
substantial Government funding and farmers increasingly will be expected to assist more in TB 
control, or else pay increasingly for control work through rates .  
At farm level, grazing management and localised possum control are the two main control 
measures currently available for farmers to implement on their own. Their effectiveness has been 
evaluated in Chapters 3 and 4 of this thesis. Management that avoids grazing TB hot-spot areas 
with cattle and deer at certain times of the year, together with localised possum control before 
livestock are put into these paddocks, are expected to reduce the direct contact between infected, 
moribund possums and livestock. This contact was found to be the most likely way of transmitting 
the disease from infected possums to livestock (Sauter and Morris, 1 995). By preventing this direct 
contact from happening, the risk of direct transmission of M.bovis from possums to cattle and deer 
was expected to be reduced and farms were more likely to get off MC. However, current farming is 
very economics-orientated, and therefore it was necessary to evaluate the profitability of such 
control measures on various types of fanns. It was found that the decision to change or adopt 
certain measures was strongly dependent on the financial pay-off (Morris et aI. , 1 995; Cullen and 
Bicknell, 2000). 
The needs of the fanners have to be met through planned disease prevention and control 
programmes, involving a range of measures, each of them offering a different degree of protection 
and requiring a different level of investment. Determining the optimum input and control level, 
therefore, is to a large extent a matter of economic decision making. This applies not only to the 
individual livestock owner, but also to national Government (Dijkhuizen and Morris, 1997). 
214 
Techniques available for analysis 
Partial budgeting is one technique available for economic analysis. It is a quantification of the 
economic consequences of a specific change in fann procedure, e.g. , a herd health programme. It is 
particularly useful for analysing relatively small changes in the business. Partial budgeting only 
considers those items of returns and costs that change as a consequence and estimates the 
difference in profit expected from the alternative option versus the base situation. It does not 
calculate the total income and the total expense for each of the options. 
A partial budget i made up of four sections:  ( 1 )  additional returns: a list of items of returns from 
the alternate plan that will not be received from the base plan, (2) reduced costs: a list of items of 
costs for the base plan that will be avoided with the alternate plan, (3) returns foregone: a list of 
items of returns from the base plan that will not be received from the alternate plan, and (4) extra 
costs: a list of items of costs of the alternate plan that are not required with the base plan 
(Dijkhuizen and Morris, 1996). All added costs and returns and reduced costs and returns are 
assumed to be caused by the introduction of the control measures. 
Aim of this study 
The objective of this analysis was to determine the economic viability of on-farm TB control 
measures under current policies and to suggest policy changes that might increase the cost­
effectiveness of these programmes. If there is an economic benefit, farmers will be interested in 
adopting on-fann control measures and therefore assisting TB control programmes. 
Insurance 
Following the economic analysis of different options the possibility of an insurance scheme was . 
discussed. In order to insure against an adverse event, the risk of this event occurring has to be 
measurable. An insurance company is making a prediction about losses that are expected to occur 
and will estimate the range of error; that way the risk is measurable. The risk for the insurance 
company is that the prediction is not accurate (Vaughan and Vaughan, 1999). The probability of 
losses occurring is one part in the measurement of risk, the potential severity and size of the loss is 
the other part. Both can be taken into account when using the concept of expected values. An 
expected value is calculated by multiplying the probability of the event happening by the amount of 
the potential loss (Vaughan and Vaughan, 1999). Other terms used in the insurance business 
include 'peril '  and 'hazard' . A peril is the actual cause of the loss, e.g. peril of hail, whereas a 
hazard is a situation that creates or increases the chance of a loss, resulting from a given peril. 
There are some situations that are both peril and hazard, such as disease. A disease is a peril 
215 
causing direct economic loss, but i t  also is  a hazard for premature culling because disease-affected 
animals may no longer justify their place in the herd. 
There are several ways to deal with risks: Some fundamental risks are dealt with through society 
and Government (e.g. police), whereas individuals deal with some particular risks by avoiding 
them, retaining, transferring, sharing and reducing the risk (Vaughan and Vaughan, 1 999). For our 
case with tuberculosis, risk sharing (e.g. within an insurance group) and risk reduction (e.g. 
prevention of loss or controlling the severity of it, such as possum control) are most important. 
Insurance creates security, it does not remove or reduce the risk, but it reduces the probability of 
severe financial loss for the individual. Insurance has two main characteristics, the first is 
transferring/shifting risk from an individual to a group and the second is the sharing of losses 
amongst all group members, whereby the burden of financial loss is spread amongst the group. 
This method of loss distribution is the basis on which insurance can exist (Vaughan and Vaughan, 
1999). 
Insurance agreements can work in two ways, either where the members of the group share the loss 
after it has occurred or on advance premium basis, whereby the members of the group pay their 
share in advance. For the second method, probability theory and predictions are of importance. By 
agglomerating a large number of homogeneous individuals/units that are exposed to the same risk, 
the insurance company can predict (using probability theory) the amount of losses that will occur in 
the group, something an individual is not realistically able to do. To achieve this prediction the 
insurance company is using past experience and establishes empirical probabilities (a posteriori 
probability). However, as these empirical probabilities will not be exactly accurate, a margin for 
error is allowed (e.g. on the basis of standard deviation of the past experiences). The more past 
events are investigated to obtain the prediction, the more accurate this prediction is. Equally, the 
more subjects are included in the group, the better the prediction fits to this popUlation (due to the 
law of large numbers) .  
In  order for a risk to be insurable some prerequisites should be met: There should be a large enough 
number of homogeneous exposed units (important for prediction); the loss should be measurable in 
terms of finances; there should be an uncertainty to whether the loss occurs or not; and the loss 
should not be catastrophic (affecting a very large percentage of insured subjects at any one time) 
(Vaughan and Vaughan, 1 999). Additionally, an insurable risk should also fulfil criteria with 
regards to moral hazard and adverse selection. Moral hazard results from dishonest characteristics 
of an insured person, such as would be the case if an individual buys an insurance policy and then 
deliberately changes behaviour to increase the magnitude or the probability of a loss. Deductibles, 
co-payment and checking the insured' s  behaviour are some ways of dealing with moral hazard. 
216 
Adverse selection occurs if individuals have a better knowledge about the magnitude or probability 
of their losses than the insurer, e.g. farmers who have an above average risk are more likely to 
obtain insurance. The way to deal with this problem is by classifying the insured persons/farms into 
different risk categories based on detailed information on magnitude and probability of loss 
(Meuwissen et aI. , 1997). 
An insurance rate is charged on a per unit basis. Insurance premiums are calculated by multiplying 
the rate by the number of units insured. The rate is usually determined from the cost of production, 
but as these are unknown the rates must be based on an estimation of future losses and costs, which 
is called rate-making (Vaughan and Vaughan, 1 999). The p emium can be divided into pure 
premium, which is intended to cover the losses, and the loading, which covers the expenses of 
operating the system. 
There are different types of rates: Class rates are most common in the insurance business, they are 
applied to all subjects that have the same set of features/qualities. Individual rates are used when 
the features of the subjects differ too much and the loss of production has to be calculated on an 
individual basis. 
Most insurance schemes available for agriculture deal with the property, crop, natural risks such as 
flood or hail, or social risks (e.g. burglary, accidents) (Ray, 1981 ). There are no examples of using 
insurance in animal production in New Zealand, and only limited examples from overseas with 
regard to the management of a contagious disease in animals. In Germany livestock insurance and 
State animal disease insurance are available to farmers (Siemienkowicz, 1984). In the Netherlands 
studies were conducted on providing insurance for classical swine fever (Meuwissen et aI. , 1997). 
Materials and Methods 
In the partial budgeting a farm without an on-farm TB control programme and with a TB problem 
was taken as the baseline for comparison. Farming with an on-farm TB control programme (as 
described in Chapter 3) and consequently a reduction of the TB problem was used as the 
alternative. The different costs and revenues were determined in consultation with experts from 
AgriQuality and experts from the farm consulting company Baker & Associates, Masterton. 
Representative farms included in the study 
Three representative farm types were used in the analysis: dairy farms, sheep and beef breeding, 
and sheep and beef finishing farms, the three types most common in the Wairarapa, a TB Vector 
217  
Control Area (VRA). It was assumed that possums were the main vector species on these farms, 
and that the role of ferrets or other infected wildlife species was of minor importance. 
Some characteristics of the typical farms used in the economic analysis are presented in Table 48 
(Ministry of Agriculture and Forestry, 1998). More details on farming operations in New Zealand 
can be found in Parker (1997) and in publications from Livestock Improvement Corporation 
Limited ( 1998). 
Table 48. Some characteristics of farms in the Hawke's Bay-Wairarapa District (from 
Ministry of Agriculture and Forestry, 1998). 
Dairy herds Beef breeding herds Beef fInishing herds 
Total herds 226 680 570 
Average herd size 243 cowslherd 2625 sheep stock units 1 682 sheep stock units 
664 cattle stock units 684 cattle stock units 
Stocking rate (in Stock Units/ha) 7.68 9.86 
Replacement rate 20 20 
Average effective hectares 98 428 240 
Average kg milksolids produced! 71 ,407 
farm 
Average Dairy company payout ($/kg 3.63 
milksolids) 
Revenues for a sheep and beef breeding farm was generated from sheep through wool production 
and lamb sales and from cattle through sales of cull cows, weaner steers and weaner heifers (rising 
one-year old animals). Revenues for sheep and beef finishing farm through cattle consisted of 
selling the fattened animals to the slaughterhouses. Sheep provided often 50% or more of the total 
farm income (Parker, 1 997). 
Tuberculous animals 
For the purpose of the economic analyses it was assumed that all tuberculous animals were 
identified through on-farm TB testing. It was assumed that no animals were found tuberculous at 
slaughter that had tested negatively in the previous test or were not tested at all. It was assumed that 
beef finishing farms tested all their cattle before sending them to slaughter. None of these herds 
was assumed to have the herd status 'Works monitored' , which can be obtained if over 90% of the 
herd is slaughtered annually. In herds with this status, no on-farm TB testing is required (see 
Appendix ill for more details, p. 334). 
218  
All animals reacting positively to the skin test are eligible for compensation, which i s  currently set 
at 65% of the 'fair market value' .  Animals that had not been TB skin tested prior to slaughter, but 
were found to be TB lesion positive at slaughter are not eligible for compensation. The farmer 
receives the slaughter value of that animal (which can only be sold for local trade, not for the 
export market). If the animal is condemned due to extensive lesions, the farmer does not receive 
any revenue from that cattle beast. Most farmers in the Wairarapa TB test their animals regularly 
and almost all tuberculous animals are identified by TB skin testing. Therefore the analysis 
assumed that all infected animals were found by TB testing. 
Details of costs and revenues used in the partial budgeting 
A summary of the assumptions and data used in the partial budgeting for the different herd types is 
presented in Table 49. 
On sheep and beef farms TB had only implications for cattle, not sheep. TB in sheep has only been 
recorded very seldom (Davidson et aI. , 198 1 ;  Cordes et aI., 1981).  Furthermore they did not expose 
themselves to 'infection' by close contact with possums to the same degree as cattle (Sauter and 
Morris, 1995), which means that sheep can be used to graze the TB hot-spot areas, thus utilising 
grass production in these areas. 
Additional returns resulting from the implementation of control measures 
Increased animal value (live animals) 
All animals sold live from a TB infected herd, have to be identified with white tags and can often 
only be sold at an estimated discount of 10% for dairy herds and 15% for beef herds from 'fair 
market price' (Pannett pers. comm., 1999). This increased animal value only applied if the herd 
achieved an off-MC status. 
Increased milk production 
This additional return was only applicable to dairy herds. Tuberculous cows had to be slaughtered 
as soon as possible after the positive reaction to the intradermal tuberculin test. Consequently milk 
production was lost if the animals were removed during lactation. If the on-farm control measures 
reduced reactor numbers, this increased milk production. It was assumed that whenever 5 reactors 
were removed from the average farm (243 milking cows) this caused a 1 % reduction on the overall 
milk yield for the current season (Pannett pers. comm., 1999). 
2 19  
Reduced costs resulting from the implementation of control measures 
Premature disposal due to TB 
All cattle suspected to be tuberculous, based on the national TB testing programme, must be 
slaughtered. Currently farmers are compensated for reactor cattle at 65% of 'fair market value' 
(determined by AgriQuality), irrespective of the 'true' TB status of the animal, determined at 
slaughter. Therefore a cost of 35% per reactor animal could be saved if the number of reactors was 
reduced due to the implementation of control programmes. Under current control policies it was 
considered unlikely that any cattle would die of TB, although in the absence of control measures 
this would occur. 
Additional costs resulting from the implementation of control measures 
Localised possum control 
Localised possum control was concentrated on five months per year (October to December plus 
March and April) and on areas with known or potential TB hot-spots. It can be done by using bait 
stations or traps. Poison and labour needed for possum control were additional costs. 
Due to larger farm sizes and rougher vegetation, these cost drivers were expected to be higher for 
beef farms than for dairy farms. On an average dairy farm it was estimated that 15 bait stations 
($ 1 0  each) were needed. Depreciation of bait stations was attained in five years, resulting in a cost 
of $30 per year. About $ 1 50 will be spend annually on poison on an average dairy farm. 
On an average sheep and beef breeding farm the localised possum control was expected to require 
an investment of $ 1 ,250 for bait stations and traps, assuming that 1 6% of the total farm area was 
possum-denning habitat, and 50% of this area was classified as hot-spots area (McKenzie, 1 999). 
Therefore, 34 hectares had to be covered in the localised possum control programme. For an 
effective possum control it was estimated that 50 bait stations at $ 10  each and 1 0  Timms-traps 
(killing traps) at $35 each had to be purchased (Pannett pers. COIDm., 1999). With a depreciation 
period of five years, this resulted in an annual cost of $ 170. Furthermore, it was estimated that $700 
will be spent annually on purchasing bait and poison. For a beef finishing farm the costs were about 
half of the costs for beef breeding farms, due to their farm size. A total cost of $430 a year was 
estimated. 
The additional labour needed for possum control was included in the additional management 
needed for the grazing management strategies. 
220 
Additional management 
The localisation of TB hot-spots on the farm for targeted control measures, managing the grazing 
strategies (using grazing records) and localised possum control all require additional management. 
Once TB hot-spot areas are identified on a farm, this part requires only limited time input. Keeping 
grazing records, setting up strategies, and conducting localised possum control once or twice a 
year, required the main time input. 
Most dairy farmers already have detailed grazing strategies in place. Therefore, exclusion of hot­
spots during certain times of the year could easily be incorporated into their existing farm 
management. Possum control was expected to take four hours per month. The adju tment of the 
grazing strategy was expected to involve one hour per week (Pannett pers. comm., 1999). It was 
expected that 100 hours of additional labour per year were sufficient to cover the time requirements 
for on-farm control programmes on an average dairy farm. 
On a beef breeding farm it was estimated that a farmer had to spend 60 hours annually on 
poisoning and trapping possums (twice a year three weeks trapping/poisoning, with 3 hrs every 
second day = 54 hr/annum). The additional labour needed for the grazing management was 
expected to be 50 hours annually, focusing on fencing and shifting cattle. The total time spent on a 
beef finishing farm (about half the size of a beef breeding farm) was estimated to be 55 hrs a year. 
Additional fencing 
On dairy farms no additional fencing was required to implement the strategy. 
Paddock sizes on beef farms are generally significantly bigger than on dairy farms. In order to 
minimise the area to be excluded from grazing, additional fencing is required around TB hot-spot 
areas. The main costs involved are fencing material and labour; the equipment needed for fencing 
is already available on most farms. Fencing expenses for a beef breeding farm were estimated at a 
total of $5,000, excluding the labour involved, which was accounted for in the management costs. 
As the fencing costs only incur once, an annual cost of $1,000 was obtained with a depreciation 
period of five years. The costs for a beef finishing farm were estimated at $500 per year ($2,500 
investment, depreciated in five years). 
Returns foregone as a result of the introduction of control measures 
No returns foregone were identified in the analysis. 
221 
Parameters used in the economic analysis 
Deterministic model 
A deterministic partial budget model was built in Microsoft Excel (Microsoft Excel for Windows 
97, Version 8.0, Microsoft Corporation, Redmond, WA). Table 49 provides a summary of the input 
variables as detailed in the previous section and used in the analyses for the three different farm 
types. 
Table 49. Summary of data and assumptions used in partial budgeting of implementing on· 
farm control programme on the three farm types in the Wairarapa. 
Dairy fann Beef breeding Beef finishing 
I ncreased milk production (kg milksolidsl 1 42.8 
cow not having to be slaughtered) 
Discount TB reactors (slaughter) 35% 35% 35% 
Annual cattle sales 50 cows 1 7  cows 70 slaughter 
6 R2yr heifers 42 weaner steers beef 
1 0  R1yr heifers 21 weaner heifers 
Discount 'white tagged' (live animals) 1 0% 1 5% 
Animal value > 2yrs $ 850 $ 695 $ 970 
Animal value 1 - 2 yrs $ 750 $ 695 $ 695 
Animal value 6wks - 1 yr $ 375 $ 315  
steers to 1 8mths $ 495 
Additional labour cost 100 hrslyr 1 1 0 hrslyr 55 hrs/yr 
Additional fencing cost $ 1 ,OOO/yr $5001yr 
Trapping/poisoning cost $180/yr $8701yr $430/yr 
In estimating the annual cattle sales a 20% replacement rate was assumed. To keep the results of 
the economic analyses comparable it was assumed that the TB reactors were dairy cows older than 
two years, beef breeding cattle older than one year of age, and beef finishing steers and non­
breeding bulls older than 1 8  months. The market value of the reactor animals was determined from 
the payout scheme of AgriQuality, paid for reactor animals .  For each animal class a maximum 
value is set for the amount which AgriQuality will be able to value the animal for compensation. 
This maximum value was used in the deterministic model. 
The spreadsheets used for the analysis in these three farm types are shown in Appendix V (p. 35 1). 
222 
In the economic analysis six scenarios, regarding the reduction in reactor numbers, achieved by the 
implementation of on-fann control measures, were considered for the three fann types separately 
(Table 50). 
Table 50. The six scenarios analysed in the study. 
Number of reactors in Number of reactors in Achieve Off-Movement 
previous year following year Control status 
Scenario 1 5 2 No 
Scenario 2 5 O.5a No 
Scenario 3 5 0 Yes 
Scenario 4 2 O.5a No 
Scenario 5 2 0 Yes 
Scenario 6 0 Yes 
aone reactor every second year 
The goal of implementing the two control measures (grazing management system and localised 
possum control) was to reduce the incidence of TB in livestock on farms, with a final goal of the 
herd achieving a TB free herd status. The analysis was therefore intended to examine the different 
scenarios described above and show their economic consequences. 
In addition the economic consequences were also evaluated under a range of different reactor cattle 
compensation levels, ranging from the current 65% to zero. 
Stochastic model 
In addition to the deterministic spreadsheet model, a stochastic economic analysis was performed 
using a risk analysis and simulation add-in for Microsoft Excel ( @ RISK, Version 3.5 .2, Palisade 
Corporation, Newfield, New York, USA). This Monte-Carlo simulation package allowed input 
variables (costs and returns) to be modelled as distributions rather than deterministic fixed values. 
Input distributions such as price of animals were sampled at each iteration. These sampled values 
were then used in the economic analysis to calculate the economic benefit for each iteration. The 
outputs were summarised across iterations at the end of the simulation and could be presented as 
histograms of expected economic returns. The simulation was conducted using 1000 iterations 
generated by Latin hyper-cube sampling, a sampling method that selects random numbers from the 
specified distributions (Vose, 1996). 
223 
Table 5 1  provides a summary of the input variables, their distributions and parameters, as used in 
the stochastic model for the three farm types. 
Cattle prices were determined from the maximum payout scheme for reactor cattle used by 
AgriQuality. This group of variables was assumed to be normally distributed with a maximum 
value, therefore a truncated normal distribution was used in the model (Table 5 1 ). 
Discount percentage on white-tagged animals was elaborated with representatives from 
AgriQuality and Animal Health Board. As the value of 1 0% for dairy and 1 5% for beef farms was 
considered the best estimate, a truncated normal distribution with a very small standard deviation 
was used in the model (Table 5 1 ). 
For dairy farms the total milk yield in kg milksolids was estimated from the published average milk 
yield per farm in the Wairarapa. This variable was assumed to be normally distributed with a 
maximum and minimum cut-off value, therefore indicating a truncated normal distribution (Table 
5 1 ). 
The price per kg milksolids for dairy farms was estimated from published data and modelled with a 
truncated normal distribution with the mean of $3.63, paid out in the financial year 1 998/99 (Table 
5 1 ). 
The hourly cost of labour was modelled as a truncated normal distribution with the mean expected 
value of $ 1 2  (Table 5 1 ). 
For the amount of time spent on grazing management and possum control, the number elaborated 
for the deterministic model was assumed to be the most commonly observed one. On this basis, a 
triangular distribution with maximum, minimum and mode parameters was used (Table 5 1). 
The distribution for the cost of fencing for beef breeding and beef finishing farms was assumed to 
be triangular with minimum, maximum and mode parameters specified as presented in Table 5 1 .  
The cost of poison and bait stations for each of the three farm types was elaborated with 
representatives of AgriQuality and farm consultants and was modelled as a triangular distribution, 
as it was reasonable to assume minimum and maximum values and assign the mode to the amount 
specified by the experts (Table 5 1 ). 
224 
Table 51. Distribution parameters for input variables for costs and returns used in the 
stochastic @ RISK partial budgeting model for the three farm types in the Wairarapa. 
Herd- Variable Distribution Mode Min Max Mean SD 
OH Animal value >2yr ($) Truncated normal 750 850 800 40 
OH Animal value 1 -2 yr ($) Truncated normal 450 750 700 40 
OH Animal value 6wks - 1 yr ($) Truncated normal 200 375 325 25 
OH % discount white-tag Truncated normal 8 1 5  1 0  2 
OH Milk yield (kg MS) Truncated normal 65,000 78,000 71 ,407 7000 
OH Price /kg MS ($) runcated normal 3.4 3.8 3.63 0.4 
OH Additional time required (hrs) Triangular 100 80 120 
OH Cost of PoisorvTraps ($) Triangular 1 80 1 50 200 
BB Animal value breeding > 1 yr Truncated normal 500 695 650 40 
BB Animal value breeding 6wks Truncated normal 200 315 270 40 
- 1yr 
BB Animal value steers 6wks - Truncated normal 350 495 450 25 
1 8  mths 
BB % discount (white-tag) Truncated normal 1 0  1 8  1 5  2 
BB Additional time required (hrs) Triangular 1 10 90 1 30 
BB Additional fencing cost ($) Triangular 1000 500 1 200 
BB Cost of PoisorvTraps ($) Triangular 870 650 1 1 00 
BD Steers > 1 8mths Truncated normal 750 970 900 50 
BD Additional time required (hrs) Triangular 55 45 65 
BD Additional fencing cost ($) Triangular 500 250 650 
BD Cost of PoisonfTraps ($) Triangular 430 350 550 
DHlBB Labour I hr  Truncated normal 7 20 1 2  2 
IBD 
aHerd type: DH: dairy herd; BB: beef breeding herd; BD: beef finishing herd 
Output variables included return minus costs for each of the six scenarios described above. The 
spreadsheets used in the analysis were the same as for the detenninistic model (Appendix V, p. 
351) .  Sensitivity analyses were performed for each output in order to determine the most important 
input variables. The rank correlation coefficient was used to measure sensitivity in an output to 
variation in the individual input distributions. 
225 
Decision analysis  
As a third step in the economic analysis, decision tree analysis was performed. For each of  the 
three farm types decision trees were constructed, using an add-in for Microsoft Excel 
(PrecisionTree, Version l .Oa, Palisade Corporation, Newfield, New York, USA). The expected 
economic outcomes of the deterministic model and a range of probabilities for achieving the 
different scenarios (Table 52) were used in the decision tree analysis. The estimates for the 
probabilities where on-farm control has been adopted were taken (as far as possible) from the 
actual data obtained in the 'focused control' farms of the study described in Chapter 3 .  However, 
especially for the scenarios starting with five reactors the actual data was less extensive than for the 
other scenarios (only five farms had five or more reactors). For the situation where no on-farm 
control programme was adopted, estimates for the probabilities were taken and adjusted from the 
'standard control' farms in Chapter 3. The actual probabilities found in the intervention study could 
not be taken unadjusted, as some standard control farms also had their own on-farm control 
programmes in place. The probability of having the same number of reactors the following year 
could not be assumed to be 1 .0 for the situation where farmers do not adopt the control measures, 
as most areas of the Wairarapa are subject to regional possum controL The costs of the current 
slaughter-levy for adult cattle was included in this part of the analysis in order to compare the 
current levy ($7.20) with the increased one ($1 3 .00), suggested in the new pest management 
strategy of the AHB. 
226 
Table 52. Assumed probabilities of reducing reactor numbers per farm if conducting on-farm 
TB control or not. 
Reactor numbers in two Probability if adopting Probability if not adopting 
consecutive years on-farm control on-farm control 
Original Adjusted" 
5 reactors to 6 reactors/yr 0.08 0 . 10 0. 1 5  
5 reactors to 5 reactors/yr 0. 1 5  0.40 0.45 
5 reactors to 2 reactors/yr 0.31 0 . 10 0.20 
5 reactorslyr to 1 every 2nd yr 0.31 0.20 0. 1 0  
5 reactors to O/yr 0. 1 5  0.20 0. 1 0  
2 reactors to 5 reactorslyr 0.0 0.21 0.25 
2 reactors to 2 reactors/yr 0. 1 2  0.21 0.25 
2 reactorslyr to 1 every 2nd yr 0.50 0.36 0.35 
2 reactors to O/yr 0.38 0.21 0. 1 5  
1 reactor to 2 reactorslyr 0. 1 2  0.28 0.30 
1 reactor to 1 reactor/yr 0.23 0.25 0.30 
1 reactor to O/yr 0.65 0.47 0.40 
" To take account of standard control farms which undertook on-fann control by fanner choice. 
Results 
In the following 'net gain' or 'net loss' are tenns used for the monetary value, resulting from 
adding the costs and returns incurred from the implementation of on-farm control measures for TB. 
Deterministic model 
Current s ituation (65% compensation) 
All three types of farms were analysed and the economic consequences are presented in Table 53. 
For example the benefit of reducing the number of reactors from five per year to two per year is 
described in the first row. 
227 
Table 53. Expected economic outcomes of the partial budgeting for reducing the number of 
reactors in three different farm types (using current compensation of 65 %). 
Dairy Beef breeding Beef finishing 
From 5/yr to 2/yr $ 1 ,068 - $ 2,460 - $ 572 
From 5/yr to 1 /2nd yr $ 2,292 - $ 2,095 - $ 62 
From 5/yr to off Mea $ 7,775 $ 3,805 $ 1 08 
From 2/yr to 1/2nd yr - $ 1 56 - $ 2,825 - $ 1 ,081 
From 2/yr to off Mea $ 5,327 $ 3,075 - $ 91 1 
From 1/yr to off Mea $ 4,51 1 $ 2,832 - $ 1 ,251 
'MC = Movement Control 
The economic outcomes show that the implementation of on-farm control measures on a dairy farm 
resulted mostly in a net gain under the current compensation scheme (65%).  Only in the scenario 
where the herd started with two reactors a year and reduced the reactor number to one every second 
year, did the implementation result in a net loss. The herd still incurred the discount of white­
tagged animals, and the reduction in reactor number was not sufficiently large to equal the costs of 
the programme. If the herd came off MC the prices for cull cows and other cattle sales increased, 
causing relatively large benefits in the economic outcomes (Table 53).  
Another retum was the increased overall milk yield, as the cows did not have to be removed from 
the farm in the middle of their lactation any more. The influence of a milk production increase was 
considerable, especially when a farm managed to reduce its number of reactors from five per year 
to zero and get off MC. The increase of 1 % in milk production calculated to result from this 
reduction in reactor numbers resulted in additional revenue of almost $2,600. 
The introduction of an on-farm TB control programme on a sheep and beef breeding farm under the 
current compensation scheme (65%) resulted in an economic advantage in those situations where 
the herd got off MC, irrespective of the number of reactors it had previously. If the herd did not get 
off MC, then only if it had a reduction of at least 14 reactors, was the break-even point reached -
with an expected net benefit of +$2 16. 
