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. E X P E R I M E NTAL STU D I ES O F H U MAN CROWDI N G : A TEST O F TWO MODELS . A thesis presented in partial fu lfi lment of the requirements for the degree of Doctor of Phi losophy in Psycho logy at Massey University KEITH FRANCIS TUFFIN 1990 ABSTRACT Fol lowing a review of the history of human crowding the attributional arousal model of crowding was examined. This mode l views the experience of crowding as spatial ly induced arousal , attributed to the closeness of others. The model i ntroduces the possibi l ity of misattribution of arousal source leading to a reductio n i n perceived crowding . A salience hypothesis was developed in order to account for the ease with which misattribution was predicted to occur. Study One attempted to test the salience hypothesis. The study varied two levels of distance and fou r levels of information concerning arousal source in attempting to partially repl icate one of the key studies on which the attributional arousal model is based. The 320 subjects were run in g roups of five. The distance manipu lation was effective in i nducing crowding in subjects seated at close distances. However, the predicted levels of i ncreased arousal fai led to emerge . This unexpected result precluded any test of the sal ience hypothesis, s ince i ncreased leve ls of arousal are necessary i n o rder for causal search and m isattri bution to occur. This result showed that the experience of crowding may occur i ndependently of arousal. The i nformation expectancy model was deve loped to explain the resu lts from Study One. This model assumes i nvasion of personal space is necessary for the i i experience of crowding, and also that confi rming of expectations wil l reduce the impact of crowding. Study Two aimed to clarify the role of information and expectations in the context of experimental crowding . A further aim was to gather psychometric data on the efficacy of measurement scales for the concepts of arousal and crowding. Study Two varied three levels of spatial information and two levels of nonspatial i nformation. The 240 subjects were run i n g roups of five and all were seated at close i nterpersonal distance. Resu lts showed the measurement scales possessed high levels of i nternal consistency. However, accurate spatial i nformation fai led to reduce the impact of crowding and this result calls i nto question the uti lity of pre-exposure i nformation . Disconfirmed nonspatial expectations increased levels of reported crowding suggesting that accurate i nformation concerning the activities which occur while in crowded conditions is important. Sex differences emerged and these contri buted to the view that women may be more adaptable than men under conditions of reduced interpersonal distance. The results of these studies were considered in terms of their impl ications for the models of crowding. Suggestions for fu rther research were discussed. These included examin ing the relationship between spatial i nvasion and crowding , and further consideri ng the impact of information on the experience of crowding. ACKNOWLEDGEMENTS I wish to gratefu lly acknowledge the unfaltering support and commitment of my fi rst supervisor John Podd. The he lpful advice and encouragement which I received from John Spicer and M i ke Sm ith is also acknowledged with the warmest appreciation . My thanks also go to the staff of the Department of Psychology who have been of assistance in so many ways. i i i TABLE OF CONTENTS ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i ACKNOWLEDGEMENTS . . .. . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i i i C HAPTER O N E I NTROD U CTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 1 OVERVI EW . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 H ISTORY OF CROWDING . .. . . . . . . . . . . . . ..... . . . . . . . . .... 2 DISTINCTIONS AND DEF IN ITIONS . . . . . . . . . . . . . . . . . . . . . . . .. 10 MODELS OF CROWDING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 18 CHAPTER TWO ATTRIB UTIONAL MOD ELS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 CHAPTER T HR E E A N EXAM I NATION O F THE TWO FACTOR M O D EL O F C ROWDING 36 SPATIAL INVASION . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 40 AROUSAL . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .. . . . . . . . . ... 42 COGNITION . . . . . . . . . . . . . . . . .. . . . . , .. .... . .......... .. 46 Attributio n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Attention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 I n fo rmation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 54 Sal ience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 56 PERFORMANCE DEFICITS AND INTERPERSONAL HOSTILITY . . . 57 RATIONALE FOR STUDY ONE . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 C HA PTER FOU R M ETHOD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 OVERVIEW AND DESIGN . . . . . .. . . . . . . . . . . . . . ...... . . . . . 64 SUBJECTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 64 EXPERIMENTAL MATERIALS . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 6 5 PROCEDURE . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . 70 ETH ICAL CONSIDERATIONS AND DEBRIEFING . . . . . . . . . . . . . . 73 UNIT OF ANALYSIS . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 74 C HAPTER FIVE R ESU LTS ........................................... 77 AROUSAL . . . . . .. .. . .. . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . 77 CROWDING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 78 PERFORMANCE .. .. . .. .. . . .. ... ... . . . . . . . . . . . . . . . .. . . 81 I NTERPERSONAL RELATIONS .. . . . .. ... ... . .. . . . . .. . . . . . 83 OTHER M EASURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 85 THE RELATIONSH IP BETWEEN AROUSAL AND CROWDING . . .. . 86 I NDIVIDUAL ANALYSIS .. .. . .. .. . . ... . ... . . . . . . . . . . .. . . . 87 C HA PTER SIX DISCUSSI ON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 88 C H A PT E R SEV E N A N I N FO R MATIO N EXP ECTANCY M O D E L O F C ROWDI NG . ... . 102 P REDICTIONS OF THE INFORMATION EXPECTANCY MODEL . . 108 HYPOTH ESES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 112 Appraisal Crowding 112 112 Arousal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 113 Annoyance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 114 Interpersonal Relationships . . . . . . . . . . . . . . . . . . . . . . . . . 115 Affect . .. . . ... .. .... .. . . . . . . . . . . . . . . . . . . . . . . . .. 116 Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Sex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 M EASUREMENT OF AROUSAL AND CROWDING . .. ......... 119 Arousal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 119 Crowding . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . 124 C H A PT E R EIG HT STUDY TWO : M ETHOD 129 OVERVIEW AND DESIGN .. .. . ... .. . .. . .. ........ . . .... 129 SUBJECTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 130 EXPERIM ENTAL MATERIAL S . . . . . . . . . . . . . . . . . . . . . . . . . . .. 132 P ROCEDURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 137 ETH ICAL CONSIDERATIONS AND DEBRIEFING . . . . . . . . . . . . . 141 CHAPTER N I N E STU DY TWO : R ESU LTS 144 PSYCHOMETRIC PROPERTIES OF THE SCALES MEASURING AROUSAL AND CROWDING . . . . . . . . . . . . . . . . . . . . . . . . . .. . 144 MANIPULATION CHECKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 146 DISTANCE MAN IPULATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 I NFORMATION MAN IPULATION AND SUBJECTS' APPRAISAL . . . 148 DIFFICULTY OF ANAGRAMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 SEX . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . ' ... . . . . . ...... 152 CROWDING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 153 ANNOYANCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 155 AROUSAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 I NTERPERSONAL RELATIONSH IPS . . . . . . . . . . . . . . . . . . . . . . . 160 PUN ITIVENESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 AGGRESSIVENESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 162 L IKING FOR OTHERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 AFFECT . . . . . . . . . -. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 164 PERFORMANCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 CHAPTER TEN STU DY TWO : D ISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 MEASUREMENT OF AROUSAL AND CROWDING . . . . . . . . . . . . 167 MANIPULATION CHECKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 I nterpersonal Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 173 Anagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 SEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 CROWDING AND ANNOYANCE . . . . . . . . . . . . . . . . . . . . . . . . .. 175 AROUSAL . . . . ... . . .. . . .. . . .. . . . .. . . . . . .. . . . . . . . . . . . 182 I NTERPERSONAL RELATIONS . . . . ... . . .. . . . .. . . .. . . . . .. 182 AFFECT . . . . . .. . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 185 PERFORMANCE ... .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . .. . 186 SUGGESTIONS FOR FURTHER RESEARCH . . . . . . . . . . . .. . . . 186 IMPL ICATIONS FOR THE INFORMATION EXPECTANCY MODEL . 190 C H A PTER ELEVEN S U MMARY AND CONC LUSIONS . . . . . . . . . . . . . . . . . . . . . . . .. 192 SUMMARY . . . . .. . .. . . . .. .. . . .. . . . . . . . . . . . . . . . .. . . . . . 192 CONCLUSIONS . . . . . . . . . . . . . . . . . . .. . . . .. . . . . . . . . . . . . . 197 R E F E RENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 202 APPENDIX A: EXPER IMENTAL PRE-INSTRUCTIONS (STUDY ONE) .. . . 216 APPENDIX B: POST-EXPERIMENTAL QUESTIONNAIRE (STUDY ONE) . 219 APPENDIX C : NOTES TO SUBJECTS (STUDY ONE) . . . . . . . . . . .. . .. 224 APPENDIX D: ANOVA TABLES FOR KEY DEPENDENT MEASURES FOR STUDY ONE (GROUP ANALYS IS) . . . . . . . . . . . . . . . . .. . . . . . . .. . . . 229 AP PENDIX E : ANCOVA TABLE FOR CROWDING CONTROLLING FOR AROUSAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 AP P ENDIX F: ANOVA TABLES FOR SELECTED DEP ENDENT VARIABLES SHOWING IND IVIDUAL ANALYSIS FOR STUDY ONE . . . . . . . . . . . . . .. 239 AP P ENDIX G : EXP ERIMENTAL P RE-INSTRUCTIONS (STUDY TWO) . . . 242 AP P ENDIX H : P RACTICE AND TEST ANAGRAMS . . . . . . . . . . . . . . . . . 245 AP P END IX I : P OST-EXP ERIMENTAL QUESTIONNAIRE (STUDY TWO) . . 248 AP P ENDIX J : CROWDING RESEARCH NOTES TO SUBJECTS (STUDY TWO) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255 AP PENDIX K: ANOVA TABLES FOR KEY DEP ENDENT MEASURES FOR STUDY TWO (GROUP ANALYSIS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 LIST O F TAB LES Table 1. Means and standard deviations for ratings of arousal . . . . . . . . . . . . 78 Table 2. Means and standard deviations for ratings of crowding .. . .. ...... 79 Table 3. Means and standard deviations for ratings of confinement . . . . . . . .. 80 Table 4. Means and standard deviations for the number of words derived from the master word . .. .... . . ... . .. . .. ..... . . ....... . . .. .. . . .. 82 Table 5. Frequencies for extent of personal space invasion . . . . . . . . . . . . .. 148 Table 6 . Means and standard deviations for number of p ractice anag rams correct pe r condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 150 Table 7. M eans and standard deviations for number of test anagrams correct per condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Table 8. Means and standard deviations for ratings of crowding as a function of sex and information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 Table 9 . Means and standard deviations for ratings of crowding as a function of sex and task difficu lty . . . . . . . . .. . ... . . ... ....... . . . . . . . . 