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 INTERRELA TIONSHrP BETWEEN SOCIAL SUPPORT, 
RISK-LEVEL AND SAFETY INTERVENTIONS FOLLOWING 
ACUTE ASSESSMENT OF SUICIDAL ADOLESCENTS 
A thesis presented in partial 
fulfilment of tbe requirements for 
tbe degree of 
DOCTOR OF PHILOSOPHY IN SOCIAL POLICY AND SOCIAL WORK 
AT MASSEY UNIVERSITY, PALMERSTON NORTH, NEW ZEALAND 
DEBORAH ANNE SMITH 
2001 
(ii) 
ABSTRACT 
This project was undertaken largely to address (a) the concern that New Zealand has 
one of the highest rates of youth suicide in the world, and (b) the limited empirical 
research available on crisis assessment and intervention for suicidal adolescents in 
New Zealand. Research on youth suicide has primarily focused upon examining 
factors which place youth at-risk for suicidal behaviour. Social support was chosen as 
a variable of interest in this study due to (a) its importance in fostering healthy 
adolescent development, and (b) its identification as an important factor in increasing 
risk of suicide, particularly if it is lacking or of a negative nature. 
The present investigation tested a model comprising three constructs: social support 
(i.e., negative and positive), assessed risk-level of suicide, and safety interventions 
(e.g., hospitalisation, respite care). Two studies were conducted: an archival study and 
a vignette study. For the archival study, a record review was conducted using acute 
assessment reports from the Child, Adolescent and Family Service (CAFS). Data 
from 50 attempter files and 50 ideator files were collected in order to establish reliable 
measures for the vignette study. The vignette study involved administering a vignette­
style questionnaire to 23 CAFS clinicians. 
With the exception of the Children's Global Assessment Scale (CGAS) scores in the 
attempter group, the interrater reliability was good on all indices for the archival 
study. The vignette study indicated adequate reliability for risk-level ratings based on 
the multi-rater kappa The archival study demonstrated that there were significant 
(iii) 
interactions between group and risk-level (recoded), group and negative support 
severity (recoded), group and positive support, and negative support severity and total 
safety interventions (recoded). The vignette study revealed significant interactions 
between negative support severity (without or with positive support) and assessed 
risk-level, negative support severity (without or with positive support) and total safety 
interventions (recoded), and assessed risk-level and total safety interventions 
(recoded). Overall, results from this study indicated that: (a) the greater the level of 
negative support severity, the higher the risk-level; (b) the greater the risk-level, the 
greater the number of safety interventions implemented; (c) the presence of positi ve 
support, in addition to negative support, appeared to result in lower risk-level 
assessments, and (d) certain risk-levels were indicative of particular safety 
interventions. Revisions to the social support model were necessary based on the 
results obtained. 
For the vignette study, clinicians' responses with respect to the decision-making 
process for risk-level and safety interventions were also explored using the principles 
from a grounded theory approach and inductive content analysis. The results indicated 
that clinicians use a methodical process when assessing risk-level and making safety 
intervention recommendations. Process models for assessing risk-level and 
recommending safety interventions are presented in relation to these findings. 
This study makes several important contributions to the research on youth suicide by: 
(a) providing evidence for reliable social support concepts - namely, that of negative 
• 
(iv) 
support, positive support and negative support severity, (b) assessing the relationship 
that both negative and positive support have with suicide risk-level, (c) examining the 
relationship between risk-level assessment and specific individual safety 
recommendations (other than hospitalisation), and (d) providing evidence of a 
relationship between negative support severity and recommended safety interventions, 
not previously tested. 
The implications of these results are discussed in tenns of their application to 
(a) youth suicide treatment and prevention., (b) current or proposed services and 
procedures for at -risk youth, and (c) future research. 
(v) 
ACKNOWLEDGEMENTS 
The completion of this dissertation was made possible through the encouragement and 
willing involvement of many people. I am indebted to my chief supervisor. Associate 
Professor Ruth Anderson, for her sympathetic understanding, advice and ongoing 
encouragement throughout my research. Her humour was much appreciated in an 
effort to help alleviate some of the stress associated with the completion of this thesis. 