In contrast to dairy and sheep and beef breeding farms, the introduction of an on-farm TB control 
programme under current policies (65% compensation) on to a sheep and beef finishing farm 
resulted generally in a net loss. It was only economically beneficial if the herd reduced its reactor 
numbers by at least 5 reactors per year (break-even point) (Table 53). 
228 
Alternative situations with reduced or no compensation for reactor animals 
A significant source of controversy is  whether or not the level of compensation for reactor animals 
should be reduced from 65%. The deterministic analysis model was run for different levels of 
reduced compensation, ranging from 60% to 0%, keeping all other variables equal to the original 
analysis. Figure 43, Figure 44, and Figure 45 show the expected net returns for each of the 
compensation levels for the three herd types; the scale of the y-axis (expected net return) was kept 
the same in all three figures, in order to facilitate comparison between the three herd types. From 
Figure 43, Figure 44, and Figure 45 it can be seen that the beef finishing farms have a lower 
expected net return than beef breeding and dairy farms. The expected net return for beef breeding 
farms is on average $3,274 lower than for dairy farms, and the expected net return for beef 
finishing farms is on average $4,030 lower than for dairy farms. Table 54 provides the outcomes of 
the analysis for zero compensation. 
1 2  
c � 9 :::J ... 
0) -
� �  
'Q) o  6 
C O  'C �  
.! c 3 (,) .-0) -
-
---
C. ..... 
>< 0 w /' 
65 
-3 
Dairy herd 
-
� "" ... � 
� 
.::.E 
60 50 40 20 0 
compensation level 
Scenario: 
--+-5 -> 2 
---- 5 -> 0 .5 
5 -> 0 
�2 -> 0.5  
�2 -> 0  
--- 1 -> 0 
Figure 43. Expected net returns on dairy farms for different reductions in reactor numbers 
using different compensation levels for reactor animals. 
c:: 1 2  � 
:::J 9 .... e -� 
"" 0  6 <» 0  C:: O ,, - 3 <» .... c:: (.) .-<» 0 Q. 
-
>< 
w -3 
-
65 
Beef breeding herd 
- � 
60 50 40 20 
compensation level 
-
� 
� 
� 
o 
Scenario: 
-+-5  -> 2 
- 5 -> 0.5 
5 -> 0 
--*-2 -> 0.5 
-*-2 -> 0 
""""- 1 -> 0 
229 
Figure 44. Expected net returns on beef breeding farms for different reductions in reactor 
numbers using different compensation levels for reactor animals. 
c; 1 2  � 
:::J .... 9 Cl) _ ... � 
Q) o  6 C; O  ,, �  3 .s c:: (.) .-Cl) 0 C. �'-
>< -
w -3 
65 
Beef fi nishing herd 
�-� , .---
J\, �� -:: 
-
60 so 40 20 o 
compensation level 
Scenario: 
-+-s -> 2 
-- S -> O.S 
S -> 0 
�2 -> 0.S 
-*-2 -> 0 
""""- 1 -> 0 
Figure 45. Expected net returns on beef finishing farms for different reductions in reactor 
numbers using different compensation levels for reactor animals. 
230 
Table 54. Expected economic outcomes in partial budgeting of reducing the number of 
reactors in the three different farm types with zero compensation 
Dairy Beef breeding Beef finishing 
From 5/yr to 2/yr $ 2 ,725 - $ 1 , 1 05 $ 1 ,320 
From 5/yr to 1 /2nd yr $ 4,778 - $ 63 $ 2,775 
From 5/yr to off MC $1 0,537 $ 6,064 $ 3,260 
From 2/yr to 1/2nd yr $ 673 - $ 2, 1 47 - $ 1 35 
From 2/yr to off MC $ 6,432 $ 3,979 $ 350 
From 1/yr to off MC $ 5 ,063 $ 3,284 - $ 620 
The reduction of compensation level from 65% to zero compensation resulted in an increased net 
gain or at least a reduced net loss for some situations for all scenarios and farm types. For dairy 
farmers it was always economically beneficial to implement on-farm control measures, for beef 
breeding farms it was beneficial if the herd came off MC, but not beneficial if the herd still stayed 
under MC. For beef finishing farms the carrying out of on-farm control measures was beneficial if 
their reduction of reactor animals was at least two animals per year. 
Alternative situation with subsidies on control costs 
In the following, a situation was modelled where a subsidy was paid to cover the costs of poison 
and bait stations . The costs of labour and fencing were still incurring. The expected net returns 
were calculated for three different compensation levels (65%, 40%, and 0%). Table 55 ,  Table 56, 
and Table 57 present the expected net returns for the three farm types for the different scenarios 
under these circumstances. As many of these expected net returns are still net losses for beef 
farmers, one can calculate the additional subsidy necessary to achieve net gains in all,  or almost all 
situations. 
23 1 
Table 55. Expected net returns with subsidy to cover costs of poison and bait stations for beef 
breeding farms and additional subsidies necessary to achieve net gain in all and all except one 
scenarios. 
Beef breeding 65% comp. 40% comp. Zero comp. 
with a subsidy of $950/yr 
Scenario 1 (5 -> 2) - $ 1 ,590 - $ 1 ,069 - $ 235 
Scenario 2 (5 -> 0.5) - $ 1 ,225 - $ 444 $ 808 
Scenario 3 (5 -> 0) $ 4,675 $ 5,544 $ 6,934 
Scenario 4 (2 -> 0.5) - $ 1 ,955 - $ 1 ,695 - $ 1 ,278 
Scenario 5 (2 -> 0) $ 3,945 $ 4,293 $ 4,849 
Scenario 6 (1 -> 0) $ 3,702 $ 3,876 $ 4, 1 54 
Additional subsidy necessary for all 
scenarios resulting in net gains $ 1 ,960 $ 1 ,700 $ 1 ,280 
Additional subsidy necessary for all 
except one scenario resulting in net gains $ 1 ,600 $ 1 ,070 $ 240 
Table 56. Expected net returns with subsidy to cover costs of poison and bait stations for beef 
finishing farms and additional subsidies necessary to achieve net gain in all and all except one 
scenarios. 
Beef finishing 65% comp. 40% comp. Zero comp. 
with a subsidy of $450/yr 
Scenario 1 (5 -> 2) - $ 1 42 $ 586 $ 1 ,750 
Scenario 2 (5 -> 0.5) $ 368 $ 1 ,459 $ 3,205 
Scenario 3 (5 -> 0) $ 538 $ 1 ,750 $ 3,690 
Scenario 4 (2 -> 0.5) • $ 651 - $ 287 $ 295 
Scenario 5 (2 -> 0) - $ 481 $ 4  $ 780 
Scenario 6 (1 -> 0) - $ 821 - $ 578 - $ 1 90 
Additional subsidy necessary for all 
scenarios resulting in net gains $ 820 $ 580 $ 190 
Additional subsidy necessary for all 
except one scenario resulting in net gains $ 480 $ 290 $ 0  
232 
Table 57. Expected net returns with subsidy to cover costs of poison and bait stations for 
dairy farms and additional subsidies necessary to achieve net gain in all and all except one 
scenarios. 
Dairy herds 65% comp. 40% comp. Zero comp. 
with a subsidy of $180/yr 
Scenario 1 (5 -> 2) $ 1 ,248 $ 1 ,885 $ 2,905 
Scenario 2 (5 -> 0.5) $ 2,472 $ 3,428 $ 4,958 
Scenario 3 (5 -> 0) $ 7,955 $ 9,017 $ 1 0,71 7 
Scenario 4 (2 -> 0.5) $ 24 $ 343 $ 853 
Scenario 5 (2 -> 0) $ 5,507 $ 5,932 $ 6,61 2 
Scenario 6 (1 -> 0) $ 4,691 $ 4,903 $ 5,243 
Additional subsidy necessary for all No additional subsidies necessary, all net returns 
scenarios resulting in net gains are net gains when cost of poison is covered 
Additional subsidy necessary for all No additional subsidies necessary, all net returns 
except one scenario resulting in net gains are net gains when cost of pOison is covered 
From Table 57 it can be seen that the dairy farms always achieve a net gain if they do not have to 
pay for poison and bait stations. For the beef farms, especially the beef breeding farms it is more 
difficult to achieve a net gain, even if the costs of poison and bait stations are covered. The beef 
breeding farms require additionally between $ 1 ,300 and $2,000 to achieve a net gain in the 
scenario with the worst financial net return, where the farm reduces its reactor numbers from two 
per year to one every second year. Beef finishing farms only require support of between $200 and 
$800 to achieve a net gain in all scenarios. 
Break-even points 
In order to find the break-even points, at which the amount of money spent on on-farm control 
measures was equal to the benefits of the programme, the additional revenues were calculated 
independently of the costs. For dairy herds under the current compensation scheme (65%) the 
additional revenues were found to be $ 8 1 6  per one animal reduction in reactor numbers plus an 
additional $5,075 if the herd came off MC. For the beef breeding herds the additional revenues 
were only $243 per one animal reduction in reactor numbers plus $5,779 if the herd came off MC. 
The higher value per one animal reduction in dairy herds versus beef herds was due to the increase 
in milk production. Beef finishing farms received additional revenue of $340 for each reactor less 
and no additional revenue if the herd came off MC (Figure 46). 
(f; Il> 
;:, 
c 
Il> _  � fh  
.... N ;;B z  
c c 
.e v -
is 
" 
« 
5000 
4000 
3000 
2000 
1 000 
88 
Herd type 
233 
Additional herd comes MC 
each animal TB 
Figure 46. Additional revenues per animal reduction in reactor numbers and if the herd came 
off Movement Control. 
The costs of the programme were $ 1 ,621 for dairy farms, $2,455 for beef breeding farms, and 
$2, 199 for beef finishing farms. This indicates that for dairy farms and beef breeding farms the 
programme was economic if the herd came off MC. If the herd did not achieve this off-MC status, 
a dairy farm needed a reduction of at least two animals in reactor numbers to make the programme 
financially viable. A beef breeding farm needed a reduction of at least 1 1  animals and beef 
finishing farms a reduction of at least 7 animals to obtain a net gain from adopting the control 
programme if the herd did not get off MC. 
If the imposed levy on slaughtered adult cattle is increased to $13 instead of the current $8.41, one 
suggested element of the new strategy (Animal Health Board, 2000), the costs of the programme 
increases to $1,850 for dairy farms, $2,533 for beef breeding farms, and $2,520 for beef finishing 
farms . If the herd did not get off MC dairy farms and beef finishing farms would require a 
reduction in reactor number of one animal more than under the current levy, whereas beef breeding 
farms would still require a reduction of at least 1 1  animals to make the programme financially 
viable. 
Stochastic model (@RISK) 
Stochastic model on dairy farms 
The stochastic model, using the parameters specified, yielded a histogram of expected net benefits 
obtained with 1000 iterations. Figure 47 presents the histograms of the fmancial output for the six 
234 
scenarios (returns minus cost for each of the six scenarios) for a dairy farm. The distribution 
parameters for the six scenarios are shown in Table 58. 
Table 58. Descriptive statistics for the probability distributions of the difference between 
returns and costs resulting from simulation modelling of the six different scenarios for dairy 
farms. 
TB reactor scenario 
From 5 to 2 From 5 to From 5 to 0 From 2 to 
1I2nd yr 1 /2nd yr 
Mean 1 001 .54 2194.06 7585.82 -1 90.98 
Maximum 1 678. 1 1  2960.91 1021 3.07 395.31 
Minimum 225.20 1296.52 5710.26 -894.34 
5th percentile 601 . 1 8  1760.04 6430.31 -580.74 
95th 1 383.55 2615.69 8965.76 161 .42 
percentile 
Dairy Farms 
from 5 to 2 reactarslyr 
from 5 10 1 reactor every second year 
O. 1 30,.--:---:,-�--: ... -:---:-.., 
� ::J 
! o 0: 0. 
from 5 te zero reaclors/yr 
from 2 10 zero reactorEJyr 
from 1 to lerO reactorslyr 
From 2 to 0 From 1 to 0 
520078 4405.77 
7843.09 7053.10 
3452.36 2699.73 
4073.1 9  3290.12 
6576.17 5764.64 
Figure 47. Economic outcome distributions from @Risk stochastic partial budgeting model 
for returns minus costs for six scenarios of reductions in reactor numbers for a dairy farm. 
235 
Sensitivity analysis showed that returns minus costs for dairy farms whose herds did not come off 
MC, were most sensitive to the cost of labour, followed by the time spent in conducting the on­
farm control measures and the increased milk production. As an example, Figure 48 shows the 
tornado graph obtained for scenario 1 ,  where the reactor number changed from five to two reactors 
per year. For the scenarios where the herd came off MC as a consequence of adopting the control 
measures, the final economic outcome was most sensitive to the chosen discount percentage (for 
white-tagged animals), followed by the hourly rate for labour and then the price of dairy cattle 
older than two years. 
. .113 
IZ�'i.�':'!:lli:i�J,':ij���!::o::4 
., ·.50 .00 
Coefficient Value 
Std b cooff 
calculated a! 
end cf oars 
...•. ........ .. � 
1 .00 
Figure 48. Tornado graph with results of sensitivity analysis showing the importance of 
influence of different input variables for the situation where a dairy farm had 5 reactors and 
reduced it to 2/yr due to the implementation of on-farm control programmes. 
Expected values, 5th and 95th percentiles obtained from the output distributions were compared 
graphically in order to determine whether the differences between the farm types were likely to 
reflect true differences in economic merit or random variation. Figure 49 shows these three 
estimates for each of the scenarios. The net benefit for the three scenarios where the herd came off 
MC (number 3, 5 and 6 in the graph) was much greater than for the scenarios where the herd did 
not come off MC. Scenario 4 (from 2 reactors/yr to 1 every second year) had the least favourable 
monetary outcome. 
236 
tn 1 0000 -tn 8000 0 0 
tn -:1 0- 6000 
C N .- Z 4000 E C Cl) .-:1 - 2000 C Cl) 
• � 
I 
> 0 Cl) 
a: 1 2 
-2000 
T 
! 
% 
3 4 
Scenario 
T 
± 1 .1. 
5 6 
+ mean 
I 95th 5th 
perc 
Figure 49. Comparison of the range of expected returns minus costs from the stochastic 
@Risk partial budgeting model for the six scenarios of reducing reactor numbers in a dairy 
fann. 
Stochastic model on beef breeding and beef finishing farms 
The same analyses were conducted for the beef breeding and beef finishing fanns. Figure 50 shows 
the distribution of the expected economic output for the six scenarios in the beef finishing herds, 
sorted for worst to best case scenario. 
237 
from 5 to zero reacio<,sJyr 
i 
Figure 50. Economic outcome distributions from @Risk stochastic partial budgeting model 
for returns minus costs for six scenarios of reductions in reactor numbers for a beef finishing 
farm (ordered from worst case scenario to best). 
For most scenarios in the beef finishing farms and the ones in the beef breeding farms, where the 
herd stayed on MC, the cost of labour, followed by the cost of fencing and the cost of poison were 
the most important input variables in the sensitivity analysis (see Figure 5 1  as an example for beef 
finishing farms).  In the beef breeding farm scenarios, where the herd came off MC the most 
influential factors were the discount percentage on white-tagged animals, followed by the hourly 
costs of labour and the time required for extra management (see Figure 52). 
238 
Regression Sensitivity for Returns - Cost 1 2  -> O.S/K23 
abor) I $IhrIF13 
Additional feneln I ·.51 
l:z;m.'!'i!t:I:am.t+Biil!t+J!'.32� 
. r--------r------
- 1 .00 -.50 .00 
Coefficient Value 
.14 
.50 
Std b coeff 
calculated at 
end of bars 
1 .00 
Figure 51. Tornado graph with results of sensitivity analysis showing the importance of 
influence of different input variables for the situation where a beef finishing farm had 2 
reactors and reduced it to one every 2nd year due to the implementation of on-farm control 
programmes. 
Regression Sensitivity for Returns - Cost / 5 -> O/J30 
t:!1ill:lliiill11mttlrn�.266_ 
,·r.GIlI$IQmL:Itl7!43'l", -.195_ 
OfcdlscountlE10 
"-"'iGAd-HT!11 aJ'ztm5 Hiim! ml1jt! �iiF; _ . 19  
l:t1mtrt:w:a.:.�{3'jJliJ..j,}rmU:tp".138" • 
1:r:1'.i,44·ft,I.i+,;iMI'tIE"!litUiiil 
•. 1 1 9  
•. 1 01 
-
-1 .00 
.073 
.00 
Coefficient Value 
.785 
Std b coeff 
calculated a' 
end of bars 
1 .00 
Figure 52. Tornado graph with results of sensitivity analysis showing the importance of 
influence of different input variables for the situation where a beef breeding farm had 5 
reactors per year and reduced it to zero due to the implementation of on-farm control 
programmes. 
239 
Comparing 65% compensation level for reactors with zero compensation 
Figure 53  shows the different economic outcomes (revenues minus costs) using current 65% (blue 
colour) and zero (pink colour) compensation levels. The three different farm types are indicated by 
their two-letter code, and the six TB reactor scenarios are indicated by the large bold numbers. For 
both compensation levels, the respective outcomes for each scenario were always higher for dairy 
than for beef breeding and beef finishing farms. For the three scenarios, where the herd did not get 
off MC (scenario 1 ,2 ,  and 4), the revenues minus costs were lowest for beef breeding farms, as the 
added revenue through increased capital prices did not occur. In the scenarios where the herd came 
off MC, the beef breeding farms achieved a higher outcome than the beef finishing farms. 
1 2000 
1 0000 I. 
-� 
N 8000 Z 
c -
J!l 6000 Cl) 0 0 
Cl) 4000 :::l , 
� . 
.� 
E 
Cl) 2000 Q) :::l 
C 
Cl) > 0 Cl) 
0::: 
-2000 
! -1-! I • I� ...... 
t ! 
!
! !' ! I , � ! !� � � T T 
± ! !! 
-4000 "--. --- --- -----
DH BB BD DH BB BD  DH BB  BD  DH  BB BD  DH BB BD  DH  BB BD  
Scenario 1 2 3 4 5 6 
I 95% DH: dairy • mean revenue at 65% camp. • mean revenue at zero camp. BB: beef breed 5% perc. BD: beef finish 
Figure 53. Expected revenues minus costs from stochastic partial budgeting (@RISK) for all three farm types and all six scenarios of reducing 
reactor numbers with 65% and zero compensation. 
N � Cl 
241 
Decision analysis 
Decision analysis was performed for all three farm types, using the expected outcomes of the 
deterministic models for the decision trees .  
Decision analysis on dairy farms 
Decision trees were constructed for the three situations, where the farm has five, two and one, 
respectively reactors per year and adopts on-farm TB control or not. Assuming the probabilities for 
reducing reactor rates as noted in Materials and Methods for a dairy farm, the results show that 
under current 65% and under zero compensation the adoption of the programme on average is the 
preferred option in all except one situation (see also Table 59). 
Table 59. Expected financial values of adoption and non-adoption of on-farm control 
measures on dairy farms under different levels of compensation and slaughter levies with 
different starting numbers of reactors. 
Starting Compensation Slaughter Expected value Expected value Decision 
reactors level levy of adoption of non-adoption 
5 65% None $ 1 ,825 $ 1 ,650 Adoption 
5 65% Current $ 1 ,465 $ 1 , 643 Non-adoption 
5 65% I ncreased $ 1 , 1 75 $ 1 ,000 Adoption 
5 0% None $ 3,480 $ 2,423 Adoption 
5 0% Current $ 3 , 120 $ 2 ,063 Adoption 
5 0% I ncreased $ 2,830 $ 1 ,803 Adoption 
2 65% None $ 1 ,781 $ 822 Adoption 
2 65% Current $ 1 ,421 $ 462 Adoption 
2 65% I ncreased $ 1 , 1 31 $ 172 Adoption 
2 0% None $ 2,6 15  $ 864 Adoption 
2 0% Current $ 2,255 $ 504 Adoption 
2 0% I ncreased $ 1 ,965 $ 2 1 4 Adoption 
65% None $ 2,351 $ 2 , 1 1 2  Adoption 
65% Current $ 1 ,991 $ 1 ,752 Adoption 
65% I ncreased $ 1 ,701 $ 1 ,462 Adoption 
0% None $ 2 ,644 $ 2 , 1 67 Adoption 
0% Current $ 2 ,284 $ 1 ,807 Adoption 
0% Increased $ 1 ,994 $ 1 ,5 1 7 Adoption 
242 
Figure 54 presents the decision tree for the five-reactor situation under the current 65% 
compensation scheme. Probabilities are presented above the branches and outcomes below the 
branches coming off the chance nodes. The expected economic value is higher when on-farm 
control measures are implemented versus situations where no control is conducted ($ 1 ,825 versus 
$ 1 ,650). The return on invested funds for this situation is 1 32% (expected net value of $ 1 ,825 on 
$ 1 ,380 invested). 
0.08 
-21 95.91 
0.1 5  
-1 380 
0.31 
1067.74 
0.31 
2291 .62 
0.1 5 
7774.57 
o 
-298 
o 
o 
o 
2447.74 
o 
3671 .62 
o 
9154.57 
Figure 54. Decision tree for expected f"mancial outcomes for adoption or non-adoption of on­
farm control measures on a dairy farm with five reactors. 
Decision analysis on beef breeding farms 
For the beef breeding farm under 65% compensation the decision tree starting with five reactors is 
presented in Figure 55 .  On average it was more economical not to adopt the control measures, as 
only the increased benefits of getting off MC were captured, but no cost for on-farm control 
incurred. The expected values for adopting on-farm control programmes versus non-adoption are 
presented in Table 60, stratified for the two compensation levels (65% and zero) and slaughter levy. 
o 
-3513.25 
o 
-3270 
o 
-2540.25 
o 
-2175.38 
o 
3725 
0.15  
-243.25 
0.45 
o 
0.2 
729.75 
0.1 
1094.63 
0.1 
6995 
243 
Figure 55. Decision tree for expected fmancial outcomes for adoption or non-adoption of on­
farm control measures on a beef breeding farm with five reactors. 
244 
Table 60. Expected financial values of adoption and non-adoption of on-farm control 
measures on beef breeding farms under different levels of compensation and slaughter levies 
with different starting numbers of reactors. 
Starting Compensation Slaughter Expected value Expected value Decision 
reactors level levy of adoption of non-adoption 
5 65% None - $ 1 ,675 $ 918 Non-adoption 
5 65% Current - $ 1 ,790 $ 803 Non-adoption 
5 65% Increased - $ 1 ,883 $ 7 10  Non-adoption 
5 0% None - $ 322 $ 1 ,551 Non-adoption 
5 0% Current - $ 437 $ 1 ,436 Non-adoption 
5 0% Increased - $ 530 $ 1 ,343 Non-adoption 
2 65% None - $ 707 $ 885 Non-adoption 
2 65% Current - $ 822 $ 770 Non-adoption 
2 65% Increased - $ 9 15  $ 677 Non-adoption 
2 0% None - $ 25 $ 91 9 Non-adoption 
2 0% Current - $ 1 40 $ 804 Non-adoption 
2 0% Increased - $ 233 $ 71 1 Non-adoption 
65% None $ 627 $ 2,336 Non-adoption 
65% Current $ 500 $ 2,221 Non-adoption 
65% Increased $ 407 $ 2,1 28 Non-adoption 
0% None $ 855 $ 2,381 Non-adoption 
0% Current $ 739 $ 2,269 Non-adoption 
0% Increased $ 647 $ 2,1 73 Non-adoption ., 
From Table 60 it can be seen that under the probabilities chosen the preferred option was always 
not to adopt the control programme, independent of compensation level or slaughter levy level. 
For a starting situation with only two reactors and current compensation level of 65%, the adoption 
of the programme was the preferred option, if it led to probabilities of 0.35 for getting the reactor 
rate down to one reactor every second year and 0.55 for getting it down to zero and therefore 
achieving off MC status and if the non-adopter had only a 5 % chance of getting off MC, but a 60% 
chance of having either an increased or the same number of reactors (see Figure 56). 
Decision 
491 .6757895 
o 
-3999.75 
0.052631579 
-3270 
0.368421053 
-2905.13 
0.578947368 
2995.25 
o 
-730 
o 
o 
o 
364.88 
o 
6265.25 
245 
Figure 56. Decision tree for beef breeding farm starting with two reactors and changed 
probabilities of reducing reactor numbers after implementing on-farm control programmes. 
Beef breeding: Including a subsidy for labour and material and payment for off-MC 
The decision tree analysis was also calculated for situations where a subsidy is paid for poison, bait 
stations and labour (e.g. voucher) as well as a yearly payment to farms that achieve to come off 
Movement Control (this payment for a maximum of three years within a 10-year period). The 
subsidies necessary to make the adoption of on-farm control programmes the preferred option in 
the decision tree analysis, depended on the starting situation of the reactor numbers; with five 
reactors the probability of getting off MC are lower than with two or one reactor, and therefore a 
higher subsidy is required. The off-MC payoff was calculated as a yearly payment over a 1 0-year 
period, e.g. a payoff of $4,000 for three years if the herd comes off MC equates to $ 1 ,200 per year 
for 1 0  years. No slaughter levies were included in this analysis. Table 6 1  presents the minimum 
amount necessary in relation to the off-MC-amount paid per year, in order to make ' adoption' the 
preferred option under current compensation level (65%) and zero compensation. For 40% 
compensation Figure 57  presents the subsidies and payments for the five-, two-, and one-reactor 
situation. 
246 
Table 61.  Yearly off-Me-payments and subsidies for material and labour necessary to make 
'adoption' of on-farm control programmes the preferred option on beef breeding farms 
under 65% and zero compensation for reactor animals, stratified for reactor starting 
situations with five, two and one reactors. 
Beef breeding farm 5-reactor situation 2-reactor situation 1-reactor situation 
Off-Me-payment 
$1 ,000 
$ 1 , 500 
$2,000 
$2,500 
$3,000 
$3,500 
$4,000 
2400 
2000 
1 600 
1200 
800 
400 
o 
o 
65% Zero 65% 
comp. comp. comp. 
$ 2,580 $ 1 ,860 $ 1 ,530 
$ 2,580 $ 1 ,860 $ 1 ,490 
$ 2,570 $ 1 ,850 $ 1 ,460 
$ 2,560 $ 1 ,840 $ 1 ,420 
$ 2,550 $ 1 ,830 $ 1 ,390 
$ 2,550 $ 1 ,830 $ 1 ,360 
$ 2,540 $ 1 ,820 $ 1 ,320 
40% compensation level 
• • • • • • • 
.. -- - -- ... 
1 000 2000 3000 4000 
Off-MC payment (in NZ $) 
Zero 65% Zero 
comp. comp. comp. 
$ 880 $ 1 ,650 $ 1 ,460 
$ 850 $ 1 ,6 1 0  $ 1 ,420 
$ 8 10  $ 1 ,580 $ 1 ,380 
$ 780 $ 1 ,540 $ 1 ,340 
$ 740 $ 1 ,500 $ 1 ,3 10  
$ 7 1 0  $ 1 ,460 $ 1 ,270 
$ 670 $ 1 ,430 $ 1 ,230 
-0- 5 reactors 
---- 2 reactors 
1 reactor 
5000 
Figure 57. Subsidies and off-Me-payments for beef breeding farms in order to make 
adoption of on-farm control programmes the preferred option in the decision analysis for the 
five-, two, and one-reactor starting situation, under 40% compensation. 
Decision analysis on beef finishing farms 
The decision analysis for the beef ftnishing farms never resulted in the decision to adopt the control 
programme. Table 62 presents the expected values for the two different compensation levels (65% 
247 
and zero), the different slaughter levies and the different reactor starting situations. In all cases the 
non-adoption of the on-farm control programme was the preferred option. 
Table 62. Expected fmancial values of adoption and non-adoption of on-farm control 
measures on beef fmishing farms under different levels of compensation and slaughter levies 
with different starting numbers of reactors. 