155 Table 10. Means and standard deviations for ratings of annoyance as a function of sex and i nformation . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . .. 156 Table 11. Means and standard deviations for annoyance ratings as a function of sex and task difficulty . . . . . .. . . ... . . . . . . .. . . . . . . . ........ . . 157 Table 12. M eans and standard deviations for arousal as a function of information and task difficulty ..... . .. .. .... ..... . . . .... . .... . . . . . . 159 Table 13. Means and standard deviations for somatic as a function of information and task difficulty ..... . .... .. . . . ... ... .. . . . . . . . . . . . . . . 160 Table 14. Means and standard deviations for the Love-Punishment scale as a function of information and task difficulty . . . . . . . . . . . . . . . . . . . . . 161 Table 15. Means and standard deviations for aggressiveness as a function of i nformation and task difficulty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 Table 16. Means and standard deviations for g roup performance as a function of i nformation and task difficulty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 LIST O F FIG U R ES Figure 1. Diagram of the attributional arousal model showing spatial invasion leading to increased arousal and interpretation . . . . . . . . . . .. .. . . . . . . . 39 CHAPTER O N E INTRO DUCTION OVERVIEW Crowding may be regarded as a human response to spatial variables. It is a subjective experience which becomes more l ikely as the distance between people is reduced. 1 The social psychology of crowding has produced several models, one of which is examined in detai l in this thesis. The attributional arousal model is unique i n predicti ng that the experience of crowding may be reduced as a result of misattribution of spatial ly induced arousal. The first experiment is an attempt to partially replicate a key supporting study (Worchel & Yohai , 1 979) and also extend the model. The fai lure to replicate basic features of the supporti ng study, or to provide support for fundamental predictions of the attributional arousal model , stimu lated the development of a new model of crowding. The i nformation expectancy model of crowding i ncorporates the personal space app roach to crowding along with an emphasis on the importance of information in deve loping accurate expectations about spatial concerns. The second study in this thesis tests the predictions of this model. 2 H ISTORY OF CROWDING In reviewing the history of crowding Altman (1 978) identifies three perspectives which have been influential in shaping the manner i n which crowding is both u nderstood and studied. Each of these perspectives has contributed to contemporary views of the nature of crowding , and for that reason are important to consider. The three perspectives (the socio logical tradition based on urban analysis , animal research , and psychological research) and thei r approaches to the study of crowding wi l l be presented. The urban soci ol ogi st s examined, with the use of correlational techniques, re lationships between high concentrations of human popu lation and i ndicators of social pathology such as mental i l lness and crime. The sociological examination of crowding was stimulated by Le Bon ( 1 903) whose seminal work, "The Crowd", analysed the processes which led to mob behaviour. Le Bon was an i nfluential figu re who, accordi ng to Moscovici (1 986) , was one of the first scho lars to consider the masses as anything other than a behavioural anomaly. This increased attention fol lowed the twin processes of i ndustrialisation and u rbanisation which resulted in more people col lecting together in cit ies. Estimates I vary about the extent to which the worlds population has shifted. The last century has seen an i ncreasing percentage of the population living in cities. Davis (in Insel & Lindgren , 1 978) estimates that in 1 850 on ly 2% of the world's population l ived in cities of over 1 00 ,000 inhabitants. In the mid 1 970s 24% of the world's population l ived in cit ies, whi le in the year 2000 an estimated 40% wil l do so. It was th is process of large numbers of people col lecting together in cities that stimu lated Le Bon (1 903) , who Moscovici (1 986) credits with providing the i mpetus for a new psycho logy : the psycho logy of the masses. Whi le reflecting increased awareness of the u rban isation of the popu lation this new psychology referred to "the crowd", rather than crowding. In this respect the psycho logy of the masses was relatively undifferentiated in considering masses, mobs and crowds as one. Le Bon's psychology of the masses was based on the view that a person does not behave in the same way when alone and when in a crowd. The extent of the crowd's influence is sti l l very much of i nterest today. Before examin i ng contemporary views of crowding, it is worth considering Le Bon's ideas, which provide an interesting perspective on the i nfluence of crowds. 3 M oscovici ( 1 986) discusses Le Bon's contribution as originating from a consideration of the crowd. Once people form a crowd they thi nk, fee l and behave differently. Three "symptoms" were apparently manifest when in a crowd : 1 . lowering of intel lectual faculties, 2. intensification of emotional reactions ( includi ng a d imming of rationality) , 3. disregard for personal profit, people becoming both unselfish and i rresponsible. Le Bon argued that in crowds people are more l ike ly to engage in antisocial acts due to the anonymity and sense of invu lnerabi l ity which the crowd affords. This process of individual submersion into the wider group subsequently came to be studied by social psychologists i nterested in the effects of deind ividuation (e.g . , Fest inger, Pepitone, & Newcomb, 1 952) . 4 Le Bon ( 1 903) was almost singu larly pessimistic in h is view of the crowd's inf luence on the individual's behaviour. He wrote that when in a crowd, the i ndividual "descends several ru ngs in the ladder of civi l isation" (p.36) . Farr (1 986) maintai ns that these views were a response to the social and political events of the t ime. Both i ntel lectual interest in the nature of crowds and the changing natu re of social influence were germinal i n the conception of social psychology. Not all authors saw the crowd in such negative terms. McDougall ( 1 921 ) , for example , reacted against the i rrationality of Le Bon by suggesting that it was on ly through participation in g roup life that "man becomes fully man" and manages to rise above the leve l of the savage. This issue of the positive or negative influence of crowding on behaviour is one which sti l l pervades the current l iteratu re. The u rban sociologists extended these early views and sought to establish re lationships between areas of high popu lation and social pathology. In exami n ing areas of high popu lation these early studies considered the re lationship between popu lation density and such i ndices of social pathology as mental i l lness (Faris & Dunham, 1 965) , suicide (Schmid , 1 955) , and crime rates (Schmid , 1 969) . These studies typically found higher levels of the indices of social patho logy i n central city areas when compared to the less densely populated suburbs. Altman (1 978) reports that these studies yielded moderate relationships which led to the conclusion that crowding was harmfu l to human wel l being. This 5 conclusion was tempered, however, by a number of criticisms. Altman discusses these as bei ng fi rstly, the problems of interpretation ; that is, covariation does not necessari ly mean that popu lation density causes social pathology. Secondly, that th is approach is unable to ascertain the underlying dynamics because the sociological analyses used large (and variable) indicators.of population density. It was rare , for example, for these studies to be based on people or fami lies. A more typical u nit was people per acre, reflecting the societal level approach to these p roblems. Thirdly, these studies had no clear conceptual framework or u nderstanding of the concept of density. Also , these studies were firmly based in the t radition of popu lation density rather than employing the concept of crowding. F reedman (1 980) has also criticised the contribution of the urban sociologists. He notes that when factors such as i ncome and educational level are taken i nto account , these studies have found no substantial relationships between density and i nd ices of social patho logy. Epstein (1 980) is sim i larly critical of these studies i n concluding that their resu lts are inconsistent, with negative effects sometimes being documented , and other times not. The animal studies of crowding, which began to appear in the late 1 950s, resulted in startl ing findings, not the least of which was a series of studies (Calhoun , 1 962) documenting the dramatic effects of long term crowding. A colony of Norwegian rats was p laced i n an experimental enclosure where they were free to mU ltip ly. When the colony reached 80, the population was held constant by the removal of new offspring . A breakdown in the use of avai lable space occurred, the rats congregated in l imited parts of the enclosure such that density levels were very high in those areas. This behavioural i rregu larity became known as a "behavioral sink". Manifestations of the behavioural sink included abnormalities in females' abil ity to reproduce and care for offspring, disrupted nestbui lding, hyperactivity , pansexuality, and cannibalistic behaviour. Calhoun concluded that these abnormalities were so severe that over t ime the breakdowns in normal reproductive functioning wou ld resu lt in the extinction of the colony. 6 Calhoun ( 1 962) had specifically designed this enclosure to encourage unequal distribution of rats in various parts of the enclosure. His research showed that unusual ly high levels of density may lead to social deterioration . Simi larly, Christian , Flyger, and Davis ( 1 960) studied deer living in overcrowded conditions and noted increased disease and death . generalise particularly when th is involves complex social behaviour. For this The animal studies were popu lar with the research community because they offered certain advantages. Specifical ly, the long term effects of population density were able to be telescoped due to shorter l ife expectancies of the animals. Detai led autopsies were possible as were experimental variations in livi ng conditions. For ethical reasons such variations wou ld not be possible with human subjects. The principal difficulty with the animal studies is the problem of general ising these results to human crowding. Freedman ( 1 972) notes that although the animal findings are fascinating and provocative , it is difficult to 7 reason the animal literature shou ld be regarded merely as a starting point. Cohen , M alpass , and Klein ( 1 980) compare the human and nonhuman literature. . . They note that a stress syndrome, typically comprised of endocrine and behavioural responses which contro l and limit population g rowth, is usually found i n nonhuman popu lations. The absence of a simi lar mechanism amongst humans has resu lted in some interesting explanations. Of some interest is the suggestion that humans may have developed cognitive mechanisms (associated with the evolut ion of symbolic behaviour) which mediate between objective density and perceived crowding. It shou ld be remembered that the animal studies examined the long term effects of h igh concentrations of population density. As wi l l be discussed shortly , the re lations hip between density and crowding is far from an exact one. One final point wh ich deserves comment is that the dramatic and exclusively negative results from the animal studies contributed substantially to the debate as to whether crowding has a positive or negative influence on behaviour. In particular, Calhoun's ( 1 962) experimental