I am also grateful to my second supervisor, Associate Professor Andrew TrIin, for his 
helpful feedback. advice and encouragement. Overall. I wish to thank my supervisors 
for their constructive criticism and excellent feedback on the various drafts of this 
dissertation. 
I am also indebted to the staff and patients of the Child. Adolescent and Family 
Service (CAFS) in Porirua and Wellington for whom this project would not have been 
possible. The staff offered a supportive environment in which I could conduct my 
research. I am grateful to Diana Kane, Senior Clinician and Consulting Psychologist, 
of the CAFS team for her support and guidance throughout the data collection phase. 
I am very thankful to Dr. Ted Drawneek, Computing Services, Massey University, for 
his help in teaching me the mysteries of SPSS. I would also like to extend my 
appreciation to Duncan Hedderley, Research Officer, Statistics Research and 
Consulting Office, Massey University, for his expertise and advice surrounding the 
statistical analysis for this project. 
(vi) 
Thanks are also extended to the following funding bodies for their support toward 
completion of this dissertation: the Massey University Scholarship Committee for 
their award of a Doctoral Scholarship, the School of Sociology, Social Policy and 
Social Work for an award from the Graduate Research Fund, and Child, Youth and 
Family (CYF) for a grant-in-aid. 
The Massey University Human Ethics Committee and the Wellington Ethics 
Committee are much appreciated for their helpful comments, feedback and approval 
of this study. 
The Head of School, Professor Robyn Munford, School of Sociology, Social Policy 
and Social Work, as well as the school's administration staffwere very instrumental 
in offering their support during the course of this project. I thank them for answering 
my many queries and for pointing me in the right direction when needed. 
I would also like to extend gratitude to my good friends, Bill and Josi Wilson, who 
were like fumily to me while I lived in New Zealand. Their caring and compassion, as 
well as their ability to offer a safe place to escape when I needed it, were of a genuine 
nature and will never be forgotten. I would like to thank my good friends Tim and 
Jane Holloway-Jones for their support and encouragement throughout the duration of 
this project. They truly believed in me and reminded me constantly that I would 
complete this project. A heartfelt thank you to the Ballentyne family for their 
hospitality, support and lasting friendship both during this project and beyond. I 
(vii) 
would also like to thank my good friend in Canada., Amanda Dematos, for her 
assistance in proofreading several drafts of this thesis. 
Finally, I would like to extend a warm thank you to my parents, Geri and Gary Smith, 
and my son, Elliot, for their kindness, love, generosity and ongoing support. 
I am indebted to my family for teaching me how to make the impossible possible. 
Deborab Anne Smitb 
May, 2001 
(viii) 
Table of Contents 
Page 
Abstract 11 
Acknowledgments v 
Table of Contents V1l1 
List of Tables XlV 
List of Figures xvi 
CHAPTER 1- INTRODUCTION AND OVERVIEW 
1.1 Overview of Youth Suicide Research and Justification 1 
for Present Study 
1.1.1 Statistics on Youth Suicide in New Zealand 1 
1.1.2 The Need for the Examination of the Assessment and Intervention 6 
Process for Suicidal Adolescents 
1.1.3 Empirical and Theoretical Research on Social Support 9 
and Suicide 
1.1.4 Government Guidelines 10 
1.1.5 Professional and Personal Experiences 12 
1.2 SummaI)' and Outline of Present Research Project 13 
1.3 Presentation of Chapters - Outline 1 5  
CHAPTER 2 - THEORETICAL FRAMEWORK OF  SOCIAL SUPPORT AND 
RESEARCH LINKING SOCIAL SUPPORT, RISK-LEVEL 
AND SAFETY INTERVENTIONS TO YOUTH SUICIDE 
2.1 Introduction 19 