Starting Compensation Slaughter Expected value Expected value Decision 
reactors level levy of adoption of non-adoption 
5 65% None - $ 593 $ 475 Non-adoption 
5 65% Current - $ 1 , 097 - $ 29 Non-adoption 
5 65% I ncreased - $ 1 ,503 - $ 435 Non-adoption 
5 0% None $ 1 ,295 $ 1 ,358 Non-adoption 
5 0% Current $ 791 $ 854 Non-adoption 
5 0% I ncreased $ 385 $ 448 Non-adoption 
2 65% None - $ 1 ,097 $ 25 Non-adoption 
2 65% Current - $ 1 ,601 - $ 479 Non-adoption 
2 65% I ncreased - $ 2,007 - $ 885 Non-adoption 
2 0% None - $ 1 45 $ 73 Non-adoption 
2 0% Current - $ 649 - $ 431 Non-adoption 
2 0% I ncreased - $ 1 ,055 - $ 837 Non-adoption 
65% None - $ 1 ,430 $ 34 Non-adoption 
65% Current - $ 1 ,934 - $ 470 N on-adoption 
1 65% I ncreased - $ 2 ,340 - $ 876 Non-adoption 
0% None - $ 1 ,096 $ 97 Non-adoption 
0% Current - $ 2,320 - $ 407 Non-adoption  
0% I ncreased - $ 2 ,006 - $ 8 1 3  Non-adoption 
Beef finishing: Including a subsidy for labour and material and payment for off-MC 
For the beef finishing farms the decision tree analysis was also calculated for situations where 
subsidy was paid for poison, bait stations and labour and the yearly payment to fanus that achieved 
off Movement Control status. No slaughter levies were included in this analysis. Table 62 presents 
the minimum amount necessary in relation to the off-Me-amount paid per year, in order to make 
'adoption' the preferred option under current compensation level (65%) and zero compensation .  
Figure 5 8  presents the subsidies and payments for the five-, two-, and one-reactor situation under 
40% compensation. 
248 
Table 63. Yearly off-Me-payments and subsidies for material and labour necessary to make 
'adoption'  of on-farm control programm es the preferred option on beef finish ing farms 
under 65% and zero compensation for reactor animals, stratified for reactor starting 
situations with five, two and one reactors. 
Beef finishing farm 5-reactor situation 2-reactor situation I-reactor situation 
Off-Me-payment 65% Zero 65% 
$1 , 000 
$ 1 , 500 
$2,000 
$2,500 
$3,000 
$3,500 
$4,000 
comp. comp. comp. 
$ 1 ,060 $ 1 0  $ 1 ,060 
$ 1 ,050 $ 20 $ 1 ,020 
$ 1 ,040 $ 20 $ 990 
$ 1 ,040 $ 30 $ 960 
$ 1 ,030 $ 40 $ 920 
$ 1 ,020 $ 50 $ 890 
$ 1 ,010 $ 50 $ 850 
40% com pensation level 
1600 -r------------------, 
1 200 +----- --------i 
800 ------------� 
t=+ � 
400 +------------� 
O --r--------,---.----,----j 
o 1 000 2000 3000 4000 5000 
Off-MC payment (in NZ $) 
Zero 65% Zero 
comp. comp. comp. 
$ 150 $ 1 , 390 $ 1 , 1 20 
$ 1 20 $ 1 ,360 $ 1 ,090 
$ 90 $ 1 , 320 $ 1 ,050 
$ 50 $ 1 ,280 $ 1 ,010 
$ 20 $ 1 , 240 $ 970 
$ 0  $ 1 ,21 0 $ 940 
$ 0  $ 1 , 170 $ 900 
-+- 5 reactors 
---- 2 reactors 
1 reactor 
Figure 58. Subsidies and ofT-Me-payments for beef finishing farms in order to make 
adoption of on-farm control programmes the preferred option in the decision analysis for the 
flVe-, two, and one-reactor starting situation, under 40% compensation. 
249 
Provision of voucher for control and off-Me-payment for the Wairarapa 
region 
Using the level of subsidies described in the previous section for beef farms, which would make the 
adoption of on-farm control programmes fmancially beneficial, the total cost for the Wairarapa was 
calculated. Currently there are 80 farms on MC (AgriQuality, pers. comm. 2000). Forty-five (45) of 
these farms are beef breeding, 17 beef fmishing, 15 dairy herds and three miscellaneous herds. As 
these miscellaneous herds are very small, they are grouped together with the dairy farms. 
Two situations were calculated; in the first situation an off-MC payment of $2,500 was made for 
three years and in the second situation this payment was only $ 1 ,000. Table 64 presents the amount 
necessary for each of the herd types and the number of reactors currently on the farm. whereby the 
two categories one or two reactors were combined. 
Table 64. Subsidies used for the three farm types under two different amounts of off-MC-
payments, stratified for starting reactor numbers under 40% and 65% reactor compensation. 
Dairy Beef breeding Beef finishing 
Reactor numbers 5 1 or 2 5 1 or 2 5 l or 2  
$2,500 off- 40% comp. $ 0  $ 0  $ 2,280 $ 1 ,460 $ 650 $ 1 , 1 80 
MC-pay 65% comp. $ 0  $ 0  $ 2,560 $ 1 ,540 $ 1 ,040 $ 1 ,280 
$1 ,000 off- 40% comp. $ 0  $ 0  $ 2,310 $ 1 ,580 $ 670 $ 1 ,290 
MC-pay 65% comp. $ 0  $ 0  $ 2,580 $ 1 ,650 $ 1 ,060 $ 1 ,390 
Assuming that 10% of MC-herds have five or more reactors, 70% have 2 to 5 reactors, 20% of 
MC-herds have one reactor per year and that 80% of the MC-herds come off MC within 5 years, 
the regional costs to the Wairarapa can be calculated. The total costs to the region from providing 
vouchers as indicated in Table 64 and two different levels off-MC-payments ($2,500 and $ 1 ,000) 
are detailed in Table 65. The off-MC-payments are paid for three years within a 1O-year period. 
250 
Table 65. Regional cost for providing vouchers for control work and two different off-Me­
payments under two different compensation levels for reactor animals. 
Dairy Beef breeding Beef f"mishing Total 
$2,500 off-Me-payment for three years and 40% reactor compensation 
Vouchers over 5 yrs $ 0  $ 344,900 $ 95,000 $ 439,900 
Off-Me-payments over 5 yrs $ 1 05,000 $ 270,000 $ 105,000 $ 480,000 
Total cost over 5 yrs $ 1 05,000 $ 614,900 $ 200,000 $ 91 9,900 
A vg. annual cost $ 21,000 $ 122,980 $ 40,000 $ 183,980 
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
$2,500 off-Me-payment for three years and 65% reactor compensation 
Vouchers over 5 yrs $ 0  $ 366,900 $ 106,400 $ 473,300 
Off-Me-payments over 5 yrs $ 1 05,000 $ 270,000 $ 105,000 $ 480,000 
Total cost over 5 yrs $ 1 05,000 $ 636,900 $ 21 1 ,400 $ 953,300 
A vg. annual cost $ 21,000 $ 127,380 $ 42,280 $ 190,660 
- - - - - - - - . - . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - . - - - - - - - - - - - - - - - - - - - - - - - - - . - - . - . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
$1,000 off-Me-payment for three years and 40% reactor compensation 
Vouchers over 5 yrs $ 0  $ 370,1 00 $ 1 03,450 $ 473,550 
Off-Me-payments over 5 yrs $ 42,000 $ 1 08,000 $ 42,000 $ 1 92,000 
Total cost over 5 yrs $ 42,000 $ 478,1 00 $ 145,450 $ 665,550 
A vg. annual cost $ 8,400 $ 95,620 $ 29,090 $ 133,110 
- - - - - - - - - - - - - - - - - - - _ . _ - - - - - - - - - . - - - - . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - . - . - - - - - - - - - - - - - - . - - - - - - - - - - - - - - - - . - - -
$1,000 off-Me-payment for three years and 65% reactor compensation 
Vouchers over 5 yrs $ 0  $ 389,850 $ 1 14,850 $ 504,700 
Off-Me-payments over 5 yrs $ 42,000 $ 1 08,000 $ 42,000 $ 1 92,000 
Total cost over 5 yrs $ 42,000 $ 497,850 $ 1 56,850 $ 696,700 
A vg. annual cost $ 8,400 $ 99,570 $ 31,370 $ 139,340 
The total cost for the region is less if only a $1 ,000 off-MC-payment is paid instead of $2,500. The 
amount paid for vouchers is very similar in both situations. 
The amount for the vouchers could also be calculated on a per hectare basis, as the amount of 
vector control work is dependent on the farm size. For the situation with $ 1 ,000 off-MC-payment 
and 40% reactor compensation the vouchers amount to $5.40/ha for BB with 5 reactors, $3.70/ha 
for BB with 1 or 2 reactors, $2.80/ha for BD with 5 reactors, and $5.38/ha for B D  with 1 or 2 
25 1 
reactors. Under 65% reactor compensation, the vouchers amounts to $6.03/ha for BB with 5 
reactors, $3.86/ha for BB with 1 or 2 reactors, $4.42/ha for BD with 5 reactors, and $5,79/ha for 
BD with 1 or 2 reactors. 
Discussion 
Factors considered and omitted in the partial budget 
The acquisition and organisation of data for the economic analysis was the most difficult aspect, as 
found in most agricultural economic decision-making (Marsh, 1999). Data about consequences of 
the implementation of on-farm TB control measures were obtained by visiting some farms, which 
were part of the intervention study that evaluated these measures (see Chapter 3). Discussions with 
experts from AgriQuality and experts from the farm consulting company ' Baker & Associates' 
were another important source of information. They provided the best available information to 
calculate or estimate data that are influenced by the implementation of on-farm control measures. 
To facilitate the comparison between the three different types of farms, equal values were assumed 
across farms for the costs of one hour of labour, costs of fencing and depreciation period. The time 
horizon (the period for which the economic analysis was undertaken) was kept to one year, as the 
costs were fixed and of short-term effect. Consideration of the time value of money (discounting) 
was therefore not necessary (Marsh, 1 999). 
In the present model two factors were included in the additional returns accrued from the 
implementation of the control strategies: ( 1 )  increased animal value, which only applied if the herd 
came off MC and (2) increased milk production, which only applied to dairy farms. It is generally 
known and accepted that white-tagged animals can often only be sold at a discount. However, the 
estimation of these discounts proved difficult, as no published information was available. The 
values used in the analysis were based on discussions with AgriQuality and experts from the AHB. 
However, we realise that these percentages chosen can be debated, and that they might need 
adjustment in further research, as more detailed information becomes available .  
The increased milk production on dairy farms was due to the reduction of the number o f  reactors. If 
this number is reduced as a consequence of adopting on-farm control measures, these animals do 
not have to be removed in mid-lactation. 
Only one factor was included in the reduced costs section of the present model: premature disposal 
due to TB, not as a consequence of the animal dying from TB, which is in the long run fatal, but 
because under the official control regulations the animal has to be slaughtered once suspected of 
252 
being tuberculous. Three other factors were considered for inclusion in the 'reduced costs' section, 
but were either regarded as unimportant or un-quantifiable in the present context: ( 1 )  Altered feed 
conversion efficiency, which occurs if the disease affects animal productivity by altering the 
metabolic processes for protein and other nutrients (Dijkhuizen and Morris, 1997). However, under 
current TB testing schemes, TB is unlikely to have a subclinical effect on milk production and live 
weight gain of the animals, because generally it is diagnosed very early. (2) Increased live weight 
gain; generally diseased animals gain weight more slowly than equivalent disease-free animals 
(Dijkhuizen and Morris, 1997). In New Zealand the existing TB control practices ensure animals 
are identified and slaughtered before clinical signs of TB show. Therefore, the increase of live 
weight was considered to be minimal and was left 'un-quantified' .  (3) More accurate genetic 
selection. If cows have to be removed from the herd due to being considered tuberculous, this 
means that more replacements have to be chosen from the available heifers, leaving fewer options 
and fewer animals for sale. Thus not only will livestock sale income be reduced, but also 
management flexibility for herd improvement will be curtailed (Dijkhuizen and Morris, 1997). TB 
infection might contribute to the loss of genetically superior animals. However, this effect was 
considered un-quantifiable in the present study. 
Quantifying the cost and time needed for implementing an on-farm control programme, involving 
vector control and altered grazing management strategy, was complicated through the fact that 
grazing management was already incorporated in any existing farm system. Additionally the 
revenues and investments that resulted from implementing such specific on-farm control measures 
were influenced by farm characteristics, such as vegetation, location of the farm and especially the 
commitment of the farmer. 
A potential additional cost-contributing factor was the purchase of supplements. Excluding TB hot­
spot areas from grazing could mean less available feed for cattle, thus necessitating supplementary 
feeding. However, dairy farms could use hot-spot areas for part of the day if necessary and beef 
farms could use sheep in these areas. Furthermore beef farmers are more likely to sell their cattle in 
times of pasture shortage, than to buy in supplements. Therefore this factor was excluded from the 
analysis. 
In the present study no returns foregone were included. A drop in milk production as a result of not 
using TB hot-spot areas, thus sub-optimal use of grazing possibilities, could have been a possible 
return foregone. Chapter 7 considers this issue, and concludes it is not important. The increase in 
milk production, due to not having to remove animals in mid-lactation, was expected to be much 
higher and as a result the overall milk production will increase. 
253 
For the sheep and beef farms the effects of adopting grazing strategies on the sheep production was 
left un-quantified. Sheep can be used to graze TB hot-spot areas, as they will not expose 
themselves readily to moribund possums (Sauter and Morris, 1 995). As a consequence of this and 
the introduction of a detailed grazing strategy, the live weight gain in sheep might be increased. 
However, these influences were difficult to measure and were left un-quantified in this analysis, but 
were considered in the whole-farm simulation model (Chapter 7).  
In the present analysis only direct impacts were taken into account. Other benefits associated with 
the eradication of tuberculosis from herds, such as the achievement of the international status of TB 
freedom, with its effects on export and trade, implications for human health and animal welfare 
have not been quantified in the study. These items can far exceed those at the farm level, but the 
analysis has been restricted to farm decisions. Macro-economic effects would require a more 
complex economic evaluation for which data availability would be a major issue (Mclnerney et aI. , 
1 992). Additionally the analysis only took into account the monetary values of costs and benefits. 
However, for farmers to decide whether or not to implement any TB control, they may also be 
influenced by non-monetary considerations - the perception of TB, its threat, and social effects. D. 
Maling ( 1993), a deer-farmer in Hawke' s Bay described a range of feelings, from anger to despair, 
as a consequence of a TB outbreak. These emotions cannot be put into monetary values and yet, 
they might have a far greater impact on the farmer and his decisions regarding TB control, than any 
expected financial outcomes of implementing control programmes. For the farmer mentioned 
above the only goal was to eradicate TB from his farm and he consequently spent $ 14,500 on blood 
tests alone, regardless of economic implications of this course of action. 
Farms used in the study and TB reactor scenarios analysed 
Using 'typical' or 'representative' farms was one way of dealing with the variety of farm practices 
and farm management systems. Management on dairy farms does not differ substantially between 
farms; however, differences might exist in replacement rates and calving percentages, which 
influence the number of cattle being sold annually. In contrast, beef breeding and beef finishing 
farms show great variety. These farms adjust their management continuously, depending on market 
prices and available feed supplies. Therefore, the number of cattle on the farm and the type of cattle 
differ considerably from year to year. In addition many farms in the Wairarapa have changed their 
stock selling policies as a result of having TB in their herd. Farms that had the grass growth 
potential moved from selling weaners to keeping less breeding stock and finishing all offspring. 
Hence, they could avoid selling white-tagged weaners at a discount of about $30 to $40 a head. 
These farms can be regarded as a combination of breeding and finishing enterprises. However, in 
the present study they were not considered as a farm type on their own, as all their stock leaves for 
254 
slaughter, therefore showing high similarities with beef finishing farms. However, the revenues 
would be slightly less as female young stock would not achieve the same weight and hence price as 
male young stock, which mostly are held on pure beef finishing farms. On beef breeding farms the 
costs of the programme would be higher than on beef finishing farms, as beef breeding farms 
usually comprise a larger area. However, the larger number of cattle sold would compensate this. 
In the present study it was assumed that all beef finishing farms tested their cattle prior to sending 
them to slaughter and that all tuberculous cattle were identified through the test, rather than being 
found lesioned at slaughter without testing positive first in the TB test. Beef finishing farms have 
the option to obtain the herd status 'Works monitored' if over 90% of the herd get slaughtered 
annUally. In herds with this herd status, no on-farm TB testing is required - all animals are subject 
to the routine slaughterhouse inspection. If a small lesion is found in such animals, the farmer 
receives the revenue of the carcass, which is downgraded to local trade price instead of possible 
export price. If the lesions found are extensive, the carcass may be condemned and the animal does 
not generate any revenue for the farmer. However, these condemned cases occur very seldom with 
current TB testing policy. As the number of farms with herd status 'Works monitored' in the 
Wairarapa was very low, the present analysis did not take this farm type into consideration. 
The six different reactor number scenarios modelled in the study, starting with five, two, one, 
respective reactor per year were intended to cover the range of observed situations in the 
Wairarapa. Three scenarios covered the situation, where the farm remained under MC and the other 
three where the herd was classified as TB free ( 'Clear' Status) as a possible result of implementing 
the control measures. It was important to evaluate the economic effects for both outcomes 
regarding MC. 
Results from the deterministic and stochastic models 
Uncertainties and lack of information about parameter values were incorporated into the 
spreadsheets by using the stochastic modelling approach, which did not assume fixed values but 
rather a range of values. Sensitivity analysis was used to evaluate the effects of changing 
parameters. In scenarios where the farm did not come off MC the net benefit was most influenced 
by the cost of labour, followed by the cost of fencing and poison in beef farms, and the time spent 
on control measures for dairy farms. In all scenarios where the herd came off MC, the outcome was 
most severely affected by the discount percentage of white-tagged animals. 
255 
Farm types 
Results from the deterministic and stochastic model showed that for dairy and beef breeding farms 
it was much more beneficial to implement control measures if the herd achieved off-MC status, as 
this resulted in a higher value for live animals. The farm was able to sell cattle without the discount 
of white-tags, assumed to be 10% and 15%. Another advantage of being off MC is the higher 
flexibility in terms of sales policies and grazing policies, such as having the opportunity to graze 
cattle on TB free farms. This is often critical to dairy farms in order to avoid damage to paddocks 
during the wet winter months. This higher flexibility was not specifically included in the partial 
budgeting, but needs to be considered too. The higher value for live animals and the higher 
flexibility was a considerable advantage for dairy farms and sheep and beef breeding farms, but had 
no effect on beef finishing farms. As beef finishing farms generally only sell their cattle to 
slaughter, farmers receive 100% slaughter value for animals with and without white MC tags, if no 
TB lesions are found. These animals are generally suitable for local and export trade. In the export 
trade, only a few countries, such as Russia, will not accept cattle from Movement Control herds 
(Atkinson, pers. comm., 2000). The fact that white-tagged cattle can yield the same revenue at 
slaughter as non white-tagged cattle, is often used to the advantage of a beef finishing farm when 
white-tagged animals are bought in at a discount and then sold to slaughter, where no discount 
applies. 
Compensation level and subsidies 
Dropping the compensation level from the current 65% to zero would make it more beneficial for 
beef finishing farms to implement control measures. If the farm achieved at least two reactors less 
through the implementation of control measures, the investment was financially worthwhile with 
an expected net benefit of $350. However, for beef breeding farms the reduction in compensation 
was still not enough to make adoption of the control measures worthwhile if the herd remained on 
MC. This was mainly due to the high costs of the programme and the continuous discount in price 
obtained for animals sold for fattening. If the costs of the programme were reduced through a 
subsidy of poison, the programme had to achieve a reduction of at least four reactor animals to 
make it worthwhile. 
It was argued that by providing compensation the actual costs associated with TB are masked, 
which will result in a reduced incentive for farmers to conduct vector control (Animal Health 
Board, 1995). As the principles of beef finishing farms apply to most mixed farms in the 
Wairarapa, the abandonment of the compensation could be beneficial on a large number of farms . 
However, due to the national and regional TB control work being done, it is more likely that the 
farms with continuing TB problems would have only one or two reactors. If the farm has only one 
256 
reactor, then the implementation of on-farm control programmes is not even beneficial under zero 
compensation. 
It is therefore concluded that other incentives, apart from reduction of compensation, have to be put 
in place. One also has to give some consideration to the compliance of farmers if compensation 
was removed. Many farmers expressed their concern that they or other farmers would not present 
all animals for testing or would remove possible reactors from the herd before the test is read (see 
Chapter 5). This therefore would harm the control process very much or would require means of 
checking the compliance of farmers. 
Results from the decision analysis 
The fmancial outcomes in the deterministic and stochastic model for beef breeding farms showed 
that it was well worth implementing on-farm control programmes if the herd achieved off MC 
status, but not so if the herd remained under MC. Thus, this would leave it up to the farmer to 
assess the risk to his farm, how likely he thought it is to achieve zero reactors. The more the 
compensation level is reduced the higher the expected financial benefit if the farm gets off MC. 
Decision tree analysis was used to evaluate these situations, taking into account elements of 
uncertainty. The choice of the preferred option is found through a process called folding back 
(Marsh, 1999). The result of the decision tree analysis with 65% compensation for the five reactor 
situation showed that the expected value of implementing control on a beef breeding farm was 
negative with -$ 1 ,675, whereas when no compensation was paid, the expected outcome was -$322. 
The decision analysis also showed that for beef finishing farms the adoption of control measures 
never resulted in a net gain, due to the costs of the programme not being counterbalanced by any 
additional return if the herd came off MC, only by the difference in value of reactor animals. The 
expected outcome of non-adoption could never fall below zero, as no additional costs were 
incurred, but benefits were obtained at zero cost if the herd reduced its reactor numbers. This 
indicated that beef fmishing farmers are probably better off investing their money in other parts of 
their enterprise rather than TB control with current policies. 
Calculating the expected values for the three farm types under a policy where no compensation was 
paid, showed that there was no change in the preferred option for dairy nor for beef farms; for dairy 
it was on average more beneficial to implement the control measures than doing nothing, both 
under the current compensation scheme, and under zero-compensation. For the beef farms it was 
never the preferred option to implement control measures. 
257 
From the decision tree analysis with the two different levels of slaughter levy on adult cattle on 
beef farms it was seen that the increase of the levy would further add to the costs of TB for the 
farm, and therefore it would be less likely that farmers adopt on-farm control programmes. The 
expected value for the adoption of the programme was always less than for the non-adoption in the 
beef farms. None of the different combinations of levy and compensation resulted in the adoption 
of the programme as the preferred option. Therefore,  other incentives have to be put in place. 
Insurance as an additional method 
The current policy where all farmers pay a slaughter levy on adult cattle and farmers obtain 65% 
compensation for any animal reacting positively to the TB test, is a form of insurance. However, 
this can lead to reduced motivation to do on-farm control, which was found in an outbreak of 
classical swine fever in Germany (Davies, 1 996). Thus it might result in using less preventive 
measures and increased reliance on large-scale control efforts (Howe and Whittaker, 1997). It also 
induces situations where farmer 'blame' the national programme for any failures on their property. 
In New Zealand, the current system of a universal slaughter levy and compensation for reactor 
cattle has worked well so far, as many herds were affected by tuberculosis or saw themselves at 
risk. 
A more comprehensive insurance scheme could involve rate-making, where farmers pay different 
insurance rates, depending on area, enterprise type, precautions taken and similar criteria CHowe 
and Whittaker, 1 997). However, traditional rate-making in insurance is subject to criticism that past 
experience does not fully represent the future. If the deviation from the past is positive, this does 
not cause a great problem, but if the deviation is negative then the rate calculated will be 
inadequate (Vaughan and Vaughan, 1999). Differentiating the premiums would also result in more 
administrative and monitoring costs, therefore necessitating evaluation of whether the efficiency 
gained through differentiation is more than these additional costs. 
Insurance companies are also reluctant to insure events with a high probability of occurrence, and 
particularly of re-occurrence on the same farm. The probability of an outbreak of an endemic 
disease is relatively high and several farmers in the same area might be affected. In the case of TB 
the proportion of farms affected is now low and declining, so companies may be willing to insure 
farmers .  However, many farmers may see the risk as low and hence decline insurance, so the risk 
would not be adequately shared and premiums for at-risk farms would be unacceptably high. 
The biggest problems of an insurance scheme would be those of adverse selection and moral 
hazards. Producers with good control measures could be less likely to buy insurance than producers 
with poor control measures, as the probability of getting infected is smaller for the former ones than 
258 
for the latter (Ekboir, 1999). Furthennore producers with a potentially higher loss could be more 
likely to take out insurance. For example a stud beef breeding fann that sells cattle nationally will 
incur very high losses if the herd is diagnosed with tuberculosis and would insure, whereas a beef 
finishing fann would not. A further problem could be that fanners who are off MC for longer 
periods underestimate the threat of TB to their farm and are less likely to buy insurance, but also 
less likely to conduct control measures. The biggest problem of an insurance scheme is moral 
hazard; if farmers who bought insurance took less care in their control programmes and therefore 
took more risk of getting infected, than if they had not bought insurance. 
Even if a way of insurance for TB is found that works well, the scheme will have high costs in 
terms of administration and audits. These costs have to be covered in the case of a private insurance 
market by the fanners choosing to insure their herds, or in the case of a state insurance scheme, by 
taxpayers' money. 
Effects for future control of TB 
From the potential problems with an insurance scheme it is suggested that increased emphasis on 
insurance against TB is not a preferred option in the national control of TB on their own properties. 
Instead, increased motivation is necessary that encourages farmers to assist the control of TB. 
Motivation can be increased through market signals. The current policy of slaughter levy and 
compensation does not give any market signals to encourage personal efforts to control TB on 
farms. However, if compensation is removed altogether there is the risk that fanners will 
circumvent the measures and that the co-operation of the farmers in the national control programme 
is lost. 
Reducing, not eliminating, the compensation will increase useful market signals. However from the 
analysis it can be seen that the increased loss per TB animal is not enough to counterbalance the 
costs of the programme. Therefore it is concluded that the reduction in compensation is not 
incentive enough. Furthermore, from the analysis it can be seen that dairy herds gain more 
financially from on-farm control programmes than beef farms (Table 66). 
259 
Table 66. Expected fmancial outcome of adopting on-farm control measures for the three 
herd types and the different reactor scenarios. 
Under 65 % compensation Under zero compensation 
Dairy Beef breed. Beef fm. Dairy Beef breed. Beef fm. 
From 5/yr to 21yr $ 1 ,068 - $ 2,460 - $ 572 $ 2,725 - $ 1 , 1 05 $ 1 ,320 
From 5/yr to 1 /200 yr $ 2,292 - $ 2,095 - $ 62 $ 4,778 - $ 63 $ 2,775 
From 5/yr to off MC" $ 7,750 $ 3,805 $ 108 $10,537 $ 6,064 $ 3,260 
From 21yr to 1/20  yr - $ 156 - $ 2,825 - $ 1 ,081 $ 673 - $ 2,147 - $ 135 
From 21yr to off MC" $ 5,327 $ 3,075 - $ 91 1 $ 6,432 $ 3,979 $ 350 
From 1 /yr to off MC" $ 4,51 1 $ 2,832 - $ 1 ,251 $ 5,063 $ 3,284 - $ 620 
An additional inequity of costs associated with TB between dairy and beef producers in the 
Wairarapa are the regional council rates .  These are set per hectare, independent of land use, but 
dependent on the existence of Regional Council vector control programmes (R.Cleary, pers. 
comm., 2000) .  However, the properties of beef producers are much larger than those of dairy 
producers, thus introducing further disadvantage to beef producers. This could also induce an 
increased reliance on Regional Councils and a tendency to blame them for problems. Beef 
producers pay proportionally more to RC and therefore 'expect' RC to control possums on their 
land. However, these rates are not set by the AHB and therefore the option of reducing RC rates for 
beef producers was not elaborated in this study. 
Based on the different gains/losses for dairy- versus beef producers, it is suggested that differential 
positive and negative signals to different producers should be put in place. A simple reduction in 
compensation does not provide such a differentiation. The following options are suggested: 
Substantially penalising TB positive animals at slaughter (culls) while leaving compensation 
for test-positive animals at 65%; for example the introduction of a charge on all TB culls of 
70% of value. This step would encourage pre-slaughter testing on farms. The skin test has 
better sensitivity than detection of TB lesions at routine slaughter (Ryan, 1 992; Corner, 1994; 
Whipple et al., 1 995). In the last phase of a disease eradication programme it is important to 
find all true positive animals, therefore it is important to avoid situations where tuberculous 
animals go to slaughter and do not get detected. The herd of origin of these animals poses a risk 
to other herds. 