theatre provided a very powerfu l stage on which to display the s inister potential of high and sustained population density. The an imal literature also functioned to stimu late psychological In terest i n crowding . This interest has further been excited by the fact that crowding is a human experience which we are al l able to relate to. The Guinness book of records (McFarlan , 1 989) reports that the largest crowd was established in 8 January 1977. An estimated 12.7 mi l l ion people attended the Hindu festival of Kumbh-Mela at Allahabad, Uttar Pradesh , I ndia. It is almost impossible to imagine what being part of such a huge mass of people is like. However, we are al l able to understand what being part of a more normal sized crowd is l ike. Reports of crowds usually rely on some est imate of the number of people, these being readi ly avai lable in the case of large sports events o r concerts. Such figu res are unavai lable, however, on crowded summer beaches, or to indicate the number of shoppers on the streets, or the number of bargain hunters in the stores. Yet these experiences may be just as relevant to being crowded as being part of a measured crowd at a contro l led event. The notion of crowding involves more than a mere estimation of the number of people i nvo lved. There is more to this particu lar human experience than simply the number of participants. Recently, envi ronmental researchers have started aski ng questions about the experience of crowding in o rder to find out more about the factors which contribute to it. These questions have been fue l led not on ly by concern about our welfare but also by intel lectual curiosity about the natu re of an experience which is readily avai lable, yet poorly understood. The concern centers around a fear of overpopulation and the possible negative concomitants of scarce resources, including food and territory. Such overcrowding may be seen as a ' th ird world' problem, but in Western industrial ised countries related problems are also of concern . Examples include the overcrowding which occurs when cities are unable to adequately accommodate thei r inhabitants , and the crowding related tragedies which have occurred at soccer matches. 9 A further spur to the development of psychological interest in studyi ng crowding was a widespread increase in awareness of environmental issues. During the 1 960s awareness of pol lution, popu lation control and conservation of resources became public issues. Psychology has never been slow to respond to the Zeitgeist and it is perhaps no accident that environmental psychology experienced its most rapid growth during the 1 960s (Holahan, 1 982) . I nterest in the effect of the environment on behaviour had been around for a long time , but the Zeitgeist of the 1 960s certain ly sharpened research interest in the fie ld . It was towards the end of that decade that saw the publication of the first professional journals in environmental psycho logy (Gifford , 1 987) . Crowding was but one topic of research interest which was carried along on the wave of activity which occurred at that time. Experimental social psycho logists began publ ishing their research findings on crowding i n the early 1 970s. Among the fi rst of these studies was work by Freedman , Klevansky , and Ehrl ich ( 1 971 ) who brought the phenomenon into the laboratory by having small groups of subjects work on a variety of tasks i n either a crowded or an uncrowded room. Laboratory studies of crowding have been criticised (Pau lus, McCain , & Cox, 1 979) on the g rounds that realistic stress levels are impossible to achieve for eth ical reasons. Paulus et al. argue that subjects know that participation in a laborato ry experiment wi l l be of short duration and this i nflates the artificiality of the experience. These drawbacks are not i nsignificant but they must be considered in the l ight of the control afforded by laboratory experiments. 1 0 The three t raditions which provide a h istorical context to the study of crowding have been presented. In chronological order these are fi rstly, the socio logical tradition wh ich documented corre lations between population density and social pathology. Secondly, the dramatic results obtained by the animal literature which further conso lidated the notion that "crowding" resulted in negative consequences. Thirdly, the youngest "tradition" to contribute to the study of crowding , the psycho logical perspective. To date this perspective has on ly been considered briefly on the basis of the question of why psychologists have become involved in crowding research. The next section wi l l consider how psychologists define crowding. D ISTINCTIONS AND DEFIN ITIONS A usefu l beginn ing to the problem of understanding what i s meant by the term "crowding" is to consider what it is not. More specifical ly, to examine how it differs from some closely al l ied terms. An important distinction exists between crowding and overcrowding. When considering crowding there may be discussion over the issue of whether positive or negative connotations prevai l . In the case of overcrowding , discussion is 1 1 unnecessary as the term has been invoked to reflect the problematic nature of the situation , and also to imply something about the cause of the problem . Prison overcrowding is a case in point. Farrington and Nuttal ( 1 980) suggest that overcrowding in prisons may produce violent or disruptive behaviour. Simi larly, it has been reported (Cox, Pau lus, & McCain , 1 984) that densely popu lated prisons have increased rates of suicide , i l lness, aggression and ru le violation. Stokols ( 1 972) made an important distinction between density and crowding. Density re lates to a physical d imension which involves the space avai lable for each person. Crowding is seen as an experiential dimension whereby the restrictive aspects of limited space are perceived by the individuals exposed to them . Density is thus a necessary but not a sufficient condition for the experience of crowding. Density is regarded as an objective physical ratio of number of people per unit of area, whi le crowding is a subjective feel ing. Stokols sees crowding as an aversive experience which develops when the avai lable space is perceived to be less than adequate. Worchel and Brown ( 1 984) regard crowding as a negative psychological state o r feel ing. The phenomenological view of crowding promoted by Stokols (1 972) has not met with u niversal acceptance. Griffit and Veitch ( 1 971 ) and also Freedman ( 1 975) define crowding as amount of square footage per person, and thereby equate crowding and density. Altman ( 1 975) is supportive of Stokols' definit ions. He argues that density is strictly a physical quality with no inherent psychological mean ing . Density in this respect seems to occur in a psycho logical vacuum. It is 1 2 no more than a measure of people per unit of space. Such measures are simi lar to the reports of crowd size ("1 8,000 people attended the game") , and are usefu l in p roviding an i ndex of the physical property of a particular crowd, but add little to our understanding of what the experience would be l ike . This last point h ig h lights one of the key distinctions between the study of crowds and the psycho logy of crowding. Crowds are typically viewed from the outside, from the perspective which emphasises the behaviour of the mass. The psychology of crowding typical ly adopts a "within" perspective which considers the individual in the mass (Kruse , 1 986). The phenomeno logical view suggests that the experiential component of crowding is of p rime importance. For example, Stockdale ( 1 978) refers to crowding as being a perceived and subjective state. Crowding wou ld, therefore , seem to be in the eye of the beholder. Holahan ( 1 982) also discusses the subjective approach to defi ning crowding but notes that the objective parameters of density are easi ly defined compared to the complex subjective elements which lead to the perception of crowding. Thus, an important aim of crowding research is to further develop the definit ion of crowding. Freedman (1 975) argues against the phenomenological view. He maintains that crowding should not be restricted to a subjective perception, but should refer to the amount of objective space available. Of course , it is possible to measure both . Freedman' s view fol lows di rectly from his assertion that the experience of crowding need not necessarily be a negative one, but that the nature of the 1 3 experience be dependent on the particu lar ci rcumstances in which it occurs. This is an i mportant issue within the l iterature as it questions a basic assumption , namely , that the experience of crowding is something which is regarded as universally and psychological ly aversive. The point is that crowding need not necessarily be experienced as aversive or noxious. Freedman's research bears testimony to this , as does the fact that we may often attend crowded events which are also h igh ly enjoyable. One view of crowding which combines both the physical and subjective aspects of crowding is the personal space approach. Numerous authors espouse this view (Sommer, 1 969; Altman , 1 975; Worchel & Teddlie , 1 976 ; and Vine, 1 982) . Spatial i nvasions are regarded as the source of density related stress. Worchel and Teddl ie, for example , defi ne crowding as an intrusion of one's personal space. Sommer ( 1 969) defined personal space as an area around a person's body into which others may not enter. Holahan ( 1 982) simi larly defi ned it as the zone around an individual into which others may not trespass. A convenient metapho r is that of a "bubble" of space which surrounds each o f us. Personal space extends above and behind the i ndividual, but its most usual function is to serve as an i ndex of i nterpersonal distance between people. There is some dispute within the literature (e.g., Patterson , 1 975) as to whether personal space or interpersonal distance ( IPO) is the more appropriate term. 1 4 Strube and Werner ( 1 982) define IPO as simply the distance between two individuals. They argue that IPO is observable to both parties in an interaction and is l ike ly to be used as a communication cue. IPO may also mediate the effectiveness of other communication cues, for example, facial expression and voice qual ity. Personal space , on the other hand, is characterised by subjective ownership . It is defined as the area surrounding a person's body, and considered to be the exclusive space of that person. Personal space is not observable and may be g reater o r smaller than the distance separating i nteractants. Simi larly, personal space may extend in all di rections and may vary in shape and size. Patterson ( 1 975) and Strube and Werner ( 1 982) wou ld prefer to see the term "personal space" abolished. They suggest a number of reasons why the concept of personal space may be misleading. Firstly, the idea of personal space involves the notion of a comfortable distance which is fairly stable. Patterson h igh l ights the variabil ity associated with personal space (th is ranging from no distance at al l to several feet) depending on factors such as gender of interactants, cu ltu re , relationship and situational constraints. This variabi l ity raises concerns about the "stabi lity" of the concept. When we talk of a person's personal space, to which particu lar personal space are we referring? Secondly, it is argued that the term "personal" is misguided, as it impl ies individual ownership as if the space is somehow attached to the individual. However, th is space ach ieves meaning only in relationship to another individual. We do not identify a personal space with respect to inanimate objects, but we may identify a distance between people. This distance is carefu l ly negotiated between individuals using a 1 5 variety of verbal and