2.2 Theoretical Conceptualisation of Social Support 20 
(ix) 
Table of Contents (continued) 
Page 
2.3 A Comparison of the Principles of Social Support 32 
2.4 Differentiating Levels of Suicidal Behaviour 37 
2.5 Family Conflict, Social Support and Suicide Research 48 
2.6 The Relationship Between Perceived Social Support and 54 
Youth Suicide 
2.7 Risk-Level Assessment and Clinical Decisions for Safety 60 
2.8 Social Support, Risk-Level and Safety Interventions: Toward 71 
A New Conceptualisation 
2.9 Summary and Synthesis of Principles and Social Support Model 86 
2.10 The Archival Study and its Relationship to Empirical and Theoretical 88 
Material on Social Support, Risk-Level and Safety Interventions 
2.11 Research Questions - Archival Study 89 
2.12 Research Hypotheses - Archival Study 90 
2.13 Relationship of the Archival Study to the Vignette Study 93 
2.14 Research Questions -Vignette Study 94 
2.15 Research Hypotheses -Vignette Study 95 
CHAPTER 3 - METHODOLOGY - ARCHIVAL STUDY 
3.1 Introduction 97 
3.2 Description of Setting and Archival Records 97 
3.3 Measurements - Operationalisation 101 
3.3.1 Negative Emotional Support 102 
3.3.2 Positive Emotional Support 103 
3.3.3 Negative Support Severity 105 
(x) 
Table of Contents (continued) 
3.3.4 Children's Global Assessment Scale (CGAS) 
3.3.5 Risk-level 
3.3.6 Safety Interventions 
3.4 Procedure and Process for Data Collection 
3.4.1 Training the Research Assistant for Coding Data 
3.4.2 Procedure of Review for Coders 
3.5 Data Analysis 
3.6 Summary 
CHAPTER 4 - METHODOLOGY - VIGNETTE STUDY 
4.1 
4.2 
4.3 
4.4 
4.5 
4.6 
4.7 
Introduction 
Description of Participants 
Data Collection Instrument - Vignette Questionnaire 
Procedure and Process for Vignette Study 
Data Analysis 
Ethical Issues - Archival and Vignette Study 
Summary of Vignette Study 
CHAPTER 5 -ARCHIVAL STUDY RESULTS 
5.1 
5.2 
5.3 
5.4 
Introduction 
Reliability Statistics 
Sample Characteristics Derived From Patient Records 
Support Variables Compared by Group 
Page 
106 
108 
110 
112 
113 
115 
118 
120 
122 
123 
123 
130 
133 
138 
142 
144 
145 
145 
149 
(xi) 
Table of Contents (continued) 
Page 
5.4.1 (A) Negative Emotional Support, Negative Support Severity, 149 
Negative Support Source and Negative Support Items 
5.4.2 (B) Positive Emotional Support, Positive Support Source 154 
and Positive Support Items 
5.5 Level of Support 157 
5.6 Impairment and Risk -Level Measures 157 
5.6.1 CGAS - Impairment Measure 157 
5.6.2 Risk-Level Measure 160 
5.7 Safety Interventions 161 
5.8 Loglinear Analysis 164 
5.9 Summary and Conclusion of Archival Study Results 168 
CHAPTER 6 - VIGNETTE STUDY RESULTS 
6.1 Introduction 171 
6.2 Reliability Statistics 172 
6.3 Experience Analysis 173 
6.4 Loglinear Model and Chi-Square Analysis for Social Support Model 176 
6.5 Chi-Square Analysis for Risk-Level Assessment and Safety 186 
Interventions 
6.6 Correspondence Analysis (CA) and Multiple Correspondence 189 
Analysis (MCA): Individual Safety Interventions and Risk-Level 
6.7 Scale Utility for Risk-Level Data: Descriptive Data 194 
6.8 Qualitative Data Analysis - Overview 195 
6.8.1 Inductive Content and Thematic Pattern Analysis: Reasons for 196 
Safety Interventions 
(xii) 
Table of Contents (continued) 
6.8.2 Inductive Content and Thematic Pattern Analysis: 
Decision-Making Steps 
6.9 Summary and Conclusion of Vignette Study Results 
CHAPTER 7 - DISCUSSION AND CONCLUSION 
7.1 Introduction 
7.2 Risk Factors - Archival Study 
7.3 Social Support - Archival Study 
7.3.1 (A) Negative Support and Negative Support Severity 
7.3.2 (8) Positive Support 
7.4 Impairment and Risk-Level Measures - Archival Study 
7.5 Safety Interventions - Archival Study 
7.6 The Social Support Model- Archival and Vignette Studies 
7.6.1 The Decision-Making Process -The Vignette Study 
Page 
203 
205 
210 
210 
215 
215 
217 