Introduce a payment per year for herds which get off Movement Control. These payments 
could be limited to a maximum of three years in a 10-year period. They could be differential, a 
260 
higher sum for beef producers than for dairy-producers, in order to achieve the same fmancial 
gain for both enterprise types. The main advantage of this payment is as an incentive for 
farmers to assist the control of TB. As the number of farms in this category is small and will 
further be reduced, this part will not be a major cost to the national programme. From Figure 
43 to Figure 45 it can be seen that the expected net return of implementing on-farm control 
programmes on beef breeding farms is around $3,000 and for beef fmishing farms around 
$4,000 lower than for dairy farms. These differences in the net return could be reflected in the 
suggested payments for coming off MC. 
In order to make the adoption of on-farm control programmes more attractive, the costs of 
material and labour have to be reduced to the individual farmer. Therefore, vouchers for 
additional RC/contractor/own on-farm control efforts or similar systems could be provided. 
Table 55, Table 56, and Table 57 provide the amount necessary for beef farms to obtain a net 
gain in all reactor number scenarios. When poison and bait stations are provided for free, dairy 
farms will always obtain a net gain. This voucher-system in conjunction with the yearly 
payment if the herd comes off MC provides a big incentive for farmers to do something on 
their farms. 
Use a phasing of different methods over the next four to five years, with decreasing 
compensation, higher penalties for tuberculous animals at slaughter (culls) and stricter 
movement control rules, such as MC farms only permitted to sell to slaughter. 
Vaccination could be used for fringe areas, where there is a risk of spreading TB into uninfected 
area. Funding from different sources could also be focused on different control measures. E.g. 
focus Government and regional rates money on broad-scale RC control efforts, and the limitation 
of spread of TB, whereas the industry money could be used for additional on-farm control efforts. 
Additionally, a significant share of TB costs at AHB level could be shifted from beef to dairy 
producers, as they are the ones who benefit most of control programmes. 
With these different combinations of penalties and incentives it is possible that the control of TB is 
achieved quicker than if the sole focus is put on large-scale intensive possum control operations. 
The intention is that a policy mix is found that provides positive and negative incentives to each 
enterprise type which will give each an appropriate set of market signals. 
The risk that Vector-Risk-Areas are still expanding requires separate methods. Reducing possum 
numbers on its own might not solve the problem, but vaccination on a broad base could be much 
more useful. 
261 
With ongoing national and regional TB control it will be more likely that farms will have only one 
or two reactors, rather than five or more. Therefore a change in compensation level from 65% to 
zero could make a difference for the beef breeding farms. Under current compensation levels, 
adoption of control measures for a farm with two reactors does not result in a net benefit, but a net 
benefit would be achieved if no compensation were paid. 
From January 2000 new cattle movement policies are in place, whereby cattle can only move to 
other herds if the herd already had one clear test and the group to be moved has passed a clear test 
within 60 days of movement. Animals from infected herds without a clear test will have to pass a 
skin test and a parallel blood test. If these conditions are met, then the animals can move subject to 
a permit, with white ear tags, and they must be tested a minimum of 90 days after the pre­
movement test. In addition the purchaser' s herd status will be suspended and only restored to Cl  
after a clear whole herd test a minimum of six months after the post-movement test. 
It is expected that this will constrain cattle sales for beef breeding farms very much if they are still 
on MC, as not many purchasers will want to lose their status and obtain a status of only Cl (Animal 
Health Board, 1 999). Especially purchasers with herds that have been clear for many years will not 
buy white-tagged animals any more. The additional requirement of a blood test for cattle from 
infected herds without a clear herd test, is a further step towards ensuring that no infected animals 
can go on to other farms. 
Another expected consequence of the changed rules is that many beef breeding farmers will change 
their trading policies towards reducing their number of beef breeding cows and fattening all the 
offspring (Animal Health Board, 1999). Therefore the same principles would apply to them as for 
the beef finishing farms, which are not affected by the change of regulation. These farms are still 
able to buy in white-tagged cattle at a heavily discounted price. If the animals pass the post­
movement test, their white tags can be removed and the cattle fattened and sold to slaughter or 
privately as normal animals. Any purchaser's herd status thereafter would not be affected anymore. 
This is also the case when farms have two different blocks of land with two separate herd numbers. 
For example a farm with a herd status of 15 can get their weaned calves tested (skin test plus 
parallel blood test) and (provided they all pass) move them on to the second block, and test them 
again within 90 days of the pre-movement test. If the animals pass this test as well, the white ear 
tags can be removed and the animals sold to any farm (e.g. C6) without affecting the purchaser's 
herd status. A stricter version of the rules might involve that any purchaser' s herd status would 
obtain the lower status of the herd where the animals came from. For example, if a herd with 
classification C5 buys in cattle from a herd classified as C3, the herd status will change from C5 to 
C3. 
262 
With the change of policy and its expected impact on the possibilities of being able to sell white­
tagged cattle, it might well be that the discount for white-tagged cattle is even greater than assumed 
in our study ( 1 0% for dairy and 1 5 %  for beef cattle). As the sensitivity analysis in the stochastic 
modelling showed, this discount value is of major importance when the herd comes off MC. This 
means if the discount is even greater, the added benefit of selling non-white tagged cattle is greater. 
In the decision analysis this higher positive outcome could therefore cause on average a higher 
expected monetary value for the implementation option than for the option of doing nothing. It 
remains to be seen if the discount will change very much as a result of the changed rules for 
movement from infected cattle herds. 
TB and its effect in the wider economy 
The part which the public sector has to play in controlling animal diseases is increasingly debated 
(Mclnerney, 1 996; Sanson and Thornton, 1997). As a result the private sector has to adopt a larger 
role in planning and implementing animal health policies (Ekboir, 1 999). However, with a disease 
complex such as tuberculosis, which has probably a greater effect on the national export trade than 
on the individual farm, one could argue about the principle of 'user-pays' . The potential losses in 
export trade in the worst-case scenario could be worth NZ$5 billion over 10 years (Parliamentary 
commissioner for the envirionment, 1994). Additionally, at farm level control on one farm will 
automatically also have effects on the neighbouring farms, termed 'external effects' or 
'externalities' .  Poor control by neighbours reduces the effectiveness of possum control by any 
particular farmer, through potential immigration of possums from neighbouring properties. The 
same applies the other way round, if a farm is surrounded by farms that conduct intensive possum 
control, then eventually this will also show a positive effect on the farm in the middle, even if that 
farm does not do any control at all. Therefore this farmer in the middle benefits from the control 
programmes of the neighbouring farms ('free-riding' )  (Ekboir, 1999). In order to avoid the problem 
of 'free-riding' and to make sure that all farmers are participating in the control effort, it might be 
necessary to impose legal requirements for participation, and compliance monitoring. 
Alternatively, market signals might be used to induce such producers to participate. 
Under the current TB control policies it does not seem to be economically beneficial for beef farms 
to eliminate TB from their herds, however as the disease does pose a threat to the national 
economy, it is very important to eradicate TB. If the aim is to put more and more responsibility for 
TB control on farmers instead of the general tax-paying community, then rules, regulations and/or 
incentives should be put in place that create positive net-benefits for all farm types. 
The need to become free of TB might increase enormously, if the persistence of TB in New 
Zealand is going to be used as a non-tariff barrier against New Zealand by trading partners. The 
263 
consequences of TB in a herd would then be much more dramatic than assumed in this study. 
Therefore, efforts should be made now to reduce the risk of TB spread and to reduce the TB levels 
on farms. As long as there is no measure available to eradicate TB from possums, or even better to 
eradicate the possum population itself, on-farm control measures, including targeted vector control 
and grazing management system, can be a helpful instrument in the fight against TB. 
Conclusions 
Table 67 and Table 68 show summaries of the results found in the study. In Table 67 the financial 
outcomes of the deterministic and stochastic partial budget are presented as 'Gain' or 'Loss' . The 
TB reactor scenarios are as described in the Materials and Methods. In Table 68 the outcomes of 
the decision tree analysis using the probabilities as stated in Materials and Methods are 
summarised. 
Table 67. Economic outcomes of deterministic and stochastic partial budgeting on adopting 
on-farm control methods for TB, stratified by farm types, compensation level and reduction 
of TB reactor numbers. 
Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 
5 down to 5 down to 5 down to 2 down to 2 down to 1 down to 
2 reactors 0.5 react. o reactors 0.5 react. o reactors o reactors 
OH with 65% Gain Gain Gain Loss Gain Gain 
comp. 
BB with 65% Loss Loss Gain Loss Gain Gain 
comp. 
BD with 65% Loss Loss Gain Loss Loss Loss 
comp. 
OH with zero Gain Gain Gain Gain Gain Gain 
comp. 
BB with zero Loss Loss Gain Loss Gain Gain 
comp. 
BD with zero Gain Gain Gain Loss Gain Loss 
comp. 
264 
Table 68. Outcomes of decision analysis whether to adopt on-farm control programmes or 
not, stratified by farm types, reactor compensation level and number of reactors to start with. 
Starting with 5 Starting with 2 Starting with 1 
reactors reactors reactor 
DH with 65% comp. Adoption Adoption Adoption 
DH with zero comp. Adoption Adoption Adoption 
BB with 65% comp. Non-adoption Non-adoption Non-adoption 
BB with zero comp. Non-adoption Non-adoption Non-adoption 
BD with 65% comp. Non-adoption Non-adoption Non-adoption 
BD with zero comp. Non-adoption Non-adoption Non-adoption 
From the results presented in Table 67 and Table 68, it is concluded that under the current 
compensation level of 65% the adoption of on-farm control measures is generally resulting in a net 
gain for dairy farms, but for beef farms only if they get their reactor number down to zero. If the 
compensation level is reduced to zero, the situation is not altered significantly. The net gain in the 
dairy farms increases, the situation in the beef breeding farms changes hardly at all and on beef 
finishing farms the adoption of control programmes becomes beneficial if at least two reactors can 
be saved. 
Therefore it is concluded that a reduction of compensation level does not create a significant 
incentive for beef farmers to adopt on-farm control measures for TB. In addition, the removal of 
compensation could result in less co-operation by farmers, and therefore lead to results contrary to 
the intended ones. Thus other incentives or stringent rules and regulations, have to be put in place 
to encourage or force farmers to conduct their own control programmes. 
Bibliography 
Animal Health Board ( 1 995). Tb Beyond 2000. Towards a National Pest Management Strategy. A 
Discussion Paper prepared by the Animal Health Board. Animal Health Board, Wellington, 
New Zealand. 
Animal Health Board ( 1999). Movement from infected cattle herds. Documentation provided by 
the Animal Health Board Wellington, New Zealand. 
265 
Animal Health Board (2000). Bovine Tuberculosis National Pest Management Strategy 2001-20 1 1 ;  
Towards a TB free NZ; A discussion paper on the future options. Animal Health Board, 
Wellington, New Zealand. 
Coleman J., Livingstone P. (2000). Chapter Twenty: Fewer Possums: Less Bovine TB. Montague 
T.L. (editor). The Brushtail Possum. Manaaki Whenua Press, Lincoln, New Zealand, pp.:220-
231 .  
Cordes, D .O. ,  Bullians, J.A., Lake, D.E. and Carter, M.E. ( 198 1 ) .  Observations on tuberculosis 
caused by Mycobacterium bovis in sheep. New Zealand Veterinary Journal, 29: 60-62. 
Corner, L.A. ( 1994). Post mortem diagnosis of Mycobacterium bovis infection in cattle. Veterinary 
Mircobiology, 40 0): 53-63. 
Cullen R., Bicknell K. (2000). Chapter Eighteen: Economic Analysis of Possum Management. 
Montague T.L. (editor). The Brushtail Possum. Manaaki Whenua Press, Lincoln, New 
Zealand, pp. :  198-207. 
Davidson, R.M., Alley, M.R. and Beatson, N.S. ( 198 1).  Tuberculosis in a flock of sheep. New 
Zealand Veterinary Journal, 29: 1 -2. 
Davies G. ( 1 996). The Role of the Public Sector in Controlling the Epidemic Diseases of 
Livestock. Goodall E.A., Thursfield M.V.E. Society for Veterinary Epidemiology and 
Preventive Medicine. Society for Veterinary Epidemiology and Preventive Medicine, Roslin, 
UK. Proceedings University College, Glasgow, UK 27th, 28th, and 29th March 199678-83. 
Dijkhuizen, A. A. and Morris, R.S. ( 1 997). Animal Health Economics: Principles and Application. 
Post Graduate Foundation in Veterinary Science, University of Sydney, Sydney, Australia. 
Ekboir, IM. ( 1999). The role of the public sector in the development and implementation of animal 
health policies. Preventive Veterinary Medicine, 40 (2): 1 0 1 - 1 1 5 .  
Howard, W.E. ( 1964). Modification of New Zealand's flora by introduced mammals. Proceedings 
of the New Zealand Ecological Society, 1 1 :  59-62. 
Howe K.S., Whittaker J.M. ( 1997). Guiding decisions on methods and responsibilities for epidemic 
disease prevention and control: perspectives from environmental and insurance economics. In: 
266 
Goodall E.A., Thursfield M.V.E. Society for Veterinary Epidemiology and Preventive 
Medicine. Society for Veterinary Epidemiology and Preventive Medicine, Roslin, UK. 
Proceedings University College, Chester, UK 9th, 10th and 1 1  th April 1997223-235.  
Jackson, R.  ( 1 995). Transmission of tuberculosis (Mycobacterium bovis) by possums. Unpublished 
PhD thesis, Massey University, Palmerston North, New Zealand. 
Livestock Improvement Corporation Limited ( 1998). Dairy Statistics 1 997 - 1 998. New Zealand 
Dairy Board, Hamilton, New Zealand. 
Livingstone, P.G. ( 1 997). Update on the New Zealand TB situation. In: Proceedings of a Seminar 
on Possum and Mustelid Control Research. National Possum Control Agencies, Wellington, 
pp. 17-30. 
Maling, D. ( 1 993). A Farmer's Perspective of a Th Outbreak. Proceedings of a Deer Course for 
Veterinarians. New Zealand Veterinary Association Deer Branch, 10: 273-278.  
Marsh, w. ( 1999). The economics of animal health in  farmed livestock at the herd level. Revue 
Scientifique Technique Office International des Epizooties, 1 8  (2): 357-366. 
McInerney, J. ( 1996). Old Economics for New Problems - Livestock Disease: Presidential address. 
Journal of Agricultural Economics, 47 (3): 295-3 1 4. 
McInerney, J.P., Howe, K.S. and Schepers, J.A. ( 1 992). A framework for the economic analysis of 
disease in farm livestock. Preventive Veterinary Medicine, 1 3 :  1 37- 1 54. 
McKenzie, J. S .  ( 1 999). The use of habitat analysis in the control of wildlife tuberculosis in New 
Zealand. Unpublished PhD thesis, Massey University, Palmerston North, New Zealand. 
Meuwissen, M.P.M., Horst, H.S., Huime, R.B.M. and Dijkhuizen, A.A. ( 1 997). Insurance against 
losses from contagious animal disease. Epidemilogie Et Sante Animale, 3 1132: 10. 13 . 1 -
10. 13.3. 
Ministry of Agriculture ( 1 998). Farm Monitoring Report - North Central Region. MAP Policy, 
Ministry of Agriculture, Palmerston North, New Zealand. 
267 
Morris, C. ,  Loveridge, A., Fairweather, J.R. ,( 1 995). Understanding why farmers change their 
farming practices: the role of orienting principles in technology transfer. Research Report 
No.232. Agribusiness and Economics Research Unit, Lincoln University, Canterbury, New 
Zealand. 
Morris, R.S . ,  Pfeiffer, D.D. and Jackson, R. ( 1994). The epidemiology of Mycobacterium bovis 
infections. Veterinary Microbiology, 40 (1) :  1 53-177. 
O'Neil, B.D. and Pharo, HJ. ( 1995). The control of bovine tuberculosis in New Zealand. New 
Zealand Veterinary Journal, 43 (7): 249-255. 
Pannett, G. ( 1 995). Possum control in the Wellington region: how successful has it  been. In: Griffin 
F, de Lisle G (Tuberculosis in Wildlife and Domestic Animals - Otago Conference Series No. 
3. University of Otago Press, Dunedin, pp. 294-296. 
Parker, W. J. ( 1997). An introduction to financial budgeting for New Zealand agriculture. Institute 
of Veterinary, Animal and Biomedical Sciences, Massey University, Palmerston North, New 
Zealand. 
Parliamentary commissioner for the environment ( 1994). Possum Management in New Zealand. 
Parliamentary Commissioner for the Environment, Wellington, New Zealand. 
Ray, P. K. ( 198 1) .  Agricultural Insurance. Theory and Practice and Application to Developing 
Countries. Pergamon Press, Oxford, UK. 
Ryan, T. ( 1992). AHB Project Code: 57/90. Ministry of Agriculture and fisheries, MAF Quality 
Management, Ruakura, Hamilton, New Zealand. 
Sanson, R.L. and Thomton, R.N. (1997). A modelling approach to the quantification of the benefits 
of a national surveillance programme. Preventive Veterinary Medicine, 30: 37-47. 
Sauter, C.M. and Morris, R.S. (1995). Behavioural studies on the potential for direct transmission 
of tuberculosis to farmed livestock from feral ferrets (Mustela juro) and possums (Trichosurus 
vulpecula). New Zealand Veterinary Journal, 43 (7): 294-300. 
Siemienkowicz, H. (1984). Changed conditions for livestock insurance for private owners of 
livestock (in the GDR). [German]. Sozialistische Finanzwirtschaft, 38 (3): 30-3 1 .  
268 
Vaughan, E. J. and Vaughan, T.M. (1999). Fundamentals of risk and insurance. John Wiley and 
Sons, Inc., New York, USA. 
Vase, D. ( 1996). Quantitative risk analysis: Guide to Monte Carlo simulation modelling. John 
Wiley & Sons, Chichester, UK. 
Whipple, D.L., Balin, CA., Davis, A.J. et al. ( 1 995). Comparison of the sensitivity of the caudal 
fold skin test and a commercial gamma-interferon assay for diagnosis of bovine tuberculosis. 
American Journal of Veterinary Research, 56 (4): 4 1 5-419. 
269 
Beef farm in the Wairarapa (modelled in Chapter 7) 
Beef farm in the Wairarapa 
270 
27 1 
Chapter 7 
FarmORACLE, a farm simulation model 
272 
273 
Abstract 
Bovine tuberculosis control in New Zealand is complicated through the existence of wildlife 
reservoirs. The Australian brushtail possum is the main vector for transmitting the disease to 
livestock. Control of tuberculosis therefore is two-fold, control of livestock and control of infected 
wildlife. Large-scale poisoning operations are the main control method.  These are expensive and it 
is not possible to cover all areas with infected wildlife. Thus, progressively greater responsibility 
will be imposed on farmers to control TB on their own properties. Tuberculous possums are mostly 
located in specific TB hot-spot areas, not evenly distributed across the farm. On-farm control 
methods found to be effective (see Chapter 3) included targeted vector control and specific grazing 
management strategies to exclude cattle and deer from grazing hot-spot areas at high risk times 
during the year. Grazing routines are often based on past experience and tradition. For adoption of 
alternative grazing strategies, that take these TB issues into considerations, it is necessary to 
provide the farmer with likely outcomes of the changed grazing strategy. A computer simulation 
model for the whole farm and all animals present on the farm, was developed to assist the process 
of evaluating alternative grazing strategies. Outcomes used for the economic evaluation included 
liveweight of animals sold, and milk production on dairy farms. 
The use of the program is described on four farms in the Wairarapa, two dairy, one beef finishing 
and one beef breeding farm. In all four farms it was possible to find a grazing strategy that resulted 
in similar or even better economic outcome than the traditional grazing regime. 
Introduction  
Tuberculosis in cattle and deer is a disease o f  great concern for agriculture in New Zealand. 
Control of the disease is complicated through the existence of wildlife vectors, such as possums 
(Trichosurus vulpecula) and ferrets (Mustela Juro). Infected brush tail possum populations are the 
major source of TB infection for farmed animals (Jackson, 1 995; Morris and Pfeiffer, 1995; 
Livingstone, 1 997). The New Zealand cattle and deer industry is based on pastoral production and 
contact between livestock and feral animals is the most likely transmission pathway for TB (Sauter 
and Morris, 1 995). Tuberculous possums are not evenly distributed across  the farm but are 
localised in clusters, commonly referred to as 'hot-spots' (McKenzie, 1 999). Excluding cattle and 
deer from grazing areas of the farm where hot-spots are located during certain times of the year, 
was shown to reduce the incidence of disease in livestock (see Chapter 3). Grazing management is  
thus a useful tool to assist farmers control TB in cattle and deer. These hot-spot paddocks excluded 
274 
from grazing can be used productively in other ways, such as preserving the pasture in the form of 
silage or hay, or by grazing them with other stock, such as sheep. 
Often, grazing management on farms is based on tradition and many years experience. Therefore, it 
is necessary to convince farmers that changing their management can result in the same or an even 
better economic outcome than the traditional grazing plan. A computer model, simulating the 
individual farm with all its animals and stock numbers could assist the process of evaluating 
alternative grazing programmes and of convincing farmers to implement these. There are several 
whole farm or feed budgeting decision support systems available, such as for example FEEDBAL 
(Lodge and Frecker, 1 990), MIDAS, described by Pannell ( 1996), and PRO Plus (McPhee et aI. , 
2000) all developed for Australian conditions, and 'The Grazing Manager' developed by the Texas 
A & M University ( 1 997) for beef farmers in the USA. The only whole-farm decision support 
system for livestock farms in New Zealand is STOCKPOL, developed by the Ministry of 
Agriculture and Fisheries. It can be used to model different management strategies, long-term 
policy changes or short-term feed budgets, but it cannot be used for paddock-level feed budgeting 
(Marshall et aI. , 199 1 ). All these models focus on maximising animal and farm production, while 
the present study required a model that had the flexibility to incorporate disease control issues such 
as managing areas of potential contact between livestock and tuberculous wildlife. 
Research has shown that there are many factors influencing the efficiency with which pasture is 
utilised on a farm. Pastures are dynamic and complex plant communities with new tissue 
continually being formed through growth and old tissue disappearing through the process of 
senescence, death and decay (Matthews, 1 994). Therefore grazing management is very complex; 
for example, pasture cannot be carried forward between grazing periods. Residual pasture left after 
grazing will contribute to initial plant growth following grazing before disappearing through 
senescence and decay. If pasture is grazed laxly in an attempt to increase animal intake, then over 
time pasture qUality will deteriorate and net herbage production will decrease resulting in reduced 
production (Matthews, 1994; Matthews, 1 995). If the pasture cover was on average down to 
1200kg DM per hectare by the end of winter this led to low animal performance in early spring 
(Coutinho et aI. , 1 998). Therefore it was recommended that grazing management should be 
changed over the winter period to focus on achieving a desired sward condition for early spring, 
rather than improving liveweight gains in winter. Following this grazing management could then 
focus again on both pasture and animal performance (Coutinho et aI. , 1 998). 
The manipulation of grazing rotation has more to do with pasture utilisation than with pasture 
production. It is the grazing rotation that manipulates average pasture cover and allows pasture to 
be transferred within the system on the short term and determines pasture availability on a day to 
day basis (Matthews, 1994). Pasture quality is maintained by control of the pre grazing levels 
275 
rather than those after grazing. If pre grazing levels are too high, there will be increased losses and 
decreased nutritive value of the sward which will result in reduced intakes of lower quality feed, 
and reduced pasture growth rates (Matthews, 1995). 
Traditionally, grazing management is controlled on New Zealand dairy farms through the control 
of animal intake. Variations in pasture growth are overcome by adjustments to intake levels and 
therefore animal liveweight and production. This way, the grazing animal acts as a buffer in the 
pasture system (Matthews, 1994). However, with the current trend of having higher animal 
performance, particularly in the dairy industry, cow requirements must be met throughout the year 
to meet these higher production targets and therefore the grazing animal can no longer act as a 
buffer. Supplements are one form of replacing the cow as the buffer in the system, which is 
determined by the post grazing residual (Matthews, 1994; Matthews, 1 995). 
All these factors mentioned in relation to grazing management indicate that there are many options 
for a grazing management plan on an individual farms. A whole-farm simulation model, which 
models pasture growth and animal intake was therefore considered to be of major importance in 
help farmers determine which grazing strategies would achieve their desired outcomes. These 
outcomes might be purely financial or they may also be associated with disease control issues, such 
as reduced exposure of cattle to infectious tuberculous possums. Regardless of the outcome, any 
grazing management programme may have a positive effect on production (Matthews, 1994). Often 
farmers are reluctant to try new approaches, due to the uncertainty of the outcome (Morris et al. , 
1995). If different grazing strategies can be modelled, and the outcomes compared against each 
other, farmers might be more likely to adopt certain strategies. FarmORACLE was developed as a 
tool for this purpose. 
Overview of FarmORACLE 
FarmORACLE is a whole-farm simulation model, designed to allow veterinarians, farm advisors 
and herd managers to select grazing management plans that minimise the risk of livestock contact 
with TB-infected feral animals. It is intended to be a stand-alone system, operating under MS 
Windows, using data input from any farm recording program. Increasingly farmers are using farm 
recording programs such as FarmTracker® or Endeveour®, to record stock movements, animal 
treatments and/or paddock events. The program is written in the programming language Delphi. 
The use of FarmORACLE to develop a grazing plan that incorporates TB control as well as 
production outcomes, involves four main steps: 
276 
(i) setting up paddock maps and animal groups (mobs) for the individual farm, 
(ii) setting up the base grazing plan, according to the way animals are normally rotated 
around the paddocks, 
(iii) setting up an alternative grazing plan, in which cattle and deer are excluded from 
known or suspected TB hot-spot areas from November to January and during July and 
August, the high risk times for livestock coming into contact with moribund 
tuberculous possums (see Chapter 3), 
(iv) comparing the production outcome between the base plan and the alternative plan. 
The current model derives its data from the farm recording program, FarmTracker®. Within 
FarmTracker, a farm paddock map can be generated by digitising fencelines shown on an aerial 
photo. The program calculates the area of each individual paddock. Each individual mob can be set 
up with its own livestock parameters and targets. 
The model uses animal numbers and their gender, breed and age, plus paddock size, pasture species 
and weather data to predict pasture growth and animal intake, using metabolic equations. Each 
individual paddock or each individual mob can be modelled separately, or the whole farm and all 
mobs can be modelled together. 
The model' s  output shows actual animal production and target levels in both graphical and text file 
format, so that it can be used in spreadsheet programs, such as MS Excel. Reports include feed 
consumption, liveweightl liveweight gain, milk production and predicted grass cover. 
The model consists of three main parts: the pasture growth sub-model, the feed intake sub-model, 
and the drafting module (see Figure 59). The fIrst two modules are similar to feed budgeting 
models described earlier (Brookes et aI., 1992). 
277 
Operator Inputs 
Output 
Paddocks (daily) 
Pasture Growth Rates 
� -) Net Pasture Supply and Demand Evaluation (Optional I Growth Rates -? SupplyfDemand Soilfremperature/ 
Rainfall/Aspect/Slope) le Difference Area Farm Pasture Cover 
Comparison of 
� grazing plans r-l using 
Stock Age/Gender GRAZING Animal Intake summary outcomes 
PLANS Feed Intake Stock Breed I -) � le [7 Liveweight Production Targets ; Stock Numbers Liveweight gain 
Reproductive Targets ,I' Milk production 
Offspring liveweight 
Stock Numbers Offspring liveweight 
Stock Sales 
gain 
Stock Purchases !7 ---'l 
Drafting module 
Stock Transactions I (Stock numbers) 
Figure 59. Graphical display of FarmORACLE. 
The model uses deterministic equations to determine the pasture growth in each paddock, and the 
average potential and actual intake of animals in a mob. From these intakes it calculates the 
average outcomes per animal. The functions used in the model are based on unpublished and 
published literature, such as (Bircham and Hodgson, 1983 ;  Thomson et al. , 1 984; Butler, 1986; 
Butler et al. , 1987; Matthew et al. , 1993; Butler and Hodgson, 1 993 ;  Hodgson et al. , 1 996; Garcia­
Muniz et al. , 1998) The model was developed over several years, and will eventually be a 
commercial product. Therefore no specific equations are presented. 
The purpose of the stock drafting module is to adjust stock numbers gradually according to sales 
over time, for example when animals get sold after they have achieved a certain target liveweight. 