nonverbal cues. IPD has the advantage of referring specifically to the distance between the individuals (and implicitly recognising that this wi l l vary). Fu rther, the term focuses attention on the relationship between individuals. Moreover, it is more closely l inked to the operational definitions used i n the measurement of the concept. Despite the advantages of using the term IPD, a number of models use the personal space notion as the term of choice. Personal space models of crowding (e.g . , Worchel & Teddlie , 1 976) see spatial vio lation leading to arousal which is then attributed to the presence of others in the environment. Thus , the experience of crowding is seen as resulting from a vio lation of personal space rather than density. This approach re lies on a concept (personal space) which includes a consideration of the question of density from the perspective of the individual. The personal space approach therefore differs from a strict density view by consideri ng the space (interpersonal distance) which separates i ndividuals rather than the number of people in the space avai lable to them (much of which may be unused). Greenberg and Firestone (1 977) supported th is view with the findi ng that perceptions of crowding can be instigated by i nterpersonal intrusion. They conclude that high density is not a necessary precondition for perceived crowding. It i s usefu l to consider the distinction between social and spatial density at this point. Social density (McGrew, 1 970 ; Loo , 1 972) is dependent on the number of people in a particular area. High social density occurs when a large number of 1 6 people occupy a specific area, and low social density occurs when a small number of people are in the same area. Thus, crowding based on social density depends on large numbers of people. Spatial density, however, holds the number of people constant and varies the avai lable space. High spatial density depends on a small spatial al location , whi le low spatial density may occur in a larger area. Social density is defi ned in terms of the number of people i n a given area, whi le spatial density is defined in terms of the available space in a particu lar setting . Gifford ( 1 987) uses the example of classroom size to clarify the distinction . Imagine a classroom with 30 students. The density (ratio of individuals to area) may be doubled by either adding another 30 students, an increase i n social density, or by partitioning the room i n half thereby increasing the spatial density. Mathematically the ratios stay the same with either of these two procedures doubl ing the density. Psychologically the two procedures will lead to different outcomes (McGrew, 1 970) . Aiel lo , Thompson, and 8rodzinsky ( 1 983) acknowledge the possibi l ity of overlap between the two. They equate h igh social density with " too many others" and h igh spatial density with " too l ittle space". A personal space defi nition of crowding depends on the operational isation of social density. Another definition which relies on the notion of social density is provided by Oesor ( 1 972) who defines crowding as excessive perception of conspecifics. Such a definit ion emphasises the perception of people, rather than the actual number of people. One l imitation of this defi n ition is that we are not g iven any real u nderstanding of what constitutes " excessive perception". 1 7 Conceptual and definitional dispute characterises the research into crowding. Insel and Lindgren ( 1 978) refer to crowding as a hypothetical construct. With regard to the problematic nature of defining the construct they see it as havi ng sim i lar status to the notion of intel l igence. They decl ine to define crowding as they bel ieve it better to simply talk about crowding as though everyone understands what it means. The precedent for this was set by researchers in the area of i ntel l igence some 50 years ago. Disputes over definitions should perhaps be expected g iven the diversity of discip l ines which have been invo lved in investigating the nature of human spatial behaviou r. The history of crowding presented in this chapter has considered the influence of three key perspectives. However, Baldassare ( 1 978), a sociologist , discusses the contributions o f various discipl ines : biologists, ethologists, anthropo logists, ecologists, envi ronmentalists and psychologists. He makes the observation that psychology has on ly recently come to the party. I n conclusion, crowding may be regarded as a subjective experience. Crowding , whi le dependent on the spatial relationships of those i nvolved , is u ltimately something which is in the eye of the beholder. This section has reviewed major distinct ions and found that crowding is distinct from overcrowding and also from density. A further distinction has been drawn between social and spatial density, with social density more closely associated with current conceptions of crowding. Recent definitions stress the physical and subjective e lements of crowding both of which are contained in an approach which emphasises the construct of 1 8 personal space. Personal space rel ies on the notion of social density. Mu lti­ discip l inary interest in spatial behaviour characterises the fie ld. This d iversity of perspective is m i rrored in the models of crowding which , i n general , attempt to provide an explanation of the processes invo lved in the e xperience of crowding. These models wi l l be considered next. MODELS OF CROWDING Epstei n and Karl in ( 1 975) discuss a social psychological model of experimental crowding based on normative expectations. According to this model , crowding occurs when the distance between i ndividuals is less than the "appropriate" d istance for that sett ing. One of the implications of this model is that expectations of what is appropriate are situationally determined. Another imp licat ion is that expectations dictate the comfortable d istance for interaction between i ndividuals. Epstein and Karl in assert that I PO mediates the experience of crowding . The basis of the Epstein and Karlin (1 975) model is the pioneeri ng work on PROXEM ICS, the scientific study of human spatial behaviour, conducted by Hal l ( 1 959 , 1 963, 1 966) . Hall identified four spatial zones which he referred to as interpersonal d istance zones. The I ntimate Distance Zone (0 - 1 8 ins) where the presence of the other is unmistakable and may be overwhelming due to increased sensory input. This zone is rich in its potential for communication , with 1 9 use being reserved for intimates. The Personal Zone ( 1 .5 - 4 ft) encompasses the characteristic spacing people use with each other. In the Social Zone (4 - 1 2 ft) i mpersonal business may be conducted. People who work together tend to use the close social zone, (each zone has both a close and a far phase) . This distance is also used by people attending casual social functions. The far phase of th is zone suggests formality in interactions. Finally, Hal l proposes a Public Zone ( 1 2 -25 ft) where communication cues are quite g ross. The far phase of th is zone is used for formal occasions , public speakers or people of very high status. Each zone includes a range of distances. Hal l (1 966) maintains that the distances are important because of the interpersonal communication possibi l ities which each offers. Space, and therefore distance , is seen as a vehicle for communication . The names of each zone serve as a clue to the type of activity and re lationship associated with each distance. Epstein and Karl in ( 1 975) fol low Hal l' s ( 1 966) lead in stating that in any situation there exists a set of expectations which dictate the comfortable distance for i nteraction. According ly, distances which fal l short of what is appropriate in a g iven situation produce social crowding. Further, disconfi rmation of expectations about appropriate spatial d istance may lead to stress reactions. This model is based on the foundation of environmental cognition , whereby situations are matched with preexisti ng views of what is personally regarded as correct for that situation. 20 Altman's ( 1 975) privacy contro l model also predicts that stress reactions wi l l result from crowding. According to Altman , crowding occurs when the privacy regu lation system fai ls , thus al lowing more social contact than is desired. This mode l assumes there are times when contact with others is sought, and other t imes when such contact is actively avoided. Crowding occurs when the desired level of p rivacy is greater than the achieved level of privacy. Altman's view has the experience of crowding being mediated by the construct of psychological contro l , the loss of which is regarded as an aversive experience. Whi le there is ample support in the literatu re to suggest that loss of control has unfortunate consequences (e.g . , Sel igman , 1 975) , it is Altman's proposal that this loss of cont rol is essential in the experience of crowding. P rivacy contro l is simi lar to Epstein and Karlin's ( 1 975) model in that crowding results from a conflict between what the individual either expects or desi res and what actual ly happens. A major difference between the two, however, is that Altman's ( 1 975) model includes the notion of contro l . The privacy regulation system is a control system whereby the individual is seen as being able to react to situat ions in an attempt to restore desi red levels of privacy. The notion of contro l may wel l be important in mediating negative psychological consequences, as demonstrated by Sherrod (1 974). Groups of eight subjects worked on tasks in either high density (smal l room) or low density ( large room). Some subjects were led to bel ieve they cou ld control their density ; they had the option of leaving and working in a less dense room , though the experimenter preferred them to remain . This simple manipu lation of control led to a reduced experience of 21 crowding by subjects i n the high density sett ing. Simi larly, aspects of control were manipu lated by Langer and Saegert ( 1 977). They used informational messages designed to provide predictive contro l. In high and low density supermarket setti ngs they provided subjects with i nformation about how they were l ike ly to fee l if the store become crowded. The man ipulation was successfu l in reducing discomfort and perceptions of crowding and was also associated with improved task performance. The concept of control is also impl icated in two further models of crowding : the information overload model and the behavioural constraint model. Overload mode ls (e.g . , M i lg ram, 1 970 ; Cohen, 1 978) postu late that perceived crowding resu lts when the amount of stimu li impinging on an individual exceeds informational processing capacity. Saegert ( 1 978) suggests that density is a very special case of cognitive overload because of the behavioural consequences of social stimulat ion. It is argued that high population levels make interactions difficu lt to predict and contro l , as we l l as i ncreasing problems of focusing on task re levant information in the environment. A basic assumption of the overload models is that i ndividuals have a finite attention capacity, which may be surpassed in a high density environment. Mi lgram (1 970) hypothesised that when large numbers of people occupy a restricted space whi le engaging in individual tasks, the amount and complexity of s ituationally relevant information can lead to attentional overload. M i lgram argued 22 that such overload may explain the impersonal detached mode of the typical u rbanite. More generally, the overload perspective emphasises the cognitive i nvo lvement of people in the experience of crowding. Density, per se, is not seen as the primary determinant of task performance, affect and social behaviour, but rather as a physical condition which can give rise to