218 
219 
220 
221 
7.7 The Revised Social Support Model - Archival and Vignette Studies 222 
7.8 Methodological Issues 235 
7.9 Implications and Future research 240 
7.10 Conclusion 248 
REFERENCES 251 
(xiii) 
Table ofCmltents (continued) 
Page 
LIST OF APPENDICES 
APPENDIXA. Coding Guide and Scoring Sheet for Archival Study 272 
APPENDIXB. Case Studies for Training 291 
APPENDIXC. Questionnaire for the Vignette Study 340 
APPENDIXD. Consent and Information Sheet for the Vignette 371 
Study 
APPENDIXE. Coding Guide for Vignette Study 375 
APPENDIXF. National Ethics Application for Vignette Study and 381 
Letters of Approval for Archival and Vignette Study 
APPENDIXG. Supplementary Analyses Removing the Effect of 405 
Abuse in the Archival Study 
APPENDIXH. Observed and Expected Frequencies for the Loglinear 410 
Model Comparing Negative Support Severity (Without 
or With Positive Support), Risk-Level and Total Safety 
Recommendations in the Vignette Study 
(xiv) 
Table of Contents (continued) 
Page 
LIST OF TABLES 
TABLE 1. Structural and Affiliative Definitions for Social Networks. 26 
TABLE 2. Summary of Study Findings - Relationship between Family 49 
Conflict and Availability of Social Support to Suicide Research. 
TABLE 3. Summary of Study Findings - Relationship between Perceived 55 
Level of Social Support and Suicidality. 
TABLE 4. Descriptive Statistics and Chi-Square Analyses for 
Demographics by Group. 
147 
TABLE 5. Stressful Life Events Preceding Suicidal Behaviour and 150 
Chi-Square Analysis by Group. 
TABLE 6. Negative Support Severity (Recoded) and Chi-Square Analysis 151 
by Group. 
TABLE 7. Negative Support Source and Chi-Square Analyses by Group. 152 
TABLE 8. Chi-Square Analyses of References in Reports to Negative 153 
Support by Group. 
TABLE 9. Positive Support Source and Chi-Square Analyses by Group. 155 
TABLE 10. Chi-Square Analyses of References in Reports to Positive 156 
Support by Group. 
TABLE 11. CGAS (Recoded) Statistics and Chi-Square Analyses by Group. 159 
TABLE 12. Risk-Level Measure (Recoded) and Chi-Square Analysis by 160 
Group. 
TABLE 13. Safety Interventions and Chi-Square Analyses by Group. 162 
TABLE 14. Frequencies for Total Number of Safety Interventions (Recoded) 163 
by Group. 
TABLE 15, Safety Interventions (Time Specificity) and Chi-Square Analyses 164 
by Group. 
TABLE 16. Frequencies for Risk-Level Responses by Vignette. 173 
(xv) 
Table of Contents (continued) 
Page 
TABLE 17. Chi-Square Analysis of Negative Support Severity 179 
(Without or With Positive Support) (Recoded) and 
Risk-Level. 
TABLE 18. Chi-Square Analysis of Negative Support Severity 181 
(Without or With Positive Support) (Recoded) and 
Total Safety Interventions (Recoded). 
TABLE 19. Chi-Square Analysis of Risk-Level by Total Safety 184 
Interventions (Recoded). 
TABLE 20. Chi-Square Analyses of Individual Safety Interventions by 187 
Risk-Level Assessment. 
TABLE 21 . Mean and Standard Deviation Statistics for Scale Utility 195 
Data by Vignette. 
(xvi) 
Table of Contents (continued) 
Page 
LIST OF FIGURES 
FIGURE 1. Current research model incorporating social support (negative 82 
support severity and positive support), risk-level 
and safety interventions 
FIGURE 2. Graphical representation of correspondence analysis (CA) 191 
results for safety interventions and risk-level 
FIGURE 3. Graphical representation of multiple correspondence analysis 192 
(MCA) results for safety interventions and risk-level 
FIGURE 4. Summary of inductive content and thematic pattern analysis for 197 
reasons related to choosing safety interventions 
FIGURES. Summary of inductive content and thematic pattern analysis for 205 
the decision-making process for safety interventions 
FIGURE 6. Schematic representation of the relationship between "elements" 232 
of the revised social support model