This allows sales to be entered with a time period over which a certain number of animals is sold. 
The drafting model works by assigning each animal an initial live weight, carcass weight and fat 
depth on the basis of pre-determined normal distributions and the co-variances between these. 
Animals are sold when they reach a certain liveweight, carcass weight or fat depth ranges. 
Following the removal of animals, the average liveweight for the mob is recalculated. 
Setting up paddocks and animal classes/mobs 
Digital paddock maps can either be imported in different software formats or created within 
FarmTracker by drawing over fencelines shown on a scanned aerial photograph on the computer 
screen (Figure 60). 
278 
.n �ilrml .ackc. urazmg PI'lOne. IS rAIN I UN - 19911/991 IUnllUed Ist.llllton.gmpJl I!!II'£JI3 
Grid: 25m (Of I) 
Figure 60. Creating a paddock map in FarmTracker. 
Although the main focus of the model is to compare different grazing management plans, it is 
necessary to enter all animal details, as each individual parameter could influence animal intake. 
The data that is entered for each mob is: 
Number 
Breed 
Age 
Gender 
Reproduction status 
Liveweight 
Stock production targets (l iveweight, milk production . . .  ) 
Reproductive targets 
Mating (how, when) 
Drying off (dairy farms) 
Calving/Lambing 
Stock movements into/from mob 
279 
As more farmers use farm recording programs, the necessary information will already be available 
in digital format. 
Table 69 presents the animal liveweight targets for dairy cows and dairy heifers. New Zealand 
pastoral livestock production is seasonal, i.e. calendar dependent, with most cattle and sheep being 
born in August each year. The annual production of a dairy cow was set to the average values 
achieved in the Wairarapa for the year 1 998/99. These were 259 kg milksolids, which was equal to 
3595 litres of milk, or 15 1  kg milkfat, using a percentage of 4.20% fat and 7.20% solids in milk 
(Livestock Improvement Corporation Limited, 1999). 
Table 69. Liveweight targets for dairy cows and heifers (in kg), used in FarmORACLE. 
Date Dairy cows Dairy heifers 
August 450 Calving 35 Birth 
November 70 Weaning 
March 120 
September 240 1 yr old 
October 433 Mating 280 Mating 
December 31 3 
March 344 
July 450 444 
September 456 2 yr old, calving 
Table 70 presents the liveweight targets for beef cattle and sheep, that were used in the model. 
280 
Table 70. Liveweight targets for beef cattle and sheep (in kg), used in FarmORACLE. 
Date Beef cows Beef heifers Beef steers Ewes Ewe Finishing 
replacements lambs 
August 450 (calving) 35 (birth) 40 (birth) 55 (lambing) 4 (birth) 4 (birth) 
October 455 (mating) 
December 1 35 53 (weaning) 25 (weaning) 25 (weaning) 
February 490 34 
March 485 (weaning) 120 181 56 (mating) 39 
May 480 236 37 46 
August 450 (calving) 282 55 (lambing) 38 
October 455 (mating) 280 (mating) 40 
December 31 3 445 53 (weaning) 50 
February 490 53 
March 485 (weaning) 344 491 56 (mating) 55 
Grazing plans in FarmORACLE 
Grazing plans can be set up after paddocks and animal mobs have been entered. The main aim of 
this series of simulations using FarrnORACLE was to compare the traditional grazing management 
plan on the farm with a revised plan, in which cattle and deer are excluded from grazing TB hot­
spot areas during certain high-risk times of the year (July and August and November to January; 
see Chapter 3). The fIrst two steps involved setting up the traditional grazing plan, and the 
alternative plan. 
Setting up grazing plans involved allocating each mob of animals to paddocks for set time periods 
(Figure 61), which can be as detailed as a daily basis. 
281 
Grazing Nobs . - -
IQI�ll r. Mob. r PI>ddock. 1997 MobN<me JuI I Aug ISep I Qct INoy IOec IJan I Feb I Mer I A!lr 110401' I Jun I ' Cows·Mike .. 94 Heife .. ' ' I '  I I  1 1  
:J � ____ . ____ .� I --- . . .. . --. --- - - - -95 I Icllc/S 00/ 1191 Ull 1 0/91 f Dates I E"ddocks T Slock 51",1 \ S/07/97 I� -'-�:-='--I i�_ FInilh 1 7/10197 I� 
I Sydem 
! Method I Rotale .:.U l [Residual OM :JI I R...wlll OM E§::] 
l - Pdk Oldel l List O,de, .:JI -� r + l .... ijin!rrlL::J -I' .i Min Tine 1 HlIIl dOl' .:11 L�-===:::::==�-==� Break fencing DaY0/8leak� r AtIow 
Dill" 0911  J98 1 7:28:00 L 
Figure 61 .  Setting up grazing plans in FarmORACLE. 
11 Paddock. ·  
, 
3 4 ') (, 
� II rs 
l i 1 h 
SI.63 ha 
I!!I • 
X Cancet 
? liet> 
Allocation of pasture to stock can be based on residual post-grazing herbage mass (kg DMlha), on 
the number of days in paddocks, according to rotational systems or set-stock grazing systems. 
Additional parameters such as pasture type, pasture growth rates, conservation of feed, and details  
on supplements (form of supplements, nutritional value, time of feeding, stock class being fed to  . . .  ) 
can also be set. If no specific pasture growth rates are known, default growth rates for areas within 
the region may be used. 
The program shows the location of each individual mob at any given time on the screen map, using 
different colours (Figure 62). 
282 
GUUlTl9 PI .... Sr'UJ.I • 
- S'4'lll Feod-1 --
'00 
o+-__ � __ � __ � __ ��� __ � 1� 30JgJ 2G11 1 2811 2W:J 2&IS 27n Date 
Hon: -46.9n, Vert 1084.Dn 
Figure 62. Location of animal groups on the farm map in FarmORACLE. 
_ [J x 
In the alternative grazing plans, cattle and deer are excluded from grazing hot-spot areas during 
November to January and July and August. On dairy farms, these paddocks were spelled, while on 
beef farms sheep were put into the hot-spot paddocks during these times. 
Output of Farm ORACLE 
The outputs for each individual grazing plan, shown in graphical and text file format, are: 
I ntake 
Liveweight 
Liveweight gain 
Milk production 
Reproductive status 
Number 
Supply and Demand 
Supply-demand d ifference 
F arm pasture cover 
Grazed paddocks 
Paddock cover 
283 
Figure 63 presents an example of the output for a dairy farm, showing animal intake (actual in 
green and potential in red) and fann pasture cover in kg DMlha (pre grazing cover in red, mean 
cover in green, and post-grazing cover in purple) . 
.. _---------==-:-----."..,------,------_._---- -------- _._-
Annual Grazing Model (Cows Milkers) f3 
Advanced Mob Model 1 1 
D ates 
� I '0 1'1 /07/97 1 to 1 1 /07/  ,I i 
Paddock - -
.--, ----.,,_ .... 1 , I Show 1 Wells 1 ....:J. 2.1 0  ha 
1 Growth I Events I Grazing 1 1 
o = 
6 
20.0..,.--------------
10 .0 -+--+--fr.fA:""""'I¥tI::-+-t--' ......... ----
5.0 -t--------------
I 
t 
'-Upper Graph--, 
r.- fintak� I (' 'Pr'og-:-intake 
" (- Liveweight 
r: LWI Gain 
(' Prog. LWG 
01107 30108 29/10 28/12 26102 27104 26106 I ,  -
Stock � 0 .0 +-�--'-�-""T�-r--"--r-�--'-�� I C Milk Prod Dat e  C Repra Status �:-�-=.=.�����============���=====.=,�- � r Numb� 
Class Cows M ilkers .;J 
Number 250 £ 
Uveweight (kg) 450.0 I ! 
Age (months) o 
Last Mating 1'1 /1 1/97 
Advanced Deta�s 
Supplement Feed [Calculate--
0 4-�--,-���--r--���_r--,-� 
-Lower Graph-
I ' r. Farm Cover t SupplylDem 
C- Grazed Pdks 
(' S,D Dill I , r Pdk Cover 
01 107 30108 29/10 28/12 26102 27104 26106 rGraph Values­
__ -,' � Smoothed Date 
r Show Key r Daily 
Figure 63. FarmORACLE output of animal intake and farm cover for dairy cows. 
Figure 64 presents the milk production for dairy cows in the top part of the graph and pasture 
supply and demand in the bottom part of the graph. During times when demand is higher than 
supply cattle have to be provided with supplementary feeds. 
284 
Annual Grazing Model [[oyjs-Milkers) . . ·13 
Advanced Mob Model 
lD�es I 
1 1 /07/97 1 to 1 1/07/� 1 J 
I Paddock -Show L...j1_W_e_lIs_1_--=�=1 
I 
2.1 0  ha 
Growth I Events 1 Grazing 1 1 
I 
Stnr.1c 
Class I Cows Mikers .:J 
r�umber 250 4-• 
Liveweigh: (kg) 450.0 + 
Age (mont1s) 0 rt 
Last Mating 1 1 /1 1 /97 -I 
Advanced Details I 
Supplement Feec [CalCulate 
"....... "0 ---""C .c. 
-:;:: 
.2 0> 
e 
""C 
� 0.. 
� 
� 
q 
1 1 "....... ""C ---Cl! .c. 
I 
---� Ll al 
e 
Cl! > Q 
E Q) 
0:: 
1 .2 
0.9 
0.6 
0.3 ��--------------�--��� 
0.0 0 1107 30108 29/ 10  28/12 20102 27t:l4 20100 
Date 
.'.-� . .  ' .- ..... 
75 
60 
'10 
30 
1 5  
::10100 29/10 �0/12 26102 271)4 26106 
Date  = 
r Show Key 
UIJIJt:1 G ldlJh 
(� firta"i(·a 
(-Piog.·lntake 
C U>'eweight 
( LY.lt Gain (" Plog. LWG 
(! Milk Prod. 
C R3pro Status 1 
f N.Jmber I 
I Lower Graph- I C !-arm Cover 
I 
c:- Supply/Den 
t G·azed Pdks 
C S·D Diff 
f Pdk Cover 
1 1 Grap1 Values 
� 
I 
C'f Smoothed 
t D3ijy 
Figure 64. FarmORACLE output of animal liveweight plus supply and demand for dairy 
cows. 
The model was verified by checking the biological correctness of the outcomes for each mob 
separately. Modifications were made to the program until the model outcomes were logical, correct 
and satisfactory. However, all mobs were modelled together to compare the different grazing plans. 
If only one mob was modelled at a time, the model assumed pre-grazing levels according to the 
pasture growth without removal by other mobs. When modelling all mobs together the model took 
into account the post-grazing levels of other stock classes that might have been grazing the same 
paddocks previously. 
The text file outputs of the model for the traditional and the alternative grazing plans were imported 
into a spreadsheet for further manipulation and comparison. Economic analyses were conducted 
within the spreadsheet using price schedules as published by the Meat & Wool Economic Service 
of New Zealand ( 1 999). 
285 
Use of FarmORACLE on some of the Wairarapa focused control 
farms 
Information on grazing routines and management was collected on all 34 farms where on-farm TB 
control interventions were conducted within the Wairarapa project (see Chapter 3 for more details). 
However, the detail obtained varied considerably between farms and years. The most detailed 
grazing information was obtained for the year July 1998 to June 1 999. Thus ,  this farming year was 
used to model the farms. Only cattle farms were included in the study. The models used on two 
dairy farms, one beef breeding and one beef finishing farm are presented here in detail. The 
traditional grazing plans (base plans) were as close as possible to the grazing routines implemented 
by these farmers . 
In setting up the grazing management plans in FarmORACLE for these project farms, no allowance 
was made for including additional forage crops.  When taking into account the losses in production 
while the paddock is in forage crop and new pasture is being established, forage crops are a means 
of redistributing feed within the year rather than significantly increasing total herbage production 
on a whole farm basis (Matthews, 1994). 
Comparing grazing plans on dairy farms 
Setting up grazing plans on dairy farms was less complex than on beef farms, as in most cases 
cattle were the only species grazed on the farm. In addition, the typical dairy farm only grazed 
dairy cows on the main farm during the lactation period, while young stock were grazed off farm, 
either on run-offs or on other people's land. Therefore, if the TB hot-spot was on the home farm, 
only dairy cows had to be modelled for the period August to May. 
Normally, all paddocks were used on a rotational basis on dairy farms. In the alternative grazing 
plan all paddocks except the ones with a TB hot-spot were used. 
Dairy farm A 
Dairy farm A was located in the south of the Wairarapa, with a total of 280 milking cows, 60 rising 
one-year heifers, and 60 rising two-year heifers . All male calves and surplus female calves were 
sold to slaughter as three-day old calves. No other stock was kept on the farm .  The farm comprised 
a total of 1 24 ha, of which 93 ha were on the main farm, and 3 1  ha on the run-off. 
Due to the high stocking rate, maize silage was usually fed to the milking cows during September 
to April at approximately 5kg DM/head/day. This supplementation had to be considered when 
modelling this farm. The time and amount of maize silage fed was used equally in all grazing plans. 
286 
modelling this fann. The time and amount of maize silage fed was used equally in all grazing 
plans. 
The main farm was situated along a river. Cattle TB testing data indicated that this riverbed was a 
suspect TB hot-spot area. This area was generally used as part of the milking platform and/or for 
one-year heifers. Figure 65 presents the paddock map and the potential hot-spot area (in grey). 
Paddocks 42 to 54 are on the run-off, and are not drawn to scale. They are only shown symbolically 
for presentation. However, the actual paddock sizes on the run-off were entered in the model. 
Pasture growth was set at a lower level for the run-off paddocks than for the main farm. No 
irrigation was used on the fann. 
Figure 65. Paddock map of dairy farm A, with TB hot-spot area in grey. 
Grazing plans used in the model for Farm A 
Figure 66 presents the traditional grazing routine during the summer of 1 998/99 and the winter of 
1 998. Both TB hot-spot areas were included in the grazing area. During summer the four-month 
old calves were put into the paddocks along the riverbed on the main property, while the 1 6-month 
old heifers rotated around paddocks on the run-off, including the hot-spot paddocks. During winter 
no animals were grazed along the riverbed, however, the dry cows were grazed in all paddocks on 
the run-off 
• Cows-milking 
• Cows-dry 
• 98 Heifers 
• 97 Heifers 
Base pllln 
December 
• Cows-milking 
Cows-dry 
• 98 Heifers 
• 97 Heifers 
Base plan 
July 
287 
Figure 66. Grazing routines during summer and winter on dairy farm A, using traditional 
grazing plans. 
These grazing plans were altered to exclude the TB hot-spot paddocks from grazing during the high 
risk times (November to January and July and August, see Figure 67). 
• Cows· milking 
Cows·dry 
• 98 Heifers 
• 97 Heifers 
Alternative plztn 
December 
• Cows· milking 
Cows-dry 
. 98 Heifers 
• 97 Heifers 
Alternative plan 
July 
49 49 
Figure 67. Alternative grazing routines during summer and winter on dairy farm A, 
excluding TB hot-spots from grazing at these times. 
Table 7 1  presents the production outcomes of the model using the base and the alternative grazing 
plans. S lightly less milk was produced over the whole lactation period in the alternative grazing 
plan, but the cull cows were heavier than in the base plan. 
288 
Table 71. Production outcomes in modelling traditional (base) and alternative grazing plans 
on Dairy farm A. 
Date Number of Base plan Alternative plan 
animals 
Total milk produced (kg 280 62,056.4 61 ,81 5.1 
milksol ids/yr) 
Average milk/cow (kg 280 221 .63 220.77 
milksolids/cow/yr) 
Liveweight of milking cows (kg) 1/5/99 250 549.7 547.7 
Liveweight of cull cows (kg) 1/4/99 30 441 .6 467.3 
2/5/99 30 444.0 467.7 
Total kg cull cows sold 26,568 28,050 
Liveweight 97 Heifers (kg) 30/6/99 60 445.9 454.3 
Liveweight 98 Heifers (kg) 30/6/99 60 194.7 223.9 
Using a milk company pay-out of $3 .42 per kg milksolids (Ministry of Agriculture, 1 999), the farm 
lost $825 .25 in income from milk over the year under the alternative grazing plan. However, under 
this plan the total weight of the cull cows was increased by 1 ,482 kg, which was equal to 74 1 kg 
meat, using a dressing out percentage of 50% for the cull cows (Fleming, 1 996). The average price 
paid to farmers in the month of sale was used for calculating the income from cull cow sales (Meat 
& Wool Economic Service of New Zealand, 1 999) . This resulted in an increased income from cull 
cows of $ 1 ,442.9 1 .  Thus, the implementation of the alternative grazing plan resulted in an overall 
net benefit of $6 1 7 .66. No additional time input had to be considered on this farm when 
implementing the alternative grazing plan versus the traditional plan. 
Dai ry farm B 
Dairy farm B was situated in central Wairarapa in a typical dairy farming area. The farm milked 
1 90 cows, and had 40 rising two-year heifers and 40 rising one-year heifers. Calving started in July 
and all male and surplus female calves were sold to slaughter as three-day old calves. No other 
stock was kept on the farm. In total the farm had 1 06 ha effective grazing area, of which 70.9 ha 
were on the main property and 35 ha on the run-off. Traditionally the new born heifers were kept 
on the main property until they were weaned in NovemberlDecember, when they were moved to 
the run-off. The rising two-year heifers were grazed on the run-off during the lactation period. 
After the adult cows were dried off in May they were shifted to the run-off while all heifers 
returned to the main property. 
289 
The TB hot-spot on this farm was located on the main property. It was traditionally used all year 
round, for milking cows as part of the milking platform, or for young stock in winter. Figure 68 
shows the paddock map and the TB hot-spot in grey. Paddocks l abelled 'rI '  to ' r I O '  comprised the 
run-off, which was situated about 1 km south of the main farm. These p addocks were not drawn to 
scale, but serve only as reference. However, the actual paddock sizes were entered for correct 
pasture growth. Pasture growth rates for the run-off paddocks was lower than for the main property 
paddocks. No irrigation was used on the farm. 
[6 17 18 19 1 10  
Figure 68. Paddock map and TB hot-spot (in grey) on dairy farm B. 
The farm was modelled from July 1 998 to July 1 999. Under the traditional grazing routine, all 
paddocks on the main property, including the TB hot-spot areas, were included in the rotation of 
milking cows between July and April. In the alternative plans these four hot-spot paddocks were 
excluded from the milking platform for the months July and August and November to January. All 
other grazing routines were kept the same. In all grazing plans the replacement heifers were kept on 
the run-off during the milking season, and were shifted on to the main property during winter while 
the dry cows were on the run-off. 
The farmer traditionally used rotational grazing, whereby the paddocks were grazed in the same 
order (list order) per rotation cycle. This grazing scheme was modelled in the plan 'Base List' . A 
second grazing plan was set up, using exactly the same criteria (rotation lengths, p addock 
allocations) as in the 'Base List' plan, except that paddocks were grazed from highest to lowest 
cover available ('Base High'). In the alternative plan ( 'Alter') the paddocks were grazed as in 
'Base High' from highest to lowest cover, but the four hot-spot p addocks were excluded from 
grazing during the high risk months. Table 72 presents the production outcomes for the three 
different plans used on dairy farm B .  
290 
Table 72. Production outcomes in modelling traditional (,Base List' and 'Base High') and 
alternative grazing plans on Dairy farm B. 
Date Number 'Base List' 'Base High' 'Alter' 
of Plan Plan Plan 
animals 
Total mi lk produced (kg milksolids/yr) 1 90 46,592.31 52,938.95 52,1 81 .59 
Average milk/cow (kg milksolids/cow/yr) 1 90 245.22 278.63 274.64 
Liveweight of milking cows (kg) 1/5/99 1 55 432.3 457.0 456. 1  
Liveweight of cull cows (kg) 1 5/2/99 5 506.4 506.4 506.4 
1 5/4/99 6 551 .4 551 .4 551 .4 
1 5/5/99 4 556.9 556.7 556.5 
Total kg cul l  cows sold 8,068.0 8,067.2 8,066.4 
Liveweight 97 Heifers (kg) 30/6/99 44 460.8 464.1 459.4 
Liveweight 98 Heifers (kg) 30/6/99 43 200.4 221 . 1 234.2 
The outcomes of the model indicated that the farm could increase its milk production by 1 2% just 
by changing the paddock order in which the milking cows were grazed. This would require the 
farmer to measure pasture height in order to identify the paddock with the highest pasture cover to 
be grazed next. 
Milk production was slightly decreased ( 1 .4% less than the 'Base High' plan outcome) by 
excluding the four TB hot-spot paddocks. 
Using a dressing out percentage of 50% (Fleming, 1 996) and the price schedules as paid in 
February, April and May 1 999 (Meat & Wool Economic Service of New Zealand, 1 999) resulted in 
the economic outcomes presented in Table 73.  The alternative plan resulted in a net loss of about 
$2,500 in comparison with the income achieved when implementing the grazing plan 'Base High ' .  
However, in comparison t o  the traditional grazing plan used on  the farm ( ,Base List ' )  this 
alternative plan still resulted in a net benefit of over $ 1 9,000. When using a grazing plan where 
milking cows were grazed in the paddock with the highest pasture cover, time for measuring the 
pasture height in each paddock had to be considered. It was assumed that it would take the farmer 
one hour per day to measure all paddocks. This then resulted in an overall time input of 304 hours 
(304 days during the lactation period July 98 to May 99). Even if farmers judged the value of their 
hour at $ 1 5 ,  this would still result in a net benefit of $ 1 4,550. 
29 1 
Table 73. Economic comparison between grazing plans used on Dairy farm B .  
'Base High' versus 'Base 'Alter' versus 'Base High' 
List' 
Schedule ($) Difference in  Price difference Difference i n  Price difference 
carcass weight ($) carcass weight ($) 
(kg) (kg) 
Pay-out/kg 3.42 6,346.63 21 ,705.48 -757.36 -2,590. 1 7  
m ilksolids 
Price Feb99 per 1 .948 o (n=5) 0 o (n=5) 0 
kg meat 
Price April99 1 .965 o (n=6) 0 o (n=6) 0 
per kg meat 
P rice May99 per 1 .928 -0.1 (n=4) -0.77 -0 . 1  (n=4) -0.77 
kg meat 
Net outcome + 21 ,704.71 - 2,590.94 
Comparing grazing plans on beef farms 
There are many more stock classes and groups of cattle on beef farms compared with dairy farms, 
resulting in more complicated grazing plans .  It was found that modelling all stock classes/mobs, 
even if some mobs were never grazing the TB hot-spot areas, provided the most reliable model 
outcomes.  This way liveweights, pre- and post-grazing pasture covers were calculated for all stock 
classes and the interactions or effects of one stock class on others were taken into account. 
Beef breeding farm 
The beef breeding farm used in this example was located in  the north-eastern part of the Wairarapa. 
Figure 70 presents the paddock map on top of an aerial photograph (Figure 69). Figure 7 1  presents 
j ust the paddock map and the suspected TB hot-spot in grey (paddocks 94 to 97). The dark green 
areas (paddocks 36 to 38)  were pine plantations, where animals were rarely grazed. These three 
paddocks were not in pine yet when the aerial photograph was taken. All paddocks in the northern 
part of the farm, with numbers 65 and higher, were on steeper hill country than the paddocks in the 
southern part of the farm. As a result, a lower pasture growth rate was assigned to the paddocks in 
the northern part of the farm. 
292 
Figure 69. Aerial photograph of beef breeding farm. 
293 
Figure 70. Aerial photograph of beef breeding farm with paddock layout. 
294 
Figure 71. Paddock map on beef breeding farm (TB hot-spot in grey, pine plantation in dark 
green). 
The farm was modelled for the time period 1 st May 1 998 to 1 st May 1999. It consisted of a total of 
1420 ha land of which 1 145 ha was effective land. At the beginning of May 1 998 stock numbers 
were: 200 mixed age cows, 63 rising two-year heifers, 128 rising one-year heifers, 9 1  rising one­
year steers, 4600 mixed age ewes, 1 860 rising two-year ewe replacements, 2200 rising one-year 
ewe replacements and 82 rams. Ewes lambed from early September onwards and weaned l 30% 
lambs in late December. All lambs were sold at or shortly after weaning, except for the ewe 
replacements. The ewe replacements were mated at an approximate age of 20 months and were put 
into the mixed age ewe mob before lambing at an age of two years. The heifers were mated at 1 6  
months of age and kept separate fi:om the mixed age cows until they were three years o f  age. 
295 
Figure 72 presents the grazing locations of cattle and sheep during the high-risk time of August 
1 99 8 .  Cattle were put into blocks of paddocks and were kept there until the desired pasture residual 
was obtained, when they were shifted to the next b lock of paddocks ('set stocking '). Sheep 
generally rotated around their blocks of paddocks. Under the traditional grazing plan the beef cows 
were grazed in the hot-spot area from 1 SI June 1 998 to 3 1  si August 1 998. During summer no cattle 
were kept in or around the hot-spot areas. Therefore the grazing only had to be altered to 
accommodate the winter high-risk period. 
Base plan August 
Figure 72. Grazing locations of cattle and sheep, using traditional grazing plans, for August 
1998 on the beef breeding farm. 
Under the traditional grazing regime, none of the young stock was kept on the northern part of the 
fann during the winter. Thus, in the alternative plan the cows were simply allocated to other 
paddocks in the northern part of the farm during the high risk time (for example using paddocks 
8 5/87/88 instead of the hot-spot paddocks 94/95/96 as in alternative grazing plan ' Alter' ) .  
Exchanging these paddocks only had minimal effect on the production outcomes (Table 74). The 
liveweight of all animals remaining on the fann were similar in both plans. 
Information was available on the dates when animals were sold. These dates were used in both the 
traditional and the alternative grazing plan to determine the average liveweights of animals sold at 
that particular time. Using the prices paid for cattle and sheep at that time, and a dressing out 
percentage of 53% for cattle and 4 5 %  for sheep, the adoption of the alternative grazing plan 
296 
resulted in a net benefit of $2,054 (Table 74). Although the cull cattle were s lightly lighter than in 
the traditional grazing plan, the sheep, mainly the male lambs (wethers), had heavier weights. 
Table 74. Summary of cattle and sheep liveweights for sale using different grazing plans. 
Base grazing Alternative Difference Economic value of 
plan grazing plan difference ($) 
Cattle 76,01 9.7 kg 75,894.1 kg -1 25.6 kg - 1 30.70 
Sheep 230,705.4 kg 232,649.0 kg + 1 ,943.6 kg + 2,1 85.07 
Net outcome + 2.054.37 
The only change made in the alternative plans was to exchange three hot-spot paddocks for three 
adjacent paddocks during July and August for the mixed age cows. Thus, no additional 
management was required to implement this change. 
Beef f in ishing farm 
The beef finishing farm was situated in the south of the Wairarapa and bordering on its east side to 
Lake Wairarapa. The farm comprised 330 ha on the main property, of which 274 ha were effective. 
Additionally the farm had a lease block of 35 ha, which was not modelled in the current study, as 
animals usually were shifted there and from there were sold directly to slaughter, without returning 
to the home property. The farm had no beef breeding cows; all cattle were bought in with the sole 
purpose of finishing them to slaughter. 
Figure 73 shows the paddock map of this farm. The area north of the hot-spot paddocks (grey 
shaded in the figure) was bush and not used for grazing. In the traditional grazing scheme cattle 
were spread over the whole farm, except the three hot-spot paddocks. In the alternative grazing 
plan the cattle were excluded not only from these paddocks, but also from hot-spot neighbouring 
paddocks (numbers 1 4  to 18) .  
297 
Figure 73. Paddock map and TB hot-spot (in grey) of the beef fInishing farm. 
The fann was modelled for the period July 1 998 to June 1 999. The stock numbers on the home 
property at the beginning of July 1 998 were: 24 rising one-year old heifers ( 1997 heifers), 5 1  rising 
one-year steers ( 1 997 steers), 750 mixed age ewes, 600 rising three-year ewes, 720 rising two-year 
ewe replacements, and 230 rising one-year ewe replacements. In October 1 998 another 43 rising 
two-year heifers ( 1 997 heifers) were bought, in March 1 999 23 rising one-year heifers ( 1 998 
heifers) and 22 rising one-year steers ( 1998 steers). 
From March to August 1999 the 1997 steers and between March and December 1999 the 1 997 
heifers were gradually sold according to their weight (all heifers with a weight greater than 520 kg 
and all steers with a weight of greater than 570 kg, drafted every 30 days). All young sheep for sale 
were taken to the run-off property fIrst and sold from there. The only sheep sold from the main 
property were 230 old ewes (termed 'gummies'), that were drafted and sold if their liveweight was 
over 52kg. 
Figure 74 and Figure 75 present the grazing locations of sheep and cattle groups during the high 
risk times of August and December 1998. Cattle and sheep were presented separately, as some 
cattle and sheep mobs were grazing paddocks simultaneously. During the high risk time in August 
298 
only the 97 steers were in paddocks adjacent to hot-spot paddocks, while during the December high 
risk period the 97 steers and the 97 heifers were in these paddocks. 
August 1998 
Sheep 
MA. Ewes 
4th 
Figure 74. August grazing plans under the traditional grazing scheme for cattle and sheep on 
the beef f"mishing farm. 
Cattle 
97 Heifers • 98 Heifers 
97 steers 
December 1998 
Sheep 
Figure 75. December grazing plans under the traditional grazing scheme for cattle and sheep 
on the beef f"mishing farm. 
In the alternative plans the sheep grazing pattern wa,s kept as ill the traditional plans, but the 
grazing for the 97 heifers and the 97 steers was changed to exclude paddocks 14 to 1 9  during high 
risk times (Figure 76). 
August 
ZJ 
Cattle 
97 Heifers . 98Hellers 
97 Steers 
98 Steers 
December 
299 
.. 
45� 
i' { 41 " 
46 47 38 42 
r 
Figure 76. Alternative grazing plans for cattle during high risk times (winter and summer). 
An alternative plan was set up (Alter), whereby the cattle were allocated to other paddocks not next 
to the TB hot-spot areas. The only criteria set for the sheep mobs was that they stayed above their 
target liveweights. Table 75 presents the liveweights of cattle and sheep and the dates these were 
drafted for sale, for the base plan and the alternative plan. 
Table 75. End liveweights of cattle and sale ewes for three different grazing plans. 
Base plan Alternative plan 
Date Number of Total Number Total 
drafted animals liveweight of animals liveweight 
drafted (kg) drafted (kg) 
97 Heifers 31 /3/99 7 3,676.2 43 22,774.4 
30/4/99 42 22,840.0 24 1 2,920. 1 
30/5/99 1 8  9,780. 1 
97 Steers 31/3/99 24 1 3,870.2 51  30,531 .0 
30/4/99 18  1 0,377.3 
30/5/99 7 4,062.2 
28/6/99 2 1 . 1 54.0 
Total kg for sale from cattle 65,204.0 66,225.5 
Cull ewes 3 1/1/99 1 98 1 1 ,3 10.4 221 1 3,312 .2 
2/3/99 8 41 8.6 0 0 
1 /4/99 1 7  908.2 3 1 56.7 
1 /5/99 7 387.9 6 331 .2 
Total kg for sale from cull ewes 13,025.1 1 3,800.1 
300 
The outcomes of the model indicate that the alternative plan achieved greater production as the 
traditional grazing plan 'Base' , without putting cattle at risk of getting infected by tuberculous 
possums in high risk areas. Both cattle groups, 97 Heifers and 97 Steers were sold quicker when 
using the alternative grazing plan than when using the base plan. Both plans used the smae drafting 
criteria. 
Using the dressing out percentages of 53% for cattle and 45% for sheep (Flerning, 1996) and the 
price schedule for the months the animals were drafted and sold (Meat & Wool Economic Service 
of New Zealand, 1999), the adoption of the alternative plan resulted in a net loss of -$ 140 .42, 
despite the fact that the alternative plan produced a higher liveweight of cull animals (Table 76). 
The net loss resulted as a consequence of the price schedule for beef, which increased in April, 
when most of the animals in the alternative plan were already drafted and sold. However, in 
deciding which grazing plan to adopt, one also has to consider the pasture being available for other 
stock classes once the cull animals have been removed. In the alternative plan, all paddocks used 
by the 97 steers would be freed up by the end of March, while in the base plan the paddocks were 
used until the end of June. 
Table 76. Difference in carcass weights and economic outcome comparing different grazing 
plans on the beef finishing farm. 
Base plan Alternative plan 
Total carcass Economic value Total carcass Economic value 
weight weight 
Heifers 18,942.4 36,856.54 18,91 8.1 36,099.77 
Steers 1 5,61 5.8 35,937.66 1 6, 1 81 .4 36, 1 81 .68 
Cull ewes 5,861 .3 5,872.89 6,2 10 . 1  6,245.22 
Total 78,667.09 78,526.67 
In beef farms the program might also be used to compare an ad hoc grazing management, with a 
detailed grazing strategy. Some farmers move their cattle only, when they see, that the pasture is 
grazed down, often resulting in a residual pasture cover of less than 1 ,000 kg DMlha. For this 
purpose, a grazing plan with a residual pasture cover of 800 kg DM was compared with one using 
1 ,600 kg DMlha, all other parameters were held constant. As Table 77 shows, the low residual 
grazing regime resulted in all liveweights being lower than if using the residual pasture cover of 
1 ,600 kg DMlha. 
301 
Table 77. Comparison of grazing regimes using 800 versus 1600 kg DMlha residual. 
Livestock weights at 30/6/99 
Residual 800 kg DMlha Residual 1 600 kg DMlha 
97 Heifers 486.4 570.8 
98 Heifers 227.2 253.6 
97 Steers 527.6 600. 1  
98 Steers 282.8 326.5 
MA Ewes 54.7 55.6 
4th replacement ewes 56.7 58.6 
2th replacement ewes 49.9 53.6 
Cull ewes 63.9 64.7 
Discussion 
The FarmORACLE software model can be thought of as  a decision support system, as it attempts 
to support decision-making using many variables ,  and allowing users to combine their 
epidemiological knowledge about tuberculosis in possums and livestock with biological, physical 
and economic factors. Using the model to evaluate the production and economic outcomes of 
different grazing strategies showed that it is not impossible to fmd a grazing regime that excludes 
TB hot-spot areas at equal or better production/financial outcomes. FarmORACLE models 
different grazing strategies in order to prevent possible direct contact with terminally ill 
tuberculous possums, a method which was suggested as main transmission pathway between 
possums and livestock (Sauter and Morris, 1995). 
In all four farms a grazing regime was found that excluded TB hot-spot areas from grazing by 
cattle during high risk times, with increased profitability. The alternative grazing strategies 
presented for each farm are only one of a multiple range of potential combinations of paddocks, 
animals and time periods. In the grazing plan on the beef breeding farm it was found that the sheep 
were gaining more weight in the alternative grazing plan. This was due to the fact that the sheep 
were grazing exclusively the hot-spot paddocks instead of sharing them with the beef cows. The 
increased sheep liveweight resulted in increased carcass weights of lambs,  which more than 
compensated for the slightly lower liveweights of cull cattle. The results on the beef fmishing farm 
indicated that the alternative plan led to a shorter fattening period for heifers and steers, however, 
the fmancial outcome was slightly reduced by $140 due to the lower price schedule for cattle at the 
302 
earlier sale date. The knowledge of the price schedule could also be in importaIit additional point 
when choosing grazing strategies. 
However, the model not only showed that grazing plans which exclude cattle from TB hot-spots at 
certain high risk times, can result in an equal or improved economic outcome, but it also indicated 
other areas of improvement. The results on dairy farm A, using the alternative grazing plan resulted 
in a higher dry-cow liveweight. Although feeding dairy cows better in winter results in an increase 
in total feed demand per cow over the winter, this was offset by the reduction in stock numbers and 
the shorter wintering period resulting in an overall reduction in supplement requirements over 
winter (Matthews, 1997). The result on dairy farm B ho ed that u ing pasture height, athe than 
the paddock order, to determine the grazing rotation on a dairy farm can result in a high increase in 
milk production, without any other changes. Using the height of the pasture uses the fact that high 
pre-grazing pasture covers will result in an increased decay and decreased nutritive values 
(Matthews, 1995). 
The model can also be used as a farm management tool for economic evaluation of 'what if 
decisions. For example the model could be used to evaluate the effect of reducing cattle numbers 
and increasing sheep numbers on beef farms, or comparing the performance of steers versus bull 
fattening or comparing grazing strategies with low versus high post-grazing levels. Results on the 
beef finishing farm, comparing grazing regimes with 800 versus 1 600 kg DM/ha post-grazing 
residual, showed that the high residual resulted in all animal groups having higher end liveweights. 
This would be in accordance with the quality of pasture declining if pasture covers gets too high 
(Matthews, 1995). Often farmers use an ad-hoc grazing regime, shifting cattle only once the 
pasture is completely grazed. The results of the model indicate that this management decision is not 
optimal. 
The use of a computer model such as the one presented, allows the farmers and other decision 
makers to evaluate different grazing strategies in terms of their production and economic outcomes. 
In agriculture more and more computer models are developed, however, they mostly focus on one 
specific part of the enterprise to try to maximise its profit. Several computer programs are available 
for dairy farms (Larcombe, 1990; Uribe et aI. , 1996; Lockhart et aI., 1997) and for pig farms 
(Vaillancourt et al. , 1992; FarmPro Systems Limited, 2000). Less programs are available for sheep 
enterprises (Gray et al. , 1 992; Parker et aI. , 1992; Parker et al. , 1998) or beef enterprises (Bircham 
and Sheath, 1986). The present model not just accommodates one particular livestock species, but 
incorporates all the different livestock enterprises present on farm. Additionally, the model can take 
epidemiological knowledge relating to tuberculosis into account. A similar combination of 
epidemiological knowledge and pastoral systems was also used in a model of ovine fascioliasis in 
sheep in Australia (Meek and Morris, 1981 ). 
303 
Any software developed for agricultural enterprises is intended to assist farm management in 
allocating limited physical, financial and human resources in order to achieve certain objectives 
(Parker et aI. , 1994). The increasing complexity of farm management and decreasing profit margins 
require more information to enhance decision-making. Therefore, more and more farmers will use 
farm recording programs, and information on existing grazing management will be readily 
available. 
Apart from being able to model all animals present on farm, another main advantage of the 
program is the representation of individual farms, rather than an 'average' farm. Especially in the 
sheeplbeef farms it was found that the 'average' farm may not be relevant to the particular case 
(Morris et aI. , 1995). With FarmORACLE each individual farm can be modelled, by entering farm 
specific parameters, such as livestock details and targets, pasture growth rates and paddock 
information. 
There are varying degrees of uptake of simulation or mathematical models. Some models 
contribute to research and understanding, but not to management decisions. For a model to be used 
in practice requires that it addresses a definite management question, and that it is actively 'sold' , 
which in turn necessitates an easy to understand graphical display and proactive demonstrations of 
the use to potential users (Barlow, 2000). In order to be used the model should be easy to use, 
flexible and capable of being linked with other programs (Brookes et aI. , 1992). One of the 
advantages of the present model is the use of many graphical features, facilitating use. For 
example, by using the paddock map and different colours for the individual stock classes, it is easy 
to describe stock movements on the farm. Rather than having to enter the paddocks on a purely 
tabular basis, such as in PRO Plus (McPhee et aI. , 2000), the end user can see the map of the 
individual farm and enter data through this interface. 
Studies in Australia on the uptake of innovations in general found that the unpredictability of 
income discouraged the adoption of new techniques (Anderson, 1982). With a program such as 
FarmORACLE it is possible to reduce the uncertainty of the outcome. Thus the uptake of control 
programmes like adopting grazing regimes that exclude cattle and deer from grazing TB hot-spot 
areas, should be increased. 
Regarding the adoption of decision support systems Lynch et al. (2000) suggested, on the basis of a 
British study in 1996 (OASIG, cited in Lynch et al., 2000) that the success of a product depends on 
the interrelationships between software developers, potential adopters of the product and the 
contexts in which the software is developed. In the present study, traditional grazing plans were set 
up with the particular farmers on ten farms, at the computer, using the actual model. These first 
304 
applications of the software resulted in many suggestions by farmers to improve the ease and 
usefulness of the model and thereafter alterations of the software. 
In using the model, the only knowledge required is about the grazing routines for the different 
stock classes on farm, information that is known by farmers. No additional information is required, 
but can be entered optionally if available. In an Australian study it was found that the low adoption 
rate of computer software within the agricultural system is also due to the fact that farmers have to 
provide data that they do not normally collect (Glyde and Vanclay cited in Lynch et al. , 2000). 
Within the current program the information required is a minimum and is readily available to 
farmers. Furthermore when modelling the traditional grazing strategies on farms, the farmers can 
compare the model outcome with the outcome they perceive their strategy achieves. If this 
comparison results in outcomes close to field reality, then the trust in the output of the model using 
alternative grazing strategies might be increased. This trust was found to be necessary for 
successful adoption of innovations, as the farmers have to make a choice between their years of 
experience and outcomes of a computer model (Lynch et aI. , 2(00). 
In order to keep the model in the present study as simple and generalisable as possible, only 
minimal field-gathered information was used. However, the program has many features already 
included that were not used in the present study, such as entering slope and aspect of each 
individual paddock, specific temperatures and rain fall data in order to calculate the pasture growth 
more exactly, conserving pasture or cropping on farm. 
In the beef farms no supplement feeding was used in the present model. Beef farms are less likely 
to buy in supplements, but use their own hay and silage. Sheep were less likely to interact with 
simulated tuberculous possums (Sauter and Morris, 1995), therefore in the model sheep were used 
in hot-spot areas during the summer and winter months of highest risk. 
Another factor which was not included in the present study, but could be included in further 
studies, was pasture damage by cattle. Often farmers put their cattle during the winter months into 
rough areas with less productive pasture, and often incidentally plenty of potential possum habitat, 
in order to save the pasture quality of better paddocks. The present study did not take any potential 
damage to pasture into account. 
Although the model devised on each of the participating farms may not have simulated the real 
situation perfectly, in comparing a standard grazing plan with an alternative plan, the assumptions 
were equal for both plans. By changing any of the underlying principles, such as pasture growth 
rates or potential animal intake, it is likely that the comparison between the two grazing plans still 
holds. Although the fmal economic output has only a guidance function, the comparison between 
305 
the plans stills shows, which plan is better, and whether excluding hot-spots from grazing can 
compete from an economic point of view. By using slope and aspect of each individual paddock 
pasture growth can be predicted more accurately. Vegetation maps can be overlaid on the paddock 
maps, giving a true percentage of effective land in each paddock. Using vegetation hot-spot maps 
can also help in identifying likely hot-spot areas (McKenzie, 1999). 
Potential further development of FarmORACLE 
In order to validate FarmORACLE it would be necessary to observe the grazing regimes, actual 
pasture growth, animal production and economic outcomes in more detail on dairy and beef farms . 
Then the outcome of the model could be compared with the actual situation on farm, and necessary 
changes made to the program. However, the data required for validation has to be detailed and 
exact and it will be difficult to get precise data in pasture based enterprises, as sward conditions and 
characteristics vary considerably between paddocks and over short periods of time. 
Further development of FarmORACLE could include adding parameters that would allow to model 
the variations in pasture digestibility, between seasons as well as between paddocks (Wilson et al. , 
1995). Additionally, FarmORACLE could be linked to EpiMAN(TB) (McKenzie, 1999), to 
automatically transfer to FarmORACLE information about TB hot-spots, using vegetation and 
slope data obtained from satellite images. EpiMANCTB) uses these two measurements to classify 
the risk of TB hot-spots into low, medium, and high risk. FarmORACLE could also be further 
developed to use satellite images to calculate the true effective size of each individual paddock. 
Control programmes for other diseases in grazing animals could also be modelled in 
FarmORACLE. By creating 'base' scenarios with standard prices etc., but without consideration of 
disease control issues, the model could then be used to assess different disease control programmes 
on equal terms. 
Conclusions 
FarmORACLE has five main benefits: 
• It can be farm specific, with specific livestock and paddock details, rather than simulating 
an 'average' farm. 
• It models all livestock enterprises present on farm, and with it the interactions between 
different stock groups and pasture availability. 
• It can be used to evaluate different grazing regimes to accommodate disease prevention 
measures or to evaluate different basic grazing principles used on the farm, such as 
306 
rotational grazing of paddocks in list order versus rotational grazing on the basis of pasture 
cover. 
• It can reduce the unpredictability of management decisions relating to grazing strategies. 
• It uses farm maps and graphical display of paddock and animal allocations, as well as text 
flles for model outputs. 
FarmORACLE has shown that grazing regimes can be found for dairy and beef farms, that exclude 
cattle from grazing TB hot-spot areas during high risk times, and still result in an equal or even 
better economic outcome than the traditional grazing regime used. 
Bibliography 
Anderson, A. M. ( 1982). Processes and implications of knowledge transmission in Australian 
Agricultural Extension. Hawkesbury Agricultural College, Richmond, NSW, Australia. 
Barlow N. (2000). Chapter Nineteen: Models for Possum Management. Montague T.L. (editor). 
The Brushtail Possum. Manaaki Whenua Press, Lincoln, New Zealand, pp. :208-219. 
B ircham, J.S.  and Hodgson, J. ( 1983). Dynamics of herbage growth and senescence in a mixed­
species temperate sward continuously grazed by sheep. Proceedings of the XIV International 
Grassland Congress 1983, 601-603. 
Bircham, J .S. and Sheath, G.W. ( 1986). Pasture utilisation in hill country. 2. A general model 
describing pasture mass and intake under sheep and cattle grazing. New Zealand Journal of 
Agricultural Research, 29: 639-648. 
Brookes, I.M., Parker, WJ. and Gray, DJ. (1992). The role of computer software in the feeding of 
farm livestock. Proceedings of the New Zealand Society of Animal Production, 52: 1 1 1- 1 14. 
Butler, B. M. ( 1986). The effect of grazing intensity and frequency during spring and early summer 
on the sward characteristics of a ryegrass-white clover pasture. Unpublished Master of 
Agricultural Science in Plant Science thesis, Massey University, Palmerston North, New 
Zealand. 
307 
Butler, B .M. and Hodgson, J. (1993). Effect of management on leaf growth and senescence of three 
contrasting grass species. Proceedings of the XVll International Grassland Congress 1993, I: 
147-149. 
Butler, B.M., Hoogendoorn, c.J. and Richardson, M.A. ( 1987). Pasture quality and animal 
performance over late spring and summer. Proceedings of the New Zealand Society of Animal 
Production, 47: 3 1 -33. 
Coutinho, H.B., Mathews, P.N.P. and Morris, S.T. ( 1998). The Effect of Grazing Management on 
Pasture and Animal Production in late Autumn to early Spring period in a one year Bull Beef 
Grazing System. Proceedings of the New Zealand Society of Animal Production, 58: 236-238. 
FarmPro Systems Limited (2000). User Guide - PigLITTER and PigMAIN Modules. FarmPro 
systems limited, Palmerston North, New Zealand. 
Fleming, P. E. ( 1996). Farm Technical Manual. Department of Farm and Horticultural 
Management, Lincoln University, Canterbury, New Zealand. 
Garcia-Muniz, J.G., Holmes, C.W., Garrick, D.J. et al. (1998). Growth curves and productivity of 
Holstein-Friesian cows bred for heavy or light mature live weight. Proceedings of the New 
Zealand Society of Animal Production, 58: 68-72. 
Gray, DJ., Lynch, G.A., Lockhart, J.c., Parker, WJ., Todd, E.G. and Brookes, LM. (1992). A 
conceptual framework for an expert system to improve drying off decisions on seasonal supply 
dairy farms. Proceedings of the New Zealand Society of Animal Production, 52: 1 1 - 12. 
Hodgson, J., Niezen, J.H., Montossi, F., Liu, F. and Butler, B.M. (1996). Comparative studies on 
pasture and animal performance and parasite infestation in sheep grazing Yorkshire fog, 
perennial ryegrass and tall fescue pastures. Proceedings of the New Zealand Grassland 
Association, 57: 89-93. 
Jackson, R. ( 1995). Transmission of tuberculosis (Mycobacterium bovis) by possums. unpublished 
PhD thesis, Massey University, Palmerston North, New Zealand. 
308 
Larcombe, M. ( 1990). UDDER: a desktop dairyfarm for extension and research. In: Proceedings of 
the 7th Seminar - Dairy Cattle Society of the New Zealand Veterinary Association, Hamilton, 
New Zealand, 22-25 May 1990. Dairy Cattle Society of the New Zealanq Veterinary 
Association, Palmerston North, New Zealand, pp. 151-2. 
Livestock Improvement Corporation Limited ( 1999). Dairy Statistics 1998- 1999. New Zealand 
Dairy Board, Hamilton, New Zealand. 
Livingstone, P.G. (1997). Update on the New Zealand TB situation. In: Proceedings of a Seminar 
on Possum and Mustelid Control Research. National Possum Control Agencies, Wellington, 
New Zealand, pp. 17-30. 
Lockhart, J., Stevenson, M.A. and Morris, R.S. ( 1997). DairyWIN Animal management program. 
Version 1 .83. Massey University, Palmerston North, New Zealand. 
Lodge, G.M. and Frecker, T.C ( 1990). FEEDBAL: an integrated expert system for calculating 
whole-farm forage budgets. Computers and Electronics in Agriculture, 5 (2): 101- 1 17. 
Lynch, T., Gregor, S. and Midmore, D. (2000). Intelligent support systems in agriculture: how can 
we do better? Australian Journal of Experimental Agriculture, 40 (4): 609-620. 
Marshall, P.R., McCall, D.G. and Johns, K.L. ( 1991). Stockpol: a decision support model for 
livestock farms. Proceedings of the New Zealand Grassland Association, 53: 137-140. 
Matthew, C ,  Black, CK. and Butler, B.M. ( 1993). Tiller dynamics of perennation in three herbage 
grasses. Proceedings of the XVII International Grassland Congress 1 993, I: 141- 143. 
Matthews, P.N.P. (1994). Grazing Management strategies for Dairy production. Dairyfarrning 
Annual, 46: 143- 1 5 1 .  
Matthews, P.N.P. (1995). Grazing Management Principles and Targets: A Case Study. 
Dairyfarrning Annual, 47: 17 1-174. 
Matthews, P.N.P. (1997). Feeding Practices on commercial farms: Recent changes in Attitudes, 
their reasons and effects. Dairyfarming Annual, 49: 67-7 1 .  
309 
McKenzie, J. S .  ( 1999). The use of habitat analysis in the control of wildlife tuberculosis in New 
Zealand. Unpublished PhD thesis, Massey University, Palmerston North, New Zealand. 
McPhee, M.J., Bell, A.K., Graham, P., Griffith, G.R. and Meaker, G.P. (2000). PRO Plus: a whole­
farm fodder budgeting decision support system. Australian Journal of Experimental 
Agriculture, 40 (4): 621-630. 
Meat & Wool Economic Service of New Zealand (1999). Annual Review of the New Zealand 
Sheep and Beef Industry 1998-99. Meat & Wool Economic Service of New Zealand, 
Wellington, New Zealand. 
Meek, A.H. and Morris, R.S. ( 1981) .  A computer simulation model of ovine fascioliasis. 
Agricultural Systems, 7: 49-77. 
Ministry of Agriculture (1999). Farm Monitoring Report - North Central Region. MAP Policy, 
Ministry of Agriculture, Palmerston North, New Zealand. 
Morris, C. ,  Loveridge, A. and Fairweather, J.R. (1995). Understanding why farmers change their 
farming practices: the role of orienting principles in technology transfer. Research Report 
No.232. Agribusiness and Economics Research Unit, Lincoln University, Canterbury, New 
Zealand. 
Morris, R.S. and Pfeiffer, D.U. ( 1995). Directions and issues in bovine tuberculosis epidemiology 
and control in New Zealand. New Zealand Veterinary Journal, 43 (7) :  256-265. 
Pannell, DJ. ( 1996). Lessons from a decade of whole-farm modelling in Western Australia. 
Review of Agricultural Economics, 1 8  (3): 373-383. 
Parker, WJ., Dooley, A.E. and Garrick, D.J. ( 1998). Sheep breeding: An enterprise budgeting 
decision support model for on-farm planning. Proceedings of the New Zealand Society of 
Animal Production, 58: 165- 169. 
Parker, WJ., Gray, DJ., Lockhard, J.C. and Townsley, RJ. ( 1994). Farm management research in 
New Zealand and its contribution to animal production. Proceedings of the New Zealand 
Society of Animal Production, 54: 357-362. 
3 1O 
Parker, WJ., Gray, DJ., Morris, S.T. and McCutcheon, S.N. ( 1992). A modelling study of the 
productivity and profitability of unmated and mated ewes on North Island hill country. 
Proceedings of the New Zealand Society of Animal Production, 52: 221 -223. 
Sauter, C.M. and Morris, R.S. ( 1995). Behavioural studies on the potential for direct transmission 
of tuberculosis to farmed livestock from feral ferrets (Mustelafuro) and possums (Trichosurus 
vulpecula). New Zealand Veterinary Journal, 43 (7): 294-300. 
Texas A & M University (1997). The Grazing Manager (CD-ROM). College Station, TX 77845-
2126, 409-845-5580, rhinnant@tamu.edu 
Tbomson, N.A., Lagan, J.G. and McCallum, D.A. ( 1984). Herbage allowance, pasture quality and 
milkfat production as affected by stocking rate and conservation policy. Proceedings of the 
New Zealand Society of Animal Production, 44: 67-70. 
Uribe, J.V., Parker, W.1., Dake, c.K.G. and McDonald, A. ( 1996). A whole farm approach to feed 
planning and ration balancing using UDDER and CAMDAIRY. Proceedings of the New 
Zealand Society of Animal Production, 56: 285-288. 
Vaillancourt, J.P., Marsh, W.E. and Dial, G.D. ( 1992). Internal consistency of preweaning 
mortality data collected by swine producers. Preventive Veterinary Medicine, 14 ( 1 -2) : 1 15-
128. 
Wilson, G.F., Moller, S. ,  Parker, WJ. and Hodgson, J.  (1995). Seasonal differences in pasture 
composition and nutritional implications. Dairyfarrning Annual, 47: 46-56. 
3 1 1  
Dairy farm in the Wairarapa 
Beef farm in the Wairarapa 
3 12 
3 1 3  
Chapter 8 
General Discussion 
3 14 
3 1 5  
Introduction 
This large and complex study has yielded some simple, and readily applicable concepts for 
adoption by the farming industry, which will assist the reduction of TB prevalence. 
Bovine tuberculosis is a global disease. However, the eradication of tuberculosis is complicated in 
some countries by the existence of wildlife species which act as reservoirs of infection and vectors 
which transmit TB to domestic stock. A notable example is the brushtail possum in New Zealand 
(Morris et aI. , 1 994) (see Introduction to this thesis for more examples). In 1998 about 23% of New 
Zealand' s  land area was classified as Vector Risk Areas, where wildlife infection is the primary 
source of livestock infection (Animal Health Board, 1998). The disease is a major threat to New 
Zealand' s  agricultural economy (Oliver et aI. , 2000). Currently 2% of New Zealand' s  cattle and 
deer herds are infected. This level has to be reduced to 0.2% in order to achieve internationally 
recognised TB-freedom. 
Control of TB in New Zealand consists of three major components: stock movement control; stock 
testing and slaughter; and vector control. Most possum control is done by poison-baiting, using 
aerial and ground application of poison baits (Morgan and Hickling, 2000). Expenditure on vector 
control was $28.4 million in 1998/99 (Coleman and Livingstone, 2000). However, despite possum 
control operations reducing TB in livestock (Pannett, 1995; Livingstone, 1997; Animal Health 
Board, 2000), a number of Vector Risk Areas have expanded (Oliver et a!. , 2000).  
While nationally and regionally operated control programmes have achieved significant reduction 
in herd incidence, there is an opportunity to complement these official control efforts with 
complementary individual farmer efforts (Oliver et al. , 2000). This study evaluated the potential of 
farmers helping the eradication of TB from their own, and consequently from New Zealand' s  cattle 
and deer herds. In order to advance such efforts, several questions have to be answered: Are there 
any control methods available to farmers? If so, are they practical and easily implemented on 
farms? Are such methods effective? Are they financially worthwhile? What are the attitudes of 
farmers towards control and towards individually contributing to TB control? What understanding 
of behaviour is important to achieve change? 
It can be proposed from the present study that the first three questions can be answered with 'yes' ; 
while the fourth question is dependent on herd type, number of TB animals on farms and 
legislation regarding TB. In answer to the fifth question there is an overall positive attitude 
amongst farmers towards TB control, but individual responsibility is still low; and in answer to the 
3 16 
sixth question a literature review showed that a change in attitude, thinking, knowledge and 
motivation is required to achieve change. 
Are there any control measures available for farmers? 
It was found that farmers generally had good knowledge about potential TB hot-spot areas on their 
farm, that farms which grazed cattle in hot-spot paddocks had higher TB incidence in cattle, and 
that grazing management was flexible to a greater degree on beef farms than on dairy farms (see 
Chapter 2). The findings of this part of the study together with epidemiological knowledge on TB 
and its transmission between vector species and livestock emanating from other research was used 
to develop potential on-farm control measures to reduce or prevent the direct contact between 
infectious vectors and livestock. This contact was the most likely transmission pathway for 
tuberculosis between the two species (Sauter and Morris, 1995; Paterson and Morris, 1995). These 
control measures consisted of targeted possum control in spring and autumn; and grazing 
management to avoid close vector-livestock contact during the high risk times of November to 
January and July and August. 
Thus there are appropriate measures that farmers can implement to assist TB control on their own 
farms. 
Are these on-farm methods practical and easily implemented? 
The analysis of the open interviews conducted with farmers in the Wairarapa who were managing 
TB infected herds (Chapter 2) showed that grazing management was flexible, especially on beef· 
farms where it often was adapted to short-tenn changes. 
The suggested control measures mainly involve time in servicing bait stations and conducting 
possum control, as well as planning appropriate grazing placement of animals on the farm. 
However, these measures are easy to implement and of relatively low cost, in comparison to the 
official Regional Council control programmes. 
In order to facilitate the adoption of grazing management that excludes cattle and deer from grazing 
TB hot-spot paddocks in summer and winter, the farm can be modelled using FarmORACLE. The 
programme can be tailored to each specific farm. It models all stock classes present on the specific 
farm, which is an advantage over using 'average' farms (Morris et al., 1 995). Thus, it can reduce 
the unpredictability of management decisions relating to grazing strategies, a major hindering 
3 17  
factor in the adoption of innovations (Morris et aI. , 1995). The graphical display of grazing routines 
and production outcomes facilitates the use of the programme. 
Thus the methods evaluated are practical and easily implemented. 
Are these on-farm control measures effective? 
Evaluation of these on-farm control measures provided evidence that they are effective in reducing 
TB in livestock (see Chapters 3 and 4). Both intervention studies achieved a higher reduction in 
cumulative TB incidence in focused control farms than in standard control farms. The Wairarapa 
project interventions achieved a higher proportion of focused control farms coming off Movement 
Control than standard control farms . 
The evaluation of the national project suggested that providing farmers with only advice on control 
methods and strategies, rather than actual control work as well, had some effect (see Chapter 4). 
The Wairarapa interventions achieved higher reductions in TB incidence than observed in the 
national intervention study. Thus it was concluded that the approach using a small team with 
hands-on practice, as applied in the interventions in the Wairarapa project, was more successful 
than using a larger team with an advisory role only. 
It is therefore proposed that the control measures evaluated are effective if adopted intensively, and 
do assist the control and eradication of TB at the farm level. 
Are these on-farm control measures financially worthwhile? 
Economic issues were found to be a major factor in deciding whether to conduct any on-farm TB 
control (Chapter 2). However, the answer to the question whether the on-farm control measures are 
financially beneficial, is dependent on herd type, the number of TB animals on farm, and 
legislation regarding TB. The economic analysis (Chapter 6) used both deterministic and stochastic 
modelling, plus decision analysis. The results showed that under the current compensation level for 
cattle of 65% the adoption of on-farm control measures generally resulted in a net gain for dairy 
farms. The control measures were economical for beef farms only if they reduced their TB animal 
numbers to zero and the herd came off MC. If the compensation level was reduced to zero, the 
situation was not altered significantly. In this case, the net gain for dairy farms increased, the 
situation in the beef breeding farms remained similar, and for beef finishing farms the adoption of 
3 1 8  
control programmes became economically beneficial only if it resulted in a reduction of two or 
more TB animals annually. 
Therefore, in contrast to dairy fanns, a reduction of compensation level does not create a 
significant incentive for beef farmers to adopt on-farm control measures for TB. In addition, 
farmers believed that the removal of compensation could result in less co-operation by farmers, and 
therefore lead to results contrary to the intended ones. Thus other incentives or stringent rules and 
regulations may have to be put in place to encourage or force farmers to conduct their own control 
programmes. It is suggested that a phased scheme is introduced that provides financial incentives 
fo fanner 0 adopt on-fann control measures. Financial incentives were found to be major factors 
in determining the adoption of innovations (Walker and Bell, 1994; Morris et aI., 1 995). In a 
second phase, compensation for TB reactor cattle can be reduced in order to create more financial 
incentive to adopt on-farm control measures. 
Thus, the current TB legislation for cattle farms does not create enough financial incentive for beef 
farms to adopt on-farm control methods, even if compensation for cattle reactors is reduced. 
Attitudes of farmers towards TB control 
Overall a positive attitude existed among farmers about the importance of TB eradication. 
However, the majority of farmers interviewed in this study were not in favour of stricter Movement 
Control regulations, removal of compensation or having to pay TB testing costs directly. Their 
main reason was the potential loss of co-operation by farmers. In a contemporary study by Corner 
et al. (2000) it was found that deer farmers were generally more in favour of reducing 
compensation and having to pay costs directly. It is noteworthy that both groups strongly favour the 
system currently in place for their industry, and oppose the system being used in the other industry. 
Although many farmers saw 'farmers in general' as being responsible for TB control, many 
farmers saw the responsibility of TB control and eradication belonged to official organisations, 
such as Regional Council, Government, and Animal Health Board. Reasons for this were manifold 
and not always based on fact. Farmers argued that TB is a national problem, not the individual 
farmer's; that farmers pay rates to the Regional Council and therefore already pay for the control; 
some farmers believed that control had to be over whole districts/regions to be effective, and that 
any control farmers themselves would do would be ineffective, sporadic and not cost-effective (see 
Chapters 2 and 5). 
3 19 
Therefore, farmers do not see the need yet for themselves to be involved in a major way in the 
control of TB. Thus, any future eradication strategy should include methods for improving farmer 
motivation in addition to an extension programme that addresses technical issues such as control 
strategies and financial implications of TB. 
What understanding of behaviour is important for achieving 
change? 
However, before on-farm control measures have reasonable prospect of being adopted, one has to 
answer other questions, such as: What change is necessary for farmers to adopt such on-farm 
control measures? What incentives have to be created for farmers to employ on-farm control 
measures? Which are the best ways to convey the information to farmers? 
Review of the literature about human behaviour indicates that the process of behaviour change 
consists of three stages: acceptance of need to change, awareness of options for change, and 
acquisition of the skills required for the implementation of the change (Greer and Greer, 1996) 
(Chapter 1) .  From the results of the questionnaires used in the present study (Chapter 5) it was 
indicated that many farmers did not see themselves as being responsible for TB eradication. They 
believed organisations such as Government, Regional Council, AgriQuality, AHB and others were. 
This observation indicates that many farmers have not accepted the need for change. Therefore 
more emphasis and extension work has to be put on explaining to farmers the whole TB situation, 
so that they have all the information on costs and benefits on their own farm, but also on the 
national scale. Alternatively the structure of the programme could be changed to encourage more 
individual farmer consequences. 
Another field of investigation should focus on how farmers can best be motivated and encouraged 
to apply these control measures, involving different extension forms and incentives/penalties, 
possibly dependent on herd enterprise type. From the economic analysis of the TB situation and the 
control measures (Chapter 6) it was found that the application of additional targeted control in beef 
breeding farms was less economical than in dairy herds. Therefore beef farmers need more 
financial incentives, such as rewards if they achieve Clear TB status, or free poison/equipment. 
Methods like this could be evaluated using intervention studies or even surveys, asking farmers if 
they were prepared to do the control if they receive certain incentives. In order to evaluate how 
many of the farmers then actually would implement the control, one could set up a study using 
randomly selected farmers from only that group that responded positive to the question. This would 
give an estimate of the proportion that could be expected to use these methods. The results of this 
320 
study would then give indications if penalties would have to be applied. One of the problems with 
promoting such schemes is to capture a commercial return on technology transfer, as it is easier to 
sell an individual product or a special service than a management concept (Joumeaux et aI. , 1 997). 
Opportunities for enhancement of on-farm programmes 
Any further study of the nature of the present one could potentially be improved by using regular 
farmer meetings/discussion groups, to obtain group adhesion and achieve active discussions 
amongst farmers (Wegener et aI., 2000). However, discussion groups may not appeal to all farmers 
(Morris et aI., 1995). In the present Wairarapa project (Chapter 3) one meeting of all farmers was 
held. Although there was a good relationship between the research team and the farmers, the 
farmers did not have much contact amongst themselves in this project. 
Additional focus could be put on the financial impact of TB on the farms by including an advisory 
person trained specifically in farm economics. In the present study it was noticeable that a lower 
proportion of national focused control farmers in the one-on-one programme (Chapter 4) than the 
Wairarapa focused control farmers (Chapter 3) or the standard control farmers in both studies, 
considered the cost of TB to be zero or unknown. These farmers had interactions within their team 
meetings with a financial person, most likely resulting in a more detailed and maybe more correct 
analysis of the financial impact of TB. 
Future research proposals emanating from this study 
Adoption of control methods studied 
In order to further support the findings in this study, it is recommended that similar studies are 
conducted in VRAs, focusing on beef breeding farms and/or non-Regional Council controlled 
areas. From the Wairarapa project (Chapter 3) it was suggested that especially the beef breeding 
farms could gain from additional targeted control above the control received of the Regional 
Council control. Therefore a study could be conducted that only involved beef breeding farms. As 
the farms in non-Regional Council controlled areas achieved a higher reduction in cumulative 
incidence in the Wairarapa project, it would be worthwhile to conduct a study like the one 
presented using only farms in non-RC areas. This could investigate if on-farm control measures 
like the ones applied in this study could achieve effectiveness similar to blanket control delivered 
by RC. However, a study like this would be very difficult, as more and more area within the VRAs 
321 
are under control. For example in  the Wairarapa 67% of farmed area was under control in 1 999 
(Animal Health Board, 1999). Furthermore, targeted control might become more and more 
important even within the Regional Council delivered control. As an example of adoption of result 
of the present study, a modelling study was conducted by a research colleague looking at the 
potential of applying targeted control rather than blanket control on a regional level, which 
indicated that a mix between different intensities of targeted control might be most cost-effective 
(McKenzie, 2000). 
FarmORACLE 
In order to facilitate the implementation of grazing strategies that exclude cattle and deer from 
certain areas during winter and summer time, computer programs such as FarmORACLE (Chapter 
7) have great potential. If farmers can be convinced that a changed grazing regime can result in 
equal or better economic outcomes, they will be more likely to adopt these changes. Therefore it is 
recommended that programs such as FarmORACLE are tested and evaluated on a range of farms 
with different locations and enterprise types. The use of programs like this is aided by the current 
trend that more and more farmers are using farm recording programs. Thus information on grazing 
routines is readily available. The ease and use of the program can be enhanced by linking the 
program to EpiMAN(TB) (McKenzie, 1 999), a program using satellite images to obtain data on 
vegetation and slope, in order to classify the risk of TB hot-spots. Using satellite images within 
FarmORACLE could also result in easier and more accurate calculation of the effective hectare 
size of each individual paddock. By linking FarmORACLE to other software, such as the economic 
models used in Chapter 6, would provide a basis for calculating costs and benefits of employing the 
control measures on each individual farm. 
Further, the use of FarmORACLE is not limited only to grazing regimes with respect to TB, but it 
could have wide-ranging applications in the control of other diseases that occur in patches, such as 
infection with fasciola hepatica (Radostits et aI. , 1994), or ryegrass staggers and diarrhoea due to 
endophyte-infected grass (Pownall et al. , 1993). 
Implications of the methods studied for TB control 
Findings from this study suggests that it is possible that farmers can and will participate in the 
control and eradication of TB. It is proposed that the best approach to achieve this, is using a 
phased scheme, including incentives and penalties. Farmers in New Zealand readily change in 
response to financial signals (Walker and Bell, 1994). Therefore, financial incentives should be 
322 
offered first, with a phased reduction of compensation for reactor animals and introduction of 
penalties for having TB animals or not applying control methods. 
In the future, in addition to targeted possum control and grazing management, vaccination of 
possums could also be part of on-farm control measures (Aldwell et aI. , 1995; Corner et aI. , in 
press). This and other newly researched practices will enhance the acceptability of farmer-operated 
vector control. 
C clu io 
Overall the results in the present study provide evidence that TB control can be assisted by farmers 
themselves by adopting appropriate vector control and farm management practices. However, the 
results also indicated areas where more research and emphasis is needed in relation to farmer 
adoption of these changes. Financial incentives are likely to be the principal drivers of change. 
These findings give direction to decision makers within the National TB control and eradication 
programme for cattle and deer in New Zealand. 
Bibliography 
Aldwell, F.E., Keen, D.L., Stent, V.C. et al. (1995). Route of BCG administration in possums 
affects protection against bovine tuberculosis. New Zealand Veterinary Journal, 43 (7): 356-
359. 
Animal Health Board ( 1998). Annual report for the year ending 30 June 1998. Animal Health 
Board, Wellington, New Zealand. 
Animal Health Board ( 1999). Regional TB Operational Plan, Wellington Region, 1 July 1999 to 30 
June 2004. Animal Health Board, Wellington, New Zealand. 
Animal Health Board (2000). Bovine Tuberculosis National Pest Management Strategy 2001-201 1 ;  
Towards a TB free NZ; A discussion paper on the future options. Animal Health Board, 
Wellington, New Zealand. 
323 
Coleman J., Livingstone P. (2000). Chapter Twenty: Fewer Possums: Less Bovine TB. Montague 
T.L. (editor). The Brushtail Possum. Manaaki Whenua Press, Lincoln, New Zealand, pp.:220-
23 1 .  
Corner, L.A, Buddle, B.M., Pfeiffer, D.U. and Morris, R.S. (in press). Aerosol vaccination of  the 
brushtail possum (Trichosurus vulpecula) with bacillus Calmette-Guerin: the duration of 
protection. Veterinary Microbiology. 
Corner, R. A, Morris, R.S. and Davies, P.R. (2000). Attitudes of New Zealand Farmers to Bovine 
Tuberculosis and the Implications for Control Strategies. Final Report to the Animal Health 
Board - Project R-20452B . EpiCentre, Massey University, Palmerston North, New Zealand. 
Greer, J. and Greer, G. (1996). Experiences of Agricultural Extension. MAP Policy Information 
Paper No. 12. 
Journeaux, P., Stephens, P. and Johnson R.W.M. (editor) (1997). The development of agriCUltural 
advisory services in New Zealand. MAP Policy Technical Paper 97/8. 
Livingstone, P.G. ( 1997). Update on the New Zealand TB situation. In: Proceedings of a Seminar 
on Possum and Mustelid Control Research. National Possum Control Agencies, Wellington, 
pp. 17-30. 
McKenzie, J.S (2000). AHB Project Report: Implementation of EpiMAN(TB). Incorporating hot 
spots into possum control programs. Massey University, Palmerston North. 
McKenzie, J. S .  ( 1999). The use of habitat analysis in the control of wildlife tuberculosis in New 
Zealand. Unpublished PhD thesis, Massey University, Palmerston North, New Zealand. 
Morgan D., Riclding G. (2000). Chapter Thirteen: Techniques Used for Poisoning Possums. 
Montague T.L. (editor). The Brushtail Possum. Manaaki Whenua Press, Lincoln, New 
Zealand, pp. : 143-153. 
Morris, c., Loveridge, A and Fairweather, J.R. ( 1995). Understanding why farmers change their 
farming practices: the role of orienting principles in technology transfer. Research Report 
No.232. Agribusiness and Economics Research Unit, Lincoln University, Canterbury, New 
Zealand. 
324 
Morris, RS.,  Pfeiffer, D.U. and Jackson, R ( 1994). The epidemiology of Mycobacterium bovis 
infections. Veterinary Microbiology, 40 ( 1 ): 1 53- 177. 
Oliver, 1., Alien, W., Kilvington, M. and Gilbert, M. (2000). The Role of Animal Health Board's 
Locally Initiated Programmes in Changing Farmer Attitudes. In: Stantiall ID, McDiarmid CM 
(eds.), Achieving Change Through Improved Knowledge Systems. Centre for Agricultural and 
Veterinary Continuing Education, Massey University, Palmerston North, New Zealand, pp. 
343-349. 
Pannett, O. ( 1995). Possum control in the Wellington region: how successful has it been. In: Griffin 
F, de Lisle G (Tuberculosis in Wildlife and Domestic Animals - Otago Conference Series No. 
3. University of Otago Press, Dunedin, pp. 294-296. 
Paterson, B.M. and Morris, RS . (1995). Interactions between beef cattle and simulated sedated 
possums on pasture. New Zealand Veterinary Journal, 43 (7): 289-293. 
Pownall, D.B., Lucas, R1., Familton, A.S., Love, B .O., Hines, S.E. and Fletcher, L.R ( 1993). The 
relationship between staggers and diarrhoea in lambs grazing different components of 
endophyte-infected ryegrass. Proceedings of the New Zealand Society of Animal Production, 
53 : 19-22. 
Radostits, O. M., Blood, D.C and Oay, CC. ( 1994). Veterinary Medicine. A textbook of the 
diseases of cattle, sheep, pigs, goats and horses. Balliere Tindall, London, UK. 
Sauter, CM. and Morris, R.S. (1995). Behavioural studies on the potential for direct transmission 
of tuberculosis to farmed livestock from feral ferrets (Mustelafuro) and possums (Trichosurus 
vulpecula). New Zealand Veterinary Journal, 43 (7): 294-300. 
Walker, A. and Bell, B. (1994). Aspects of New Zealand's Experience in Agricultural Reform since 
1984. MAF Policy Technical Paper 94/5, Wellington, New Zealand. 
Wegener, P., Frank, B. and Maroske, M. (2000). A Planning and Self-Evaluation process for dairy 
farmer discussion groups. In: Stantiall ID, McDiarmid CM (eds.), Achieving Change Through 
Improved Knowledge Systems. Centre for Agricultural and Veterinary Continuing Education, 
Massey University, Palmerston North, New Zealand, pp. 77-86. 
325 
Appendix 
326 
Appendix I :  Questionnaire used for the interviews (Chapter 2) 
The questions were used as guidelines and key words only. 
I FARM MANAGEMENT QUESTIONNAIRE I 
1. Farm Details: 
a) Name of owner: 
b) Address: 
c) Interview date: 
d) Property details 
Farm ID: 
Herd No: 
Type of operation: 
Labour units: 
Years farm owned: 
Farm structure: 
1. Home farm 
location: 
Size (effectiveltotal) 
Ownership: 
2. Run-off 1 
Location: 
Own/lease? 
Used Since? 
3. Run-off 2 
Location: 
Own/lease? 
Used Since? 
2. Stock numbers (30 June 1995) 
Cattle: 
Milking cows: 
Breeding cows: 
R 2yo heifers: 
R 2yo steers: 
R 2yo bulls: 
R l yo heifers: 
R lyo steers: 
R l yo bulls: 
=� ���J 
MASSEV UNIVERSITY 
327 
328 
Breeding bulls: 
Sheep: 
Ewes: 
Ewe hoggets: 
Wether hoggets: 
Rams: 
Deer: 
Hinds: 
Weaners: 
Stags: 
Other animals: 
3. TB history 
Test how often? 
TB history the way fanner told it: 
TB in one particular stock group? 
TB infected on homefannlrun-off? 
4. Grazing management 
Start AI: 
Start Mating: 
Start calving: 
Wean calves: (how heavy are calves in kg / or how old are calves in months) 
Dry cows off: 
Grazing management for Calves 
Grazing management for Yearlings 
Grazing management for Cows 
Grazing management for Heifers 
Grazing management for Steerslbulls 
5. Major changes to grazing management? 
Yes/no, what changes? 
6. Influence of change on TB. 
Yes/no, what way? 
7. High risk areas? 
Are there any high risk areas? 
What animals are grazed there?/when? 
8. High risk areas and grazing management 
does fanner take the high risk into account? 
9. Grazing bush 
Yes/no, what sort? When grazed? What animals graze it? 
10. Pasture shortage 
When are times of pasture shortage? 
1 1. Farmer possum control 
any possum control done by farmer or farmer employed people? 
12. Herd replacement/purchasing 
13. TB and purchasing 
anything bought that had TB/white tags? 
14. Off-farm grazing 
yes/no, when and what animal groups? 
15. Grazing other owner's cattle 
Yes/no, when, what sort of animals, from where? 
16. Stock selling policies 
1 7. Classes of stock 
were there any changes in the classes of stock over the years? Due to TB? 
18. Feral animals 
any feral animals on farm? Any found with TB? 
19. Views on TB 
20. Cost of TB 
21. Methods of possum control 
22. Sources of information on TB 
23. Other comments on TB 
329 
330 
331  
Appendix 1 1 :  List of categories and codes used in  WinMAX98 
List of categories and codes used in the interview analysis of TB history and grazing management. 
The list of codes was taken out of WinMAX98, which stores the codes in alphabetical order. 
Farm Management 
Changes 
Reasons 
Economic 
General Comments 
General Management 
Mating/Calving 
Herd Replacement 
Off-farm Grazing 
Reasons 
Times 
Other land 
Other owner's cattle 
Purchasing 
TB bought 
View on affecting TB 
Selling 
By TB affected 
TB consequences 
Loss of ftexibility 
Reasons for accepting TB 
Grazing 
Changes 
Effect on TB 
Flexibility 
General 
General motives 
Grazing Bush 
Reasons 
Grazing hot-spots 
Mobs 
332 
TB 
Bulls 
Cows 
TB reactors 
Younger cattle 
Pasture shortage 
Management 
Supplements 
Control 
Grazing 
Motivation 
Possum Control 
Effect 
On TB 
On wildlife 
Farmer Control 
Convenience 
Awareness 
RC Control 
Views on methods for possum control 
Cost 
Control 
Non-control related 
Farmer's perception 
Regarding farm management 
Regarding MAF/RC control 
Farmer's theories 
TB history 
Reactors 
Circumstances 
Grazing/management (reactors) 
Subgroups 
Time 
Hot-spots 
Areas 
Grazing 
TB cattle 
Vegetation 
Wildlife 
Interesting quotes 
Sources of information 
Suggestions 
TB testing 
Farmer's view 
Wildlife 
General wildlife 
TB infected wildlife 
333 
334 
Appendix I l l :  Regu lations regarding TB control and testing 
The following definitions and rules are taken from the 'Five year national bovine tuberculosis pest 
management strategy' 1998 (Animal Health Board, 1998). 
TB classification 
Classification of Areas: For TB control reasons New Zealand is classified into 'TB Vector Risk 
Areas' and 'TB Vector Free Areas' .  Vector Risk areas (VRAs) are where M. bovis infection is 
present in wild animals, and where there is epidemiological evidence of clustering of infected herds 
and/or the persistence of TB in herds despite an intensive test and slaughter program. 
Herds are classified into 'Clear' , 'Suspended' and 'Infected' as a measure of the risk of TB 
infection: A herd changes from Clear to Suspended if a TB test positive animal or lesions in 
slaughter animals are found. All test positive animals have to be slaughtered unless an ancillary test 
is conducted. This herd changes back to Clear if the slaughter or ancillary test provide no indication 
of tuberculosis. The status changes to Infected (I) if TB has been confirmed by testing, post­
mortem or other approved diagnosis. A newly infected herd will be classified as 1 1 ,  with the 
number incremented for every year it remains infected. Herds with no evidence of TB are classified 
as Clear (C). Again, the number will be incremented for each year the herd has been confirmed by 
testing to be remaining clear of infection. For a herd to shift from Infected to Clear 1 (C l )  all test 
positive animals have to be slaughtered and the herd must have two consecutive clear whole herd 
tests with a minimum of six months between tests, with no further evidence of disease. Under 
normal circumstances it will take at least one year for a herd to change from Infected to Clear 
status. 
Beef dry stock herds have the option of obtaining the status 'Works Monitored' ,  depending on the 
area they are in and the percentage of cattle going to slaughter within 12  months. These herds 
require no on-farm testing. 
TB testing 
The standard test for TB in cattle is the Caudal Fold Skin Test, whereby 0. 1 ml of bovine tuberculin 
is injected intradermally and the test is read 72 (±6) hours after the injection. A positive test under 
standard interpretation is any palpable/visible reaction at the site of the injection. Any animal 
reacting positive to this test is termed 'reactor' . Similar testing procedure and interpretation applies 
to deer, but at the mid-cervical site. 
335 
As an ancillary test in cattle the Comparative Cervical Test (CCT) may be used, injecting bovine 
and avian tuberculin intradermally. The test is read 72 (±6) hours after the injection and under 
standard interpretation a positive test is present if any reaction at the site of the bovine injection is 
greater than any reaction at the site of the avian tuberculin. 
Another ancillary test in cattle is the Gamma Interferon Test (marketed as BOVIGAM), which is 
approved as an ancillary serial test in cattle, to be used only on Caudal Fold test positive cattle. The 
test has to be performed 13-30 days after the injection of tuberculin into the caudal fold. A positive 
test is obtained if the bovine antigen minus the nil antigen is 2: 100, AND the bovine antigen minus 
avian antigen is 2: 100. Ancillary tests in deer include BTB and post-skin test ELISA (for more 
details on these tests see (Griffin et al. , 1993). 
In cattle herds with Clear or Works Monitored TB status ancillary serial testing is used on skin test 
positive cattle using BOVIGAM only (if less than four skin test positive animals to be re-tested) or 
using a combination of BOVIGAM and CCT or CCT alone (if four or more skin test positive 
animals are to be re-tested). In this case, the CCT is read under the modified interpretation, where a 
difference in skin thickness of at least 4 mm exists between the two sites of injection (bovine and 
avian). 
The application of ancillary serial testing in cattle herds with an Infected or Suspended TB status 
depends on the TB area classification and the control/eradication objectives in this area. In Vector 
Free Areas and VRAs where the objective is  eradication of TB, ancillary tests are normally not 
applied on herds with Infected/Suspended status. Exemptions are made when specific 
epidemiological information are evident, such as evidence of skin TB, or lohnes vaccinated cattle, 
unexpected large number of test positive animals considering the TB history, or test positive 
animals in a management or age group, that is normally not associated with infection. 
In VRAs where the objective is control only or where there is no vector control program, ancillary 
tests can be used in areas where the skin test positive rate is very low or within herd exemptions 
apply. The standard interpretation of the CCT is used under these conditions. 
The TB status of a herd is monitored through on-farm TB testing and/or post-mortem inspection of 
all cattle processed in a slaughterhouse. TB testing in cattle herds is paid through the Animal 
Health Board (through levy funded by farmers per head of cattle slaughtered) and time frames are 
set for regular testing, depending on herd status and area classification. Deer farmers pay directly 
for TB testing costs. 
In cattle herds with a Clear2 status or better in VF As, all animals over 24 months of age have to be 
tested at least every 3 years. In herds with a status of Clearl all animals over three months of age 
336 
have to be tested at least every 12 months. Beef dry stock herds have the possibility to obtain 
'Works Monitored' status, whereby no on-farm testing is required. This is provided if 45% of 
animals in a triennial testing area, or 60% of animals in a biennial testing area go to slaughter over 
a 12 months period. 
In VRAs all animals over the age of 3 months have to be tested at least once every 6, 1 2  or 24 
months, which is defined in regional TB plans. All herds with a status of C l  have to be tested 
annUally. For Beef dry stock herds works surveillance can be applied when over 90% of the 
animals go to slaughter within 12 months. Test eligibility in deer herds varies slightly from that of 
cattle. 
Compensation 
All test positive cattle have to be slaughtered as soon as possible, within 30 days after diagnosis. 
For all reactor cattle slaughtered, the owner gets paid 65% of 'Fair Market Value' of the animal. 
This value is assessed at the time of testing and considers the value the animal had at that time, 
assuming it was TB free. The maximum value payable is determined by the New Zealand Dairy 
Board (for dairy cattle) and Meat New Zealand (for beef cattle). The money for compensation and 
TB testing is partly obtained through a levy per head of cattle being slaughtered, paid by all 
farmers. 
Deer farmers do not pay a levy per slaughtered animal, and no compensation is paid for deer 
reactors. 
TB Movement Control Restrictions 
Cattle and deer movements from herds with an 'Infected' status are restricted in order to prevent 
the spread of disease. Movement restrictions are lifted once the herd achieves a 'Clear' status again. 
In infected deer herds all animals are only allowed to move to slaughter directly. Owners of 
infected cattle herds have to obtain a Permit-to-Move, for all cattle leaving their farm except for 
cattle that go directly to slaughter. This permit is issued if all cattle six weeks or older are tested 
negative within 60 days of the movement date and all are identified with an official (AHB) white 
Movement Control ear tag ( 'white-tagged' cattle). The purpose of the white ear tags is to indicate a 
potential risk. Often this 'white-tagged' animals can only be sold at a discount, as the market for 
these animals is limited, since farmers in TB non-endemic areas are reluctant to buy these animals.  
If a TB reactor is diagnosed at a pre-movement test a permit is only issued if the lesion reactor rate 
at the pre-movement test is less than 0. 1 % or if the animals go to an Infected herd within a VRA. In 
both cases all the animals have to be white-tagged. All reactors have to be identified with official 
orange TB Reactor ear tags and can only be moved to slaughter directly. 
337 
An Infected herd is classified as 'Infected (High Risk)' if the TB lesion incidence in any test is 5% 
or greater if  the herd size is at least 50 cattle, or if  five or more lesioned animals are found in  a 
single test where the herd is less than 50 animals. For Infected (High Risk) herds the same rules 
apply, except that all animals to be moved and all animals in contact with them within the previous 
60 days have to be tested negative. Additionally if TB reactors are identified in a pre-movement 
test, the remainder of the tested animals has to stay on the property or move only to slaughter. 
Suspended herds have to fulfil the same requirements as Infected herds if the cattle are to be moved 
before the final result on the TB reactor is obtained and if there is a 'reasonable probability 
(�30%)' that the suspected animal is tuberculous. Otherwise suspended herds obtain a movement 
permit without the requirement of a pre-test. Movements directly to slaughter never require a 
permit. 
For all white-tagged cattle the movement has to be traced and they have to be tested again within 
60 to 120 days after leaving their herd of origin, but with at least 90 days between the pre- and 
post-movement test. 
As from January 1 SI 2000 a new policy is in place, whereby cattle can only move to other herds, if 
the herd already had one clear test and the group to be moved passed a clear test within 60 days of 
movement. Animals from infected herds without a clear test will have to pass a skin test and a 
parallel blood test. If these conditions are met, then the animals can move subject to a permit, with 
white ear tags, and they must be tested a minimum of 90 days after pre-movement test. In addition 
the purchaser's herd status will be suspended and only restored to Cl after a clear whole herd test a 
minimum of six months after the post-movement test (Animal Health Board, 1999). 
Bibliography 
Animal Health Board ( 1998). Five Year National Bovine Tuberculosis Pest Management Strategy 
to 30 June 2001 .  Animal Health Board, Wellington, New Zealand. 
Animal Health Board ( 1999). Movement from infected cattle herds. Documentation provided by 
the Animal Health Board, Wellington, New Zealand 
Griffin, IF.T., Chinn, D.N., Mackintosh, e.G., Rodgers, C.R. and Buchan, G.S. ( 1993). Serial 
ELISA and Skin Testing in Deer for Diagnosis of Mycobacterium bovis Infection. 
Proceedings of a Deer Course for Veterinarians. Deer Branch, New Zealand Veterinary 
Association, lO: 288-296. 
338 
Appendix IV: Questionnaire used on Wairarapa and national 
project farms 
TB Control Programme Questionnaire 
339 
Please fill in the blank spaces and tick or circle the appropriate box 
Example: when meaning YES 
Ye@ NoD o or Yes d NoD Massey 
University Name: 
Date: __ 1 __ 1 __ 
1) Introduction andiGene�ij\J))formation 
.��""��<-,:::};?�:"£L;:..�_���_ "._ ' 
1 .  What is the size of your farm? Please fill in the size of your farm (both total size and 
effective size) and other land used by your farm. (where stock is shifted by truck, land is 
classified as 'Other' , please indicate the distance in km from your home property for these 
areas) 
Total size Effective Size Distance from 
(in ha) (in ha) home property (in 
km) 
Owned Land Main property --------
Other owned land 
Leased Land Locally leased --------
Other leased land 
2. How many labour units are working on the farm? Full-time workers ____ _ 
Part-time workers ____ _ 
3.  Which of the following categories describe your enterprise type (multiple answers are possible, 
please give approximate percentage of total farm income) 
Enterprise type 
Dairy herd 
Town supply 
Seasonal 
Others (e.g. goats, pigs, plantations) 
Please 
% of gross farm 
income 
340 
4. How many animals do you run on your fann? (Please give the numbers as at 30th June 1998 
for each of the groups - If numbers have changed over the last 3 years, what were they at 30th 
June 1 995?) 
Cattle Female 
Male 
Deer female 
male 
Sheep Ewes 
Hoggets 
Rams 
others 
5. 
6. How important is off-fann grazing to your fann? 
Crucial 0 Important 0 Moderately ID Minor 0 Unimportant ID 
7. How important is trading of cattle and deer (e.g. through private sales, sale yards, but not to 
slaughter) to your fann? 
Crucial 0 Important 0 Moderately 0 Minor 0 Unimportant ID 
8. Numbers of stock (only cattle/deer) leaving the fann in the last 12 months. If you are currently 
on MC, please fill in also the last column, indicating the number of animals you would sell if 
your herd were not infected. 
?N.U¥fJt&.:if7f1fi.@i�1:.�f! �Yjy{[¥ifiiijJiiiJJjjp�,:f�:;�:I ,:,:� iir..�?/jfW#':liqi lijje.i}ji#.8 .. �-
leaving for Deer Cattle Deer Catte 
slaughter 
private sale / sale yards 
9. How do you replace your herd? Please tick the appropriate box and indicate the percentage 
each type contributes to the total replacement of your herd. 
Cattle Deer 
Own breeding % % 
Buying in % % 
Other (please specify) % % 
341 
10. How often do you purchase cattle/deer? 
Every year 0 every now and again 0 neverD Other (please specify) 0 __ _ 
1 1 . From how many different herds do you buy cattle/deer in a normal year? 
1-3 herds 0 more than 3 herds 0 from sale yards or similar source 0 
12. Number of animals (cattle/deer), which entered the farm in a normal year. Please indicate the 
time of the year when the animals were brought on to the farm, the class of stock and the 
number of animals: 
Bought animals 
Grazers 
13 .  Did you buy any white-tagged cattle? Yes 0 No 0 
If yes, what were your main reasons to buy white-tagged animals? _______ _ 
14. Which of the following factors do you consider as having contributed to your current or past 
TB infection? (Please tick the box, if you think it contributed to your TB problem) 
Contributed? 
Bought-in cattle/deer 0 
Cattle/deer grazed on-farm 0 
Cattle/deer grazed off-farm 0 
Problems with the TB-skin test D 
Neighbours (incl. DOC and RC land) D 
TB infected feral animals 0 
Possums 0 Ferrets 0 
Others (specify please) • 
15. Do you have any crops on your farm YesO NoD 
If 'Yes' : what sort of crop _____________________ _ 
342 
I. On-farm Stock Grazing Management 
16. Are there areas of your farm, which you know or believe, pose a high TB risk? 
Yes 0 No D Unknown D 
If 'Yes' do you graze your cattle/deer in these areas? Yes 0 No 0 
If 'Yes' , what time of the year do you graze your cattle/deer there? ____ _ 
What factor do you take into account when you graze these areas (open question)? 
17. Do you graze any cattle/deer in a bush area? Yes 0 No D 
What sort of bush? Heavy dense bush 0 
Other (please specify) 0 
Open bush 0 
18 .  Do you graze any cattle/deer near a bush area? Yes 0 No 0 
19. What is your average stock density per ha? 
I Overall Cattle Deer Sheep Stock densitylha 
20.As a management system, do you 
o move animals regularly between groups to meet grazing needs 
o keep animals together in the same groups as far as possible 
2 1 .  Do you identify your cattle/deer mobs permanently? (ear tags, by breed) Yes 0 No 0 
22. Does each mob graze a specific part of the farm? Yes 0 No 0 
23. During which months of the year do you normally experience your most severe shortage of 
pasture? (Please circle appropriate months) 
Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec 
24. How often does it happen that your cattle/deer get mixed up with those of your 
neighbours? 
Never 0 1-5 times a year 0 more than 5 times a year 0 
25. Within the last five years, have you had any neighbours that were infected with TB? 
Yes D No D Unknown D If yes: what year(s)? _____ _ 
11. Grazed in Stock and Stock sent off farm grazing: 
26. Are cattle/deer belonging to another owner grazed on your farm? Yes D No 0 
27. From how many farms do you graze stock? ________ _ 
343 
28. Within the last 3 years, did you send any stock for off-farm grazing? Yes 0 No 0 
29. What stock classes were sent for off-fann grazing? ____________ _ 
In. Weather influence 
30. What changes would/did occur due to drought (with regards to grazing/stock/general fann 
management)? ________________________ _ 
I. TB in your herd 
3 1 .  How often do your animals get tested for TB? If you are currently off MC please indicate in the 
last column how often your animals were tested while the herd was infected. 
,Cattle 
Currently While herd 
Twice a year 
Once a year 
Once every two years 
Other (please specify): 
32. How would you classify your TB problem? 
Deer 
Currently , 
Continuous 0 on and off 0 seldom 0 no problem any more 0 never 0 
How likely do you think it is that you get TB back on your farm? 
Less than 25% chance D Between 25% and 50% 0 
Between 50% and 75% D Over 75% D 100% 0 
33. Where did your TB-reactors get infected? Homefann D current Run-off D 
34. Has one class of stock been more affected than others with regards to TB reactors? 
Yes D No D Unknown D 
other 0 
If 'Yes' : which class/classes? ___________________ _ 
35. Has one breed of cattle/deer been more affected than others with regards to TB reactors? 
Yes D No D Unknown D 
If yes? What breed? ______________ _ 
344 
36. Have you changed your cattle/deer selling policies as a result of having TB on your farm? 
YesD NoD 
37. Have you changed the mix of classes of cattle/deer on your farm as a result of having TB? 
YesD NoD 
38. Over the last five years, did the number of TB reactors on your farm 
Increase 0 Decrease 0 Stayed about the same D 
If there was a change, to which factors do you contribute this change? ___ _ 
D. Cost of TB 
39. How much do you spend annually on TB control (e.g. possum control)? 
On poison: $ _____ _ 
On labour: $ ____ _ 
40. What do you expect to be the annual cost of TB to your farm, if you have TB on your farm or if 
you had TB back on your farm? (e.g. lost income) $ ___ _ 
Due to what reasons? ___________________ _ 
ITI. On-farm Feral Animal Control 
41 .  Are the following wildlife species present on your farm (please tick the appropriate boxes) and 
how likely do you think is contact between your stock and the wildlife species? 
. 
, : �... :-
� • • #. 
' . " ..... ' :-. 
possums 
ferrets 
wild cats 
wild pigs 
wild deer 
wild cattle 
hedgehogs 
rabbits 
Others 
(please 
specify) 
--. - . - - , -�'--" " .� -.- . - ..  - . -" �PJ!c,.i��Jfre$�l :, ' .. '� �()n�ac� . ; 
-;' . .  'on-jarm .' -: . . ,imli/iidy . 
0 0 
0 D 
0 D 
0 D 
0 0 
D 0 
D D 
D D 
D D 
" Coiitati .��-� 
. :possi.bi� � : 
0 
0 
0 
D 
D 
D 
D 
D 
D 
, Contact 
.. . �. . 
'" ve.T'Y"likilY· . 
D 
0 
D 
D 
0 
0 
D 
D 
0 
- . 
. U1fknpwn. 
-- -
D 
D 
D 
0 
D 
D 
D 
D 
D 
'" 
42. Has feral animal control been undertaken on your fann at any time within the last 5 years 
Yes 0 No 0 Unknown 0 
43. If 'Yes ' ,  when? Please tick all years when possum control was conducted. 
1993 D 19940 1995 0 1996 0 1997 0 1998 0 1999 0 
44. If 'Yes' who performed this work (may nominate more than one) 
Unknown 
Regional Council / MAF 
Fanner group (please specify who did the work) 
You, your family, fann worker 
Person from outside, employed for the work 
Others (please specify) 
o 
B 
B 
o 
345 
45. What type of control was performed over this period. Please distinguish between work done by 
the Regional Council (first column) and work done by yourself or persons you employed 
(second colUlnn). 
Shooting 
Trapping 
Poison 
If you know which 
Poison, please tick 
appropriate box 
Other (please specify) 
Phosphorus 
1 080 
Talon 
Cyanide 
"iJ()�trl;l!Ipne 
·· b . .  ;RC 
o 
o 
o 
o 
o 
o 
o 
o 
46. Extent of the TB control work conducted over the last 12 months. How many days per year 
would you (first column) or the Regional Council (second column) have done each form of 
control (trapping, poisoning, shooting . . .  ) (please specify in days per year) Please tick if the 
work is being done on a regular basis, meaning at least once a months) 
Trapping 
Poisoning 
Shooting 
Others (specify) 
How many days would you have worked on possum control three years ago? ___ _ 
47. What feral species was the control aimed at? 
Possums 
Ferrets 
o 
o 
Cats 
Rabbits 8 Others (please specify) 0 
346 
48. What proportion of the farm is subject to wildlife control (%)? 
Percentage of your total area controlled by Regional Council 
Percentage of your total area controlled by yourself 
----_% 
----_% 
49. Intensity of wildlife control over the last 5 years, distinguished between what the Regional 
Council did and what was done by the fann itself. Has the intensity of wildlife control on your 
farm 
RC initiated: 
Farm initiated: 
increased 0 
increased 0 
decreased 0 
decreased 0 
stayed stable 0 
stayed stable 0 
50. Which fonns of possum control do you consider best in reducing TB in your cattl�? 
Poison 0 Trapping 0 Shooting 0 
5 1 .  Which of the two types of programmes do you consider better in reducing TB in your ca tle? 
Regional progranunes 0 Individual Farm Efforts 0 
52. Where do you think individual farm efforts could play an important role? (open question) 
53. How important is it to you to eradicate TB from your herd 
crucial 
o 
important 
o 
moderately 
D 
minor 
o 
not important at all 
o 
54. Do you believe TB can be eradicated on your farm? Yes 0 NoD UnknownD 
55. If 'no' what are the factors hindering the progress? (open question) _________ _ 
56. Who do you see as having responsibility for eradicating TB from infected farms? (tick them 
please and rank them in tenns of your priority, with ' 1 '  being the highest priority.) 
All farmers 
Only farmers with the problem 
All landowners 
Government 
MAF/AHB 
Regional Council 
RAHC 
Local farming action group 
Local veterinarian 
Others (please specify) 
):R�'� ;"hlibii?'E� ·:<;:��'�-JHilJ..:-;.·�;;" -
o 0 
o 0 
o 0 
o 0 
o 0 
o 0 
o 0 
o 0 
o 0 
o 0 
347 
57. Who do you see as having responsibility for doing the work required for eradicating TB from 
infected farms? 
Only farmers with the problem 
all landowners 
Government 
MAF/AHB 
Regional Council 
RARC 
Local farming action group 
Local veterinarian 
Others (please specify) 
o 
o 
o 
o 
o 
o 
o 
o 
o 
58. If you were asked to do your own TB control on your farm where would you expect to get help 
from, and what would be your expectations about the nature of this help? 
MAF/AHB 
Regional Council 
Government 
RARe 
Neighbours (incl. forestry . . .  ) 
Veterinarian 
Other (specify) 
None 
B D 
B 
o o 
None of these, I would get out of farming cattle 
What would you expect? 
B 
59. Do you think movements of cattle from MC farms should be more strictly controlled? (e.g., 
only animals to slaughter) Yes 0 No 0 Less D Not known 0 
Why? ______________________________________________________ __ 
60. If compensation for reactors was removed altogether, what do you believe the effect on TB 
control would be? TB eradication achieved quicker 0 slower 0 no change 0 
Why? ______________________________________________________ __ 
61 .  Do you believe that control of TB in cattle would be achieved quicker if cattle farmers, like 
deer farmers, had to pay testing costs directly? Yes 0 No 0 Not known D 
Why? ________________________________________________________ __ 
348 
62. Do you think the current control scheme is satisfactory for controlling TB in New Zealand? 
YesO NoD Not known 0 
What improvements could be made to the current control policy/ what incentives could be 
put in place? ________________________________________________________ _ 
63. Over the last three years were there any management changes in relation to livestock policies. 
grazing management and pest control for your farm? (open question) 
64. What factors influence you most in deciding whether to conduct any TB control on your farm? 
65. Gender 
..  --. 
Increased farm profitability 
Decreased farm profitability 
Adverse climatic events 
Time (conflicts with other farm priorities) 
Reduction of the externally funded 
programmes (AHBIRCIFarmer group) 
Pressure from other farmers 
Others (please specify) 
male 0 female 0 
increase - -0- -- --
D 
D 
D 
D 
D 
D 
-� . .  
Decrease . . -cr--
D 
D 
D 
D 
D D 
66. What age group are you in? <20 years 0 20-30 0 30-40 D 40-50 0 50-60 D >60 D 
67. Do you live on the farm Yes 0 No 0 
68. Do you have any other employment commitments? Yes 0 No 0 
69. Can you cover your living expenses from your farm income? Yes D No D 
70. What is your relation to the property? 
Owner D 
Share milker D 
manager D 
lessee D 
other (specify) D 
(50% D. 39% D. 29% D. others D) 
7 1 .  Who makes decisions concerning grazing managementlbuying in animals? 
you solely 0 
you in partnership with someone else 0 
others (please specify their function) 0 
72. How frequent is contact between the decision-maker and you? 
349 
Daily 0, at least once a week D, at least once a month D, less than once a 
month 0 
73. How long have you been working on this farm? _______________ _ 
74. How many years have you worked in farming? _______________ _ 
75. How many years have you worked in non-farm jobs? _____________ _ 
76. What is your farming background 
started farm job without having any farming background D 
brought up on farm D 
other (please specify) D 
77. What kind of farm-specific formal qualifications do you have 
none 
technical agricultural institute 
diploma/degree 
other (please specify) 
B 
B 
We would like to hear any comments you have on the current TB control policy, what could be 
improved, what could be changed, or any comments to the questionnaire. 
This is the final question and here we would like you to give us an assessment of your personal 
characteristics. These questions were created for a dairy farmer in England who has got six dairy 
farms . Despite the similarity of the farms (sheds, number of animals, location, breeding material) 
he got different milk yields for each of the farms and he attributed this to the managers. Therefore 
he asked a psychology professor to set up a questionnaire, looking at personal characteristics of his 
managers and farm management. After conducting this questionnaire with many farmers and 
managers in England, they found that there is a certain zigzag pattern which makes a good animal 
person, others which make a good communicator and so on. With this question we would like to 
study if the same principles also hold for New Zealand farmers. 
350 
For this purpose, please try to describe yourself on the scale in relation to the following 
characteristics. (For example, if you think you are very easy going please tick the box to the right, 
if you think you are in between 'easy going' and 'not easy going' please tick the box in the 
middle.) 
You do have the option not to fill in these questions, if you feel uneasy about it, but I assure you 
that all information is treated anonymously. 
Not easy going D D DDD Easy going 
Meek D D DDD Not meek 
Patient D D DDD Impatient 
Unsociable D D DDD Sociable 
Not modest D D DDD Modest 
Persevering D D DDD Giving up Easily 
A worrier D D DDD Not a worrier 
Cheerful D D DDD Grumpy 
Talkative D D DDD Not talkative 
One who speaks one's mind D D DDD One who keeps quiet 
Difficult to get on with D D DDD Easy to get on with 
Lacking confidence D D DDD Confident 
Liking change D D DDD Suspicious of  change 
Forceful D D DDD Giving in easily 
One who prefers machinery D D DDD One who prefers animals 
One who prefers buying a new D D DDD One who prefers choosing a new animal machine 
Dislike to learn D D DDD Very keen to learn 
Still learning 0 0 000 Very knowledgeable 
One who likes to avoid hard work D D DDD One who values hard work 
One who dislikes using records D D DDD One who likes using records 
One who values traditional ways D D DDD One who likes adopting new ideas 
One who does not like to set targets D D DDD One who likes setting targets for himlherself 
One who likes to look after his/her D D DDD One who likes to strictly monitor favourite animals a bit better than performance of the herd 
the rest 
If you have any questions, please feel free to contact me any time. 
Thank you very much for your co-operation, which is very much appreciated. 
A I B I c I D I E I F I G I H I I I J I K I l I M I N 
c-!- Dairy farm: Partial lJlldgetiNg oN.fllrm control programmes 
� compensation: 65% reactor scenario 
� Returns 5 -> 2 5 -> 0.5 5 -> 0 2 -> 0.5 2 -» 0 1 -> 0 
+ Number saved 3 4.5 5 1 .5 2 1 � Increased prices reactors: % $Ihd � Dairy cattle » 2 yrs 35% $ 850 returns ($) 893 1 339 1 488 446 595 298 
� 
-4- OFF MC 0 0 1 0 1 1 9 %discount 1 0% 
10 Increased prices "white tagged' Number $Ihd rt1 Dairy cattle » 2 yrs 50 $ 850 returns $ - $ - $ 4 ,250 $ - $ 4 ,250 $ 4 ,250 
12 Dairy cattle 1 -2 yrs 6 $ 750 $ - $ - $ 450 $ - $ 450 $ 450 
13 Dairy Cattle 6wks - 1 yr 1 0  $ 375 $ - $ - $ 375 $ - $ 375 $ 375 14 
15 f-:1s % change 0.006 0 .009 0.01 0 .003 0 .004 0.002 f-:J7 Kg Ms $IkgMs in milk production 
18 Increased milk production 71 401' $ 3.63 return ($) 1 ,555 2,333 2 ,592 778 1 ,037 51 6 19 20 21 Total additionlll return. 2.447.74 3,671.62 9,154.57 1,223.87 6,7(16.33 5,891).91 22 i-='- Costs 23 
-
$I"r 4 firs costs 
� Management (extra labor) 1 00 $ 1 2  $ 1 ,200 $ 1 ,200 $ 1 ,200 $ 1 ,200 $ 1 ,200 $ 1 ,200 
4 est cost 
4- Poisonltraps tor possum control 1 80.00 $ 1 80.00 $ 1 80.00 $ 1 80.00 $ 1 80.00 $ 1 80.00 $ 1 80.00 � cl!- Total additioNal costs $ $ $ $ $ $ � 1,38(1 1,380 1,380 1,38(1 1,380 1,380 � 
tJ » (\) "C (j) "C 
� (I) :l 
5' a. 
Ci)' _. ..... >< 
()' :c:: 
3 en 8. "C .. !l CD 
0- I» a. .., tn 
� =r 
�' (I) CD 
� ut 
� c: tn (I) (I) a. 
.... 0 .. 
(I) (') 
0 :l 
0 
3 
_. (') 
I» 
:l 
1L 
Returns - Cost 1,067.74 33 2,291.62 7,774.57 -156. 13 5,326.83 4,510.91 
I» 
-< tn (I) tn 
� � 
3 � 5 S-7 8 9 10 11 12 13 14 -'--'--� � JL 
1 8  
19 -=--
R 4 � � 4 � � :£L � � 
30 -
A I 8 I c I D I E I F I G I H 
Beef breeding farm: P,,,till  blldg.tllfg Olf.fllrlfl cOlftrol progrll",,,, .. 
Returns 
Increased prices reactors: 
8eet breeding callle > 1 yr 
%discounl 
Increased prices "White tagged": 
8eet breeding callle > 1 yr 
Beef breeding 6 wks - 1 yr 
steers8.bulls 6 wks-1 8mnlhs 
Breeding bulls 
Costs 
Managemenl (extra labor) 
AddKional fencing 
Poisonl\raps for possum conlrol 
compens8lion: 65%'" 
% $Ihd 
35% $ 
1 5%'" 
Number $Ihd 
1 6  $ 
21 $ 
42 $ 
Number saved 
695'" returns 
OFF MC 
695'" returns 
31 5 
495 
$ 2,798 
TOtllllldditio"III ,..tll,.". 
no of IIrs $Il1r costs 
1 1 0'" $ 1 2  
est cost 
$ $ 1 ,000 
$ $ 950 
Totllllldditiolflll com 
Returns - Cost 
I I I J 
reactor scenario 
5 -> 2 "' 5 -> 0 .5 
3 4.5 
$ 730 $ 1,095 
0'" 0 
$ - $ -
$ - $ -
$ - $ -
$ 730 $ 1,095 
$ 1 ,320 $ 1 ,320 
$ 1 ,000 $ 1 ,000 
$ 950 $ 950 
$ 3,270 $ 3,270 
-2,540.25 -2, 115.38 
I K I L I M I N I 0 
5 -> 0 ... 2 -> 0.5 2 -> 0 1 -> 0 1 6  -> 2 
5 1 .5 2 1 1 4  
� w VI IV 
iD 
3 
5' 
e;;' 
$ 1,216 $ 365 $ 487 $ 243 $ 3,406 :::-(')' 
1'" 0 1 1 0 3 
() 
� 
$ 1 ,668 $ - $ 1 ,668 $ 1 ,668 $ -
$ 992 $ - $ 992 $ 992 $ -
-
0-
$ 3,1 1 9  $ - $ 3,1 1 9  $ 3,1 1 9  $ - .., 
tr 
m $ $ $ $ 6,995 365 6,265 6,022 $ 3,406 -to tr 
Ci 
(I) 
$ 1 ,320 $ 1 ,320 $ 1 ,320 $ 1 ,320 $ 1 ,320 9-:, ca 
$ 1 ,000 $ 1 ,000 $ 1 ,000 $ 1 ,000 $ 1 ,000 ar 
$ 950 $ 950 $ 950 $ 950 $ 950 3 
(I) 
$ 3,270 $ 3,27 $ 3,270 $ 3,270 $ 3,270 
3,125.00 -2,905. 13 2,995.25 2,152.00 135.50 
- - - - --------------------- -- � � - --
A I B I c I D I E I F I G I H I I I J I K I l I M 
� Beef finishing farm: Partial blldg�ting on.farm control programm� 
2 compensation: 65% reactor scenario 
-
3 Returns 5 -» 2 5 -» 0 .5 5 -» 0 2 -> 0 .5  2 -> 0 1 -» 0 
-
3 5 1 .5 2 � Number saved 4 .5 1 � Increased prices reactors: % $Ihd � steers&non-breeding bulls>1 8mnths 35% $ 970 returns $ 1 .01 9 $ 1 .528 $ 1 .698 $ 509 $ 679 $ 340 r1- steers&bulls<c1 8 mnths 35% $ 695 � 
9 f-:!O Total additional r@turns $ 1,019 $ 1,523 $ 1,693 $ 509 $ 679 $" J4t) I-'-'-� Costs 1 2  � hrs. $Ihr costs '14 Management (extra labor) 55 $ 1 2  $ 660 $ 660 $ 660 $ 660 $ 660 $ 660 � est cost Ts Additional fenCing $ 500 $ 500 $ 500 $ 500 $ 500 $ 500 $ 500 '17 '18 Poisonltraps for possum control $ 450 $ 450 $ 450 $ 450 $ 450 $ 450 $ 450 
: 1 9  
1 20 Total additional costs $ $ 1,61t) $ 1,610 $ 1,61t) $" 1,610 $" 1,610 � 1,610 
1 22 
1 23 Returns - Cost ·591.50 ·82.25 81.50 ·1, 100.15 ·931.00 · 1,210.50 24 I--
354 
.- . 
355  
AND LAST, BUT NOT LEAST 
Thanks is also due to all chocolate manufacturers - their products kept me going for all these years. 
I would not dare to count up the kilograms of chocolate I have eaten during the years of my PhD, 
but it is certainly more than ten times that ofthe average New Zealander, 
who eats 2.2 kg chocolate per year3. 
3Gray, A. (2000) The World Cocoa Market Outlook, http://www .acri-cocoa.org/acriJLMCrep1 .pdf.