attentional overload when a person is engaged in active transactions with , or scanning of, the environment . Overload models assume that crowding necessari ly entai ls increased information and this may lead to difficulties in coping with the extra input of information . Support for attentional overload is rather mixed with Neiser (1 976) suggesting that there is no evidence that attentional capacities are so easi ly reached. Attention is a subjective notion , and in this respect is similar to Altman's ( 1 975) not ion of privacy control . C learly, individuals decide for themselves the level of p rivacy which is desirable. Similarly, they may decide at what point they can no longer cope with further information. Behavioural constraint models of crowding assume that crowded situations p lace behavioural constraints on people. Decreased behavioural freedom may lead to a loss of contro l , which may in turn exacerbate the negative consequences of crowding. Proshansky, Ittleson , and Rivl in ( 1 970) p ropose a constraint model which they label a goal blocking model. The model suggests that crowding is experienced when a person's goals are blocked by the density of the environment. Simi larly, Karl in ( 1 980) suggests that when certain conditions successfu l ly block goals then people wi l l experience negative emotions and wi l l 23 seek to explain the cause of these. While not invalidating the notion of behavioural constraint , crowding is viewed by this model as an exclusively negative phenomenon. It is defined in terms of its interruptive influence on the ach ievement of behavioural goals and the resulting negative affect. This need not necessarily result from crowding related processes and therefore should be regarded as a rather restrictive view of crowding. Karl in ( in Cohen et aI . , 1 980) also suggests that in seeking the cause of negative emotion people may either look to situational factors or to dispositional properties of the others involved in the setting. It is proposed that dispositional attributions wi l l resu lt in anger and aggressive or withdrawn behaviour. If, however, situational attributions are made then the environment may be label led as crowded. I ndividuals wi l l then explai n thei r responses, and the responses of others , as being caused by crowding. Such an attribution to the situation (crowding) may resu lt in attempts to exert contro l over the environment by trying to change it or escape from it. Behavioural constraint does not mean that ind ividuals rel inquish attempts to exercise control . Karl in further suggests that if escape or modification of the situation proves to be impossible then individuals may experience a loss of control with accompanying feel ings of hopelessness and depression . Under these circumstances, goal attainment becomes even more difficult. Behavioural constraint models , by definition , emphasise the negative aspects of crowding . They are based on the extent to which the experience of crowding 24 l imits people's options, induces helplessness (loss of control) and negative affect. One criticism of the behavioural constraint mode ls is that they focus too much on the negative aspects of crowding. This is not to deny that there are times when experiencing crowding wi l l be a totally negative experience. For example, I spoke to one couple who had attended an indoor motor show in Birmingham , Eng land in 1 978 where more than twice the maximum number of people expected attended. People became so crowded that, in this case , the couple's legs were raised off the g round and they completely lost control of the direction in which they wanted to travel . They were simply forced to go in the same general di rection as that which the crowd took them. They described thei r experience as extremely frightening . The behavioural constraint model is relevant to this sort of experience. The criticism , however, is that by focusing on the negative features of the phenomenon these may be seen as the on ly features. Thus, this particular mode l has nothing to say about a crowding experience where people's enjoyment of the experience is actually enhanced by the presence of others. A "good crowd" of people attending a party, for example, wi l l often faci l itate one's enjoyment of the event, more so than if the event is poorly attended. An interesting feature of the Karl in ( 1 980) model is the implied use of attributional processes. Attributional models of crowding wi l l be discussed more fully fol lowing the review of general models. This review wi l l conclude with a consideration of Freedman's ( 1 975) density intensity model . Freedman ( 1 975) l inks density and crowding , suggesting that density operates to 25 make others who are present more important stimul i in the situation . He acknowledges the importance of others in any situation, but argues that with low density their importance is minimised si nce there may not be interaction among them. Likewise there may be no competition for resources or no necessity for dealing with them in any way. As density i ncreases, people play a g reater role in the actions and reactions of others. I t is proposed that responses to others under conditions of high density wi l l be stronger and more intense. In essence, the medium of physical distance serves to magnify the impact of the stimu lus properties of the person. The intensification notion may be analogous to any stimulus as it increases in importance. Enjoyable music, for example , may be even more enjoyable when amplified. However, increasing the volume of music which we disl ike wi l l only i ntensify our dislike. What is suggested by the density-intensity model of crowding is that high density may have either good or bad effects. Schiffenbaur and Schiavo ( 1 976) demonstrated such an intensification effect by varying both i nteraction distance and quality of i nteraction. They found that close partners were l iked more in positive interactions and less in negative interactions. S imi lar support was forthcoming from a study by Storms and Thomas ( 1 977) where close distances intensified subject's l iking for friendly or similar confederates. The models of crowding reviewed include: Epstein and Karlin's ( 1 975) model based on normative expectations; privacy contro l (Altman , 1 975) based on the notion of personal contro l in re lation to crowding ; i nformation overload models 26 which invoke the assumption that attentional capacity may be exceeded by the conditions associated with high density ; behavioural constraint mode ls which emphasise the extent to which behavioural options are restricted by i ncreased density ; and finally the density-intensity model (Freedman , 1 975) which regards the process of increased density as being akin to magnification of social stimul i . The notions of normative expectations, perceived contro l , stimulus overload, behavioural constraint and density-intensity may be seen as competi ng positions, but may also reflect the fact that different processes are involved in the experience of crowding. Crowding is a complex phenomenon and it is of little surprise that such a diverse range of models has developed. Adding to the range of psychological explanations of crowding are the attributional models which wi l l be the subject of the next chapter. 27 CHAPTER TWO ATTR I B UTI O N A L M O D ELS - In 1 976 three separate publications appeared in which attributional p rocesses were involved in models of crowding. Schiffenbauer and Schiavo ( 1 976) ; Patterson ( 1 976) ; and Worchel and Teddl ie ( 1 976) , each independently embraced cogn itive explanations of crowding. Attribution research blossomed in the 1 970's, with its widespread impact felt in many areas of social psychology. The area of crowding was no exception, with the attribution literature being responsible for the birth of a new class of attributional models of crowding. These mode ls are important to consider as they are representative of a major contemporary stance in social psychology. These three mode ls are dist inct from the models presented in the previous chapter, because they share the assumptions that fi rstly, arousal mediates the experience of crowding ; and secondly, that i ncreased arousal triggers an attributional search. Sc hiffen bauer an d Sc hi avo (1 976) exami ned hypotheses derived from Freedman's ( 1 975) notion of density intensification. They suggest that subjects are aroused by close approaches and crowded environments. This arousal i nitiates an envi ronmental search for a reason, in which any cues may be used as an explanation for arousal. When positive cues are p resent, the arousal wi l l receive a positive label whi le the reverse wi l l occur when negative cues preside. 28 Schiffenbauer and Schiavo had confederates g iving subjects either positive o r negative comments on thei r performance on a problem solving task. I n the negative evaluation condition arousal was expected to be negatively labe l led with negative arousal being attributed to the confederate, who under these conditions is also thought likely to be disl iked. This explanation of crowding is based on the pivotal assumption that close interaction distances lead to increased arousal . This assumption has previously received some support (Middlemist, Knowles, & Matter, 1 976 ; McBride , King, & James, 1 965; D'Atri , 1 975 ; Zajonc, 1 965). P atterson ' s (1 976) model relates to a variety of nonverbal dimensions (including personal space) , but not specifical ly to crowding. It is important to conside r because it relies on s imi lar constructs as the other models. Patterson's basic proposal is an outgrowth of Argyle and Dean's (1 965) equ i l ibrium model of nonverbal behaviour. This latter model proposes that both approach and avoidance forces underl ie the display of nonverbal behaviours in social interactions. Approach forces derive from affi l iative needs and the desire for social feedback. Avoidance forces, on the other hand, i nclude fear of self­ disclosure and rejection . Approach and avoidance forces govern a numbe r of nonverbal dimensions i ncluding interpersonal distance , gaze, smi l ing and body posture. Collectively. these dimensions determine the level of intimacy, also referred to as the point of mutual comfort. The model proposes that i ntimacy can be maintained at an appropriate level (equi l ibri um) by compensatory changes in any one dimension counteracting imbalances in any of the other d imensions. Thus, a person who attends too closely to another may be met by a reduction in 29 gaze or an increase in interpersonal distance . Once a comfortable level of int imacy has been reached, any change requires a reciprocal change in order to maintain equi l ibri um . Patte rson's ( 1 976) arousal model of interpersonal intimacy is a major extension of equi l ibrium theory. The model proposes that sufficient changes in the i ntimacy behaviours of one person wi l l produce arousal changes in the other. Patterson argues that arousal changes may be usefu l in explain ing results which have indicated reciprocal reactions, that is, the complete opposite of that predicted by equ i l ibri um theory. Based on the assumption that reduced i nterpersonal distance wi l l i ncrease arousal , the model proposes that either substantial i ncreases or decreases in the immediacy (exceeding some range of subjectively rated comfortable i ntimacy) of one person wi l l result in a change in the arousal level of the other person. Depending on several variables (nature of the relationship, environmental setting , perceived control over the situation) , Patterson suggests the resu lting arousal wi l l be labe l led either positively or negatively. When change in i mmediacy resu lts in positively label led arousal , such behaviours wi l l be reciprocated. Conversely, when immediacy changes produce negatively label led arousal , such changes wi l l lead to compensatory reactions. Based on Schacter's ( 1 964) two factor theory of emotions, Patterson ( 1 976) argues that the mediating emotional states are a product of both arousal change and cognitions. A further assumption is that cognitions exercise considerable i nf luence on the emotional states, which in turn determine specific behavioural 30 responses. Patterson ( 1 978) emphasises the importance of arousal change suggesting that reduced arousal may be more important, especially in a situation where the person is fearfu l or distressed. Wo rchel an d Teddl ie (1 976) propose a two factor model of crowding , suggesti ng that the initial stage in the experience of crowding involves violations of personal space. Violations of personal space are assumed to cause the i ndividual to feel aroused. According to this view, spatial violations need not occur in high density situations. Personal space violations may occur even under conditio ns of low density, for example, when people congregate in a corner of a large avai lable area (high social density). A unique implication of the personal space approach to crowding is the possibi lity of one person crowding another. Whi le two people do not constitute a crowd the subjective nature of the crowding experience means that one i ndividual may sti l l invade another's space to the extent of inducing crowding . The other side of such views of crowding is the acknowledgement that spatial invasions do not necessarily resu lt in fee ling crowded. For example , i nvasions often occur at major rock concerts or sports events , but people do not report feel ing crowded at such events. Often such crowds are seen as exhi larating and adding to the enjoyment of the experience (Tuan , 1 977; cited in Worchel & Cooper, 1 983). Events l ike these run the risk of losing much of their excitement if viewed from a stadium which is nearly empty. I n an attempt to explain the experience of crowding , Worchel and Teddlie ( 1 976) focused on the attributional p rocess. They noted Schacter and Singer's ( 1 962) 31 work which suggests that arousal leads the individual to search the environment in o rder to explain and label the arousal. Schacter and Singer suggest the label which is selected wi l l be determined by envi ronmental cues. Accordingly, Worchel and Teddlie assume that spatial violation arouses the individual who is then motivated to search for the cause of arousal. Crowding wi l l occur if, and on ly if , the individual decides that the arousal is due to others being too close. If arousal is attributed to another cause , crowding wi l l not result. Thus, the spatial invasions which may be experienced at a football match wil l arouse spectators who wi l l attribute the sou rce of arousal to the excitement of the game and wi l l not feel crowded. Whi le there is some simi larity among these models, namely that they all re ly on the i nvolvement of arousal and corresponding causal search , there are also some important differences. A key difference is that Patterson ( 1 976) invokes attribution in the sense of attaching a label or interpretation to an emotional state. In contrast, Worchel and Teddlie ( 1 976) emphasise the causal search which is thought to occur before label ling. A second difference is that Worchel and Teddlie's model deals specifically with spatial considerations and is therefore d irectly l inked with crowding. Patterson's model accounts for changes across a number of nonverbal d imensions and may be regarded more as a model of i nterpersonal (dyadic) int imacy. Thirdly, Worchel and Teddlie maintain that causal m isattribution is easily induced, while the other two models provide for no such possibi lity. A final difference among these mode ls is that both Schiffenbauer and Schiavo ( 1 976) and Patterson maintain that arousal wi l l be either label led positively or negatively dependi ng on a variety of factors. The Worchel and Teddlie model does not include the labe ll ing of arousal as positive o r negative. Arousal is s imply attributed to one source or another. Three attributional models have been introduced. Recent developments and research pertinent to these approaches wi l l now be reviewed before one of the attributional models is considered in more detai l . 32 Morasch , Groner, and Keating ( 1 979) note that attributional models of crowding generally assert that perceptions of crowding result when a person experiences arousal, the cause of which is attributed to the presence of others in the environment. Morasch et al. discuss the possibil ity that such an attribution may or may not be accurate. They suggest that inaccurate causal attributions to others as the source of arousal may be made if it is more comforting than an accurate attribution would be. Morasch et al . attempt to l ink the attribution component of the perception of crowding with a motivational bias, the tendency for success to be attributed to internal factors, while fai lure is more l ikely to be attributed to external factors (Zuckerman , 1 979). Attributions to others in the environment as causing arousal are external attributions. Morasch et al. therefore predict i ncreased attributions to others in the envi ronment (and a corresponding i ncrease in perceptions of crowding) when fai lure is experienced. They predict a positive correlation between perceptions of crowding and fai lure. Their results support this p rediction ; however, the positive correlation between perceptions of crowding and fai lure on ly occurred when the fai lure was recorded with an activity which was rated as important. They conclude that fai l ing at an unimportant activity wi l l produce little or no arousal, thereby precluding the need for an attributional search. 33 Gochman and Keating (1 980) simi larly propose that individuals wi l l experience crowding when they become aroused and attribute this to some aversive aspect of the high density environment. Gochman and Keating suggest that individuals who are aroused by a source they are unable to identify wi l l actively search for an explanation for the arousal. If an external cue is readi ly avai lable as a label for the arousal , the person will attribute their arousal to that source. They propose a l imited scope for this search , as arousal wi l l cause the individual to focus on the most salient features in the envi ronment. Consequently, these sal ient features are most l ike ly to be judged as causes of arousal (Pryor & Kriss, 1 977; Taylor & Fiske, 1 978). Gochman and Keating ( 1 980) cite research evidence (e.g . , Ross, Rodin , & Z imbardo, 1 969; Storms & Nisbett , 1 970; Val ins & Nisbett , 1 972) indicating the ease with which individuals wil l make incorrect attributions for the ir arousal when such explanations are reasonable and apparent. They suggest that density could be mistakenly identified as the cause of nondensity i nduced arousal if it is sal ient relative to other possible explanations. They also h ighl ight the possibi l ity of the plausibi l ity of a crowding attribution being enhanced by the widespread cultural be lief that density negatively influences behaviour. It is also worth noting that this belief has found its way into the major theories of crowding. 34 Gochman and Keating (1 980) examined the proposition that arousal caused by facto rs unrelated to density can be misattributed to crowding , given a moderate ly high density environment. This hypothesis was strongly supported by their resu lts. I n high density environments, i ndividuals who experienced both disconfirmed expectancies and unattained goals made greater attributions to crowding than those whose expectations or goals were met. Vine's ( 1 982) model combines a personal space approach with an attributional element. The model is centered around a series of sequential steps , the first being a spatial i nvasion. This stimu lates an assessment of the perceived discrepancy between distance sought and the actual IPO (simi lar to the normative expectations model of Epstein and Karl in , 1 975). It is assumed that any discrepancy wil l result in arousal. The arousal stimulates an assessment of the causes of the arousal . Vine postu lates that what subsequent ly occurs depends on the assessments of responsibi lity for the spatial invasion and its personal significance. Vine argues that people react differently to "neutral thwarting" as opposed to "personal thwarting". Thwarti ng is used as if it is synonymous with the process of spatial invasion. Personal thwarting is perceived to be di rected to oneself, or at least the intruder is seen as personally responsible . Neutral thwarting is situational ly necessitated and i ndiscriminate. At this point, the model diverges significantly from other attribution models. It suggests aversion to either the environment or to the i ntruder. Following this, coping strategies are developed and perceptions of the i r success are also considered. 35 A number of attributional models have been reviewed, (Sch iffenbauer & Schiavo , 1 976; Patterson , 1 976 ; Worchel & Teddl ie, 1 976 ; and Vine, 1 982) . These models and the subsequent research which has developed from them has been considered. The models are all individual approaches ce-ntred around a common theme and sharing a common genesis. That genesis is the relationship between arousal and attribution in the context of crowding. This relationship and its most archetypical manifestation wi l l be considered in detai l i n the next chapter. 36 CHA PTE R THREE AN EXAMI NATI O N OF THE TWO-FACTOR MODEL OF C ROW D I N G The two-factor attributional arousal model (Worchel & Teddlie , 1 976) is one of several attributional models. A unique aspect of th is model is that i t al lows for the possibi lity of m isattribution of arousal source leading to a reduction in perceived crowding. Worche l ( 1 978) refers to the possibil ity of reducing the experience of crowding without increasing the amount of space avai lable to the individual. He regards this as particu larly important in view of the fact that on earth we have a fi n ite amount of space , yet the population doubles every 35 years. Thus , the experience of crowding may be reduced by cognitive rather than spatial means (Worchel & Yohai , 1 979). For this reason the model warrants carefu l consideration . Fi rst, an overview of the model wi l l be presented. Then, specific e lements of the model wi l l be examined and the major research question and rationale for Study One presented. The two factor mode l of crowding (Worchel & Teddl ie , 1 976) identifies invasion of personal space as the spatial variable associated with crowding, and attribution as the cognitive process through which the experience of crowding occurs. It is asserted that arousal wil l fo l low an invasion of personal space. Once aroused , the i ndividual wi l l attempt to find an explanation for the arousal . If the arousal is attributed to others being too close, the individual wil l experience crowding and will subsequently attempt to reduce this negative state. Crowding is assumed to be an aversive experience , such that its occurrence wi l l immediately resu lt in attempts to avoid, reduce, o r l imit the experience. 37 Worchel and Yohai ( 1 979) maintain that the experience of crowding wi l l result in performance decrements on complex tasks and also increased interpersonal host i l ity . Fo llowing the attribution of arousal to crowding , they suggest attention wi l l focus on alleviating the negative state of crowding . This demand on the individual's attention is predicted to cause performance decrements on complex tasks. Presumably these attentional demands are insufficient to cause any disruption to performance on simple tasks, although the distinction between simple and comp lex tasks is never made clear. Thus , the attribution to crowding impl ies disrupted performance due to the individual's mobi l isation of effort to reduce the negative state. I nterpersonal relations are predicted to suffer fol lowing an attribution of crowding because the focus on reducing the experience of crowding increases the salience of others i n the envi ronment. These others are not on ly viewed as di rectly causing the individual to feel crowded but also as frustrating attempts to reduce crowding . Two methods are proposed for manipu lating attribut ions in order to reduce the experience of crowding. Fi rstly, it is suggested that this may be effected by reducing the salience of others in the environment. Worchel and Teddlie ( 1 976) interfered with the attribution process by diverting attention away from the other people in the environment. This was achieved by invoking "attribution inhibitors" (placing interesting pictures on the wal ls of the room , showing a movie, or havi ng 38 some other distracting event occur). Worchel and Teddlie found that the presence of attribution inh ibitors reduced the extent to which subjects reported feel ing crowded. Additionally, they found increased performance (on an anagram task) and decreased agg ressiveness. The second proposed method of reducing the experience of crowding invo lves encouraging misattribution by means of providing an alternative source of arousal. It is possible , of cou rse, that the two methods overlap, as the pictures provided by Worche l and Teddlie ( 1 976) may have had an arousing effect for subjects. Worchel and Yohai ( 1 979) provided subjects with an alternative sou rce of arousal , by misinforming them about being exposed to a "subsonic" (inaudible) yet arousing noise. In fact no noise was used. The expectation of receiving an arousing noise was effective , however, in reducing perceptions of crowding and improvi ng task performance relative to subjects who were provided with no alternative explanation. The authors conclude that the process of attribution is a major component inherent i n the experience of crowding. Diagrammatical ly the Worchel and Yohai ( 1 979) model may be represented as shown i n Figure 1 . Each of the components of this mode l wi l l now be examined in detai l . ATTRIBUTION TO ALTERNATIVE SOURCE OF AROUSAL INVASION OF PERSONAL SPACE AROUSAL � I NTERPRETATION 39 --7 ATTRIBUTION TO CROWDING REDUCTION IN CROWDING ATTENTION FOCUSED ON ALLEVIATION Figure 1 . PERFORMANCE DEFICITS & INTERPERSONAL HOSTILITY Diagram of the attributional arousal model showing spatial i nvasion leading to i ncreased arousal and i nterpretation . 40 SPATIAL INVASION Hall ( 1 966) suggests that interactions are characterised by the distance between participants. Appropriate interaction distances wi l l be smaller for friends than for strange rs . Stress is predicted to fol low vio lation of the appropriate interaction distance. Research supports the view that spatial vio lations can create d iscomfort and arousal. Sommer ( 1 969) found spatial violations resu lted in avoidance of eye contact, changing of body position , and physical departure in o rder to re-establish personal space. Evans and Howard (1 973) and McBride et aI . , (1 965) systematically considered the effects of invasion on physiological arousal. For example, in the McBride et al . study , e levated galvanic skin responses were found when the personal space of subjects was vio lated. I n an unusual study, Middlemist et aI . , ( 1 976) demonstrated that violations of personal space can result in physiological arousal. They hypothesised that arousal , due to spatial i nvasion , wou ld result in delayed onset of urination . In a men's toi let they used a hidden periscope to measure how long it took men to begin u rination and how long they urinated when alone, when another man stood one urinal away, and when another man stood in the adjacent uri nal . In accordance with predictions, urination took longer to begin and also lasted less time when someone stood closer. The personal space approach to crowding implies that arousal may be generated by spatial invasion. Another implication of this approach to crowding is that sex of subject is an important variable. There is some suggestion in the literature that 41 females have smal ler personal spaces than males (Wi l l is, 1 966; Aie l lo & Aie l lo , 1 974; Evans & Howard, 1 973). These studies report that males typically fee l uncomfortable and may become hosti le in high density situations, whereas females do not. Evans and Howard note that research has shown repeatedly that the appropriate i nteraction distance is smal ler for females than for males. That is , the distance at which females feel comfortable when i nteracting is smal ler than the distance at which males feel comfortable. A personal space view of crowding therefore suggests that females wou ld have to be more densely packed together than males before experiencing personal space vio lations and crowding . Vine ( 1 982) maintains that crowding involves an invasion of a body-centered personal space zone. This zone is seen as an area which the individual wi l l regard as their own and seek to keep free from unwanted i ntrusions. Violation is l ikely to occur when spatial and/or social density is h igh . Vine argues that the strength of such an approach is that many sal ient factors influencing perceived crowding operate at the level of the interaction between any subject and specific others . Some of these may apply even in low density situations where no more than two people need be i nvolved. Thus, crowds are not an essential aspect of crowding , since dyadic encounters constitute a microcosm for most types of crowding experience. The personal space dimension of the experience of crowding has received general support in the research literature. Greenberg and Firestone ( 1 977) found that perceptions of crowding and also the experience of stress could be 42 instigated by interpersonal intrusion. In an interview situation , subjects whose personal space was intruded (they sat so close their knees were touching), reported increased crowding related stress, and also increased perceptions of crowding . Levitt and Leventhal ( 1 978) found that both density and intrusion of personal space concurrently affected the perception of crowding. Walden and Forsyth (1 981 ) similarly concluded that perceptions of crowding were influenced by interpersonal distance. Those subjects seated closer together rated themse lves as more crowded than those subjects seated further apart. An invasion of personal space is the basic foundation of the two factor attribution mode l of crowding . The research evidence suggests spatial invasion may produce discomfort and arousal . An implication of the personal space approach is that male subjects, because of thei r greater spatial requirements, wi l l more easily experience crowding than females. The suggestion being that a more rigorous test of any theoretical predictions shou ld be conducted with female subjects. The personal space view of crowding has received general support in the research literature . I t also has the advantage of permitting research into the experience of crowding without requiring crowds of people, a point which has not been lost on those researchers favouring laboratory experiments. AROUSAL Worchel ( 1 978) notes that with the exception of Freedman and col leagues (e .g. , Freedman , 1 975) most investigators agree that there is an arousal o r stress component i n crowding. This comment is not intended to suggest that arousal and stress are the same, but rather reflects the problems inherent in the measurement issues in this area. Worchel's model is predicated on the notion that arousal is i nvolved in the experience of crowding . 43 Evans ( 1 978) reviews four lines of evidence supporting the notion that crowding is mediated by arousal. Fi rstly, task performance data are generally in accordance with the Yerkes-Dodson law. Evans suggests the research has i ndicated s l ight facil itation or no effects of spatial impingement on simple task performance with decrements on complex tasks under the same conditions. Second ly , the psychophysiological data indicate heightened arousal under conditions of crowding and spatial invasion. These measures include skin conductance , cortiso l leve ls, blood pressure, heart rate and micturation. Thi rdly, some of the individual differences found in reactions to spatial l imitations are consistent with the proposed arousal mechanism. For example, extroverts allow much closer personal space approaches and become less aroused under such condit ions than introverts. Finally, observational data on tension and gaze behaviour are consistent with the view that crowding is mediated by arousal . Karl in and Epstein ( 1 979) refer to crowding as a "re liable" method for stress inductio n i n humans. Based on the physiological measures alone, they conclude that crowded subjects are almost invariably more aroused than thei r noncrowded counterparts. 44 Although there may be some consensus on the apparent involvement of arousal in the experience of crowding, this is not matched by agreement on what actually constitutes arousal under these circumstances. Worchel and Yohai ( 1 979) discuss arousal (without defin ing it) as being something which results from an invasion of personal space, something which subsequently needs to be explained, and something which if attributed to the closeness of others wi l l be label led as crowding. They further suggest that misattribution of the cause of arousal can resu lt in reduced reports of crowding . Finally, when presenting their results under the subheading "Arousal" they use the terms arousal and st ress interchangeably, and measure this notion of arousal with one question which asked subjects about the extent to which they fe lt "relaxed or il l at ease". The interchangeable use of terms may be described as careless and at the very least wi l l do little to foster progress in the area. Regarding the measurement of arousal, Worchel and Yohai ( 1 979) note that their results apply on ly to perceived arousal asserting that demonstrations of physiological arousal are unnecessary because perceived arousal is more di rectly li nked to the process of attribution. Worchel and B rown ( 1 984) expand on this point by stating that the model is concerned with how i ndividuals attribute perceived arousal. They note that the relationship between actual and perceived arousal is beyond the scope of the model. They maintain that the attribution process is set into motion by the "perception of arousal". Therefore , it is asserted that the relationship between actual and perceived arousal (despite its importance) has no di rect bearing on the question of whether the experience of 45 crowding is mediated by attributions. The lack of a clear definition of arousal is regrettable , especially as the concept is a pivotal one for the model. Defin itions of arousal are avai lable elsewhere in the crowding literature. For example, Epstein ( 1 980) refers to arousal as a state of activation . The general conclusion to be drawn from a consideration of the relationship between spatial invasion and arousal is that spatial i nvasion may resu lt in increased physiological arousal and also increased awareness of physiological indicators such as increased heart rate and skin conductance. Spatial invasion may also be accompanied by feeli ngs of discomfort which may be h igh ly correlated with physiological arousal. The need to explain such arousal is based on the assul11ption that subjects are careful assessors of their internal arousal states (Schacter & Singer, 1 962). Calvert-Boyanowksy and Levanthal ( 1 975) are critical of this assumption particu larly as it impl ies that the same care is not appl ied to complex external envi ronments. In the context of the two factor model it suggests that possible variables which may wel l contribute to increased arousal in an experimental setting (for example , uncertainty in the setting and performance evaluation anxiety) wil l not be seriously considered as potential sources of arousal in the attributional search. Final ly, the model asserts that individuals wi l l need to explain their spatially induced arousal. This assumes that arousal precedes cognition . The reverse 46 order seems just as p lausible, but this possibi lity is not addressed by the mode l . COGNITION In the diagram of the attributional arousal model (Figu re 1 ) th is section is represented as " interpretation". Th is refers to the cognitive aspects of the model and includes the processes of attribution and attention, and the related issues of pre-exposure information and sal ience. Attribution The two factor model l inks arousal via the process of attribution to the experience of crowding. If arousal is attributed to other people being too close then crowding wi l l be experienced. However, if arousal is attributed to another sou rce crowding wi l l not be experienced. The model suggests that when arousal is experienced the i ndividual wi l l search the external envi ronment and attribute the arousal to a sal ient and plausible source which is present. Schacter and Singer ( 1 962) pioneered research which examined the cognitive elements invo lved in the experience of an emotion. After arousing their subjects with injections of epi nephri ne , they informed them the drug wou ld either cause them to feel aroused (informed condition ) , or wou ld have effects other than arousal (misinformed condition) , or wou ld have no noticeable effects ( ignorant condition). They fou nd subjects were more l ike ly to report emotional sta�es consistent with environmental cues (euphoria or anger) in the misinformed and ignorant 47 conditions than in the informed and p lacebo (no drug) conditions. Schacter and Singer suggest that where subjects did not anticipate arousal , it was necessary for them to locate a source in the envi ronment that cou ld be used to cognitive ly label the internal state. For the informed and placebo conditions, labels were either provided by the experimenter o r were unnecessary when no arousal was experienced. Simi lar resu lts have been found with situationally induced arousal. In such studies (often referred to as misattribution studies) , arousal is induced by leading subjects to expect an imminent environmental stressor (e.g . , electric shock) . Misattribution is then induced by introduci ng another possible sou rce of arousal . Behavioural or self report measures are then taken to indicate how arousal has been labelled. Val ins and Ray (1 967) were able to reduce snake phobias by convincing subjects that their arousal was due to impending electrical shock and. not snakes. Simi larly, Storms and Nisbett (1 970) were able to reduce insomnia by making subjects believe thei r arousal was due to a placebo pi l l rather than to thei r personal problems. Ross et al. ( 1 969) found subjects were more wi l li ng to work for monetary reward than escape from impending electrical shock when they were convinced that their arousal was due to the presence of a loud noise rather than the shocks. The misattribution studies have been criticised by Calvert-Boyanowsky and Levanthal ( 1 975) who suggest the resu lts have been misinterpreted. They argue that no c lear evidence exists to show that misattribution was invo lved in these 48 studies. I nstructions in misattribution studies are designed to affect attributions but they also provide subjects with differing information about what to expect. Calvert-Boyanowsky and Levanthal suggest that it is the difference i n information which results in misattribution effects and not actual changes in subjects attributions. They examined this hypothesis and concluded that it was indeed the information in the instructions which accounted for the misattribution effects. They also found that g iving subjects accurate information resulted in less anxiety than g ivi ng inaccurate information. Similarly, Manstead and Wagner (1 981 ) reviewed the degree of support for Schacter and Singer's ( 1 962) two factor theory (on wh ich the two factor model of crowding is based) , and concluded that there is a disconcerting lack of empirical support for the theory. These criticisms imply that the attributional arousal model of crowding may not be based on the most solid of fou ndations. If in fact, Calvert-Boyanowsky and Levanthal are correct, it may be that the two factor model of crowding is based on a misinterpretation of results. The idea that the evaluation of crowding may be an attributional response to arousal has developed more recently in the literature (e.g . , Worchel & Teddlie , 1 976 ; Worchel & Yohai , 1 979) . Worchel and Teddl ie examined the hypothesis that the addition of attention distractors wou ld alleviate the experience of crowding . The attention distractors were six posters placed on the walls of the experimental room. It was proposed that these wou ld serve as attributional i nh ibitors and would distract the individual from making the attribution to crowding. Further, it was expected that the pictures wou ld only have an effect for individuals who were aroused and i n a position to make an attribution for the cause of this arousal. In fact, the addition of pictures to the environment had exactly the predicted effect, with the experience of crowding and the stress associated with it both bei ng reduced. 49 The strength of the Worchel and Teddlie ( 1 976) study is that it di rectly examines, and subsequent ly provides support for, the personal space approach to crowding. However, Worchel (1 978) acknowledges that this experiment offers on ly indirect support for the attributional hypothesis. One criticism is that the addition of pictures to the walls may have served to reduce arousal, or that they merely distracted subjects. That is , the pictures could have acted to influence processes other than the predicted attribution - crowding process. Paulus ( 1 980) s imi larly criticises the attributional approach in noting the difficulty in predicting whether subjects wi l l use density or other "coincident" stimuli for the attribution of their arousal. Pau lus also h ighl ights the apparent conflict between the resu lts of the Worchel studies and those of Langer and Seagert (1 977) who found that focusing subjects' attention (and presumably attributional tendencies) on density factors reduced the impact of the density. Worchel and Yohai ( 1 979) designed a study which more di rectly examined the role of attribution i n crowding. They also wished to demonstrate that crowding perceptions could be reduced by allowing subjects to misattribute the cause of their arousal. They reasoned that if subjects were provided with an alternative explanation for arousal, un related to crowding, they wou ld be less l ikely to 50 attribute the stress caused by violations of personal space to crowding . They had groups of five same sexed subjects sit in a circle. Subjects were seated at either a close or a f a r interpersonal distance and were informed that the aim of the study was to test the effects of subl iminal stimul i on g roup performance. A "transmitter" was placed in the corner of the room. I n the Arousing condition , subjects were told that a subliminal noise wou ld be p layed while they worked. They were informed that the noise wou ld be inaudible, but previous studies had shown that it had often caused i ndividuals to experience some discomfort and distress. Identical instructions were issued to the subjects in the Relaxing condition , with the exception that they were told the noise would have a calming and relaxing effect on them. A No - Explanation condition was also run i n which subjects were not told of the subliminal noise. Fol lowing these i nstructions , subjects worked on a series of tasks similar to those used by Worchel and Teddlie ( 1 976). Worchel and Yohai 's ( 1 979) results were consistent with their hypotheses (even though they are open to alternative interpretations). Worchel and Yohai interpret thei r results as supporti ng the attributional crowding hypothesis ; that is , the contention that crowding is experienced through an attributional process, whereby the individual becomes aroused by spatial vio lation and searches the environment to determine the cause of arousal. When personal space was violated (as in the close condition) the experience of crowding was significantly reduced by providing an alternative explanation for the arousal. These resu lts suggest that the subjects in the arousing condition attributed the cause of thei r 51 arousal to the subliminal noise and, therefore , felt less crowded. Worchel and Brown ( 1 984) showed that misattribution may result from a "pool i ng" of arousal. They had subjects watching arousing or nonarousing fi lms under conditions of close or far interpersonal spacing. For subjects at close distances those who saw arousing fi lms reported less crowding than those who saw nonarousing fi lms. Worchel and Brown interpret this to mean that the spatial ly induced arousal was attributed to the fi lms and thus the experience of crowding was reduced. Again the resu lts are open to alternative interpretations. It could be, for example, that showing fi lms of explicit sexual acts or the severe beat ing of a boxer were simply more i nvolving or demanding of viewers attention than a fi lm about ecological evo lution. Such an i nvo lvement or attention hypothesis wou ld suggest that viewers seeing "arousing" fi lms were less concerned about those seated next to them. This possibi l ity wou ld not rely on arousal mechanisms or attributional processes, neither of which were actually demonstrated by Worchel and B rown. Aie l lo et al . ( 1 983) report a study in which subjects in conditions of h igh spatial density were exposed to humour and unlike the Worchel and Brown ( 1 984) study found that these subjects reported feeling more crowded and confined. Accord ing to the two-factor model the humour shou ld have provided the plausible alternative explanation for arousal and served to reduce perceptions of crowding. Accord ing ly, this study provides no support for the two-factor mode l . 52 There is mixed support for the ro le of attributions in the process of crowding. In a nutshel l , the two factor model suggests that even when there exists sufficient spatial conditions for crowding, the attributions which individuals make about the ir state of arousal wi l l determine whether or not they feel crowded (Worchel & Brown , 1 984). One criticism of the two factor model is that it overemphasises the role of cognitive processes. It seems possible, for example, for individuals to fee l crowded because they observe themselves to be part of a crowd. Such a possibi lity places the experience of crowding with in the context of a crowd, without necessari ly re lying on either arousal or cognitive processes which may or may not operate. Attention One of the implications derived from the two factor mode l of crowding concerns the apparent plasticity of att ributions. Worchel and Teddlie (1 976) suggest that these attributions may be moulded by either redi recting the individual 's focus, o r by providing alternative "explanations" for arousal . A key assumption underlyi ng this suggested plasticity is that attention di rects the focus of attributions. The effect of crowding on attentional processes is the basis of the overload theory of crowding (Mi lgram, 1 970) . Other researchers have also documented the role of attention. Baum and Greenberg (1 975) found withdrawal and decreased attention to others among subjects who were anticipat ing a crowded situation. Hel ler, G roff, and Soloman ( 1 977) suggest that in high density 53 conditions i ndividuals have greater demands p laced on their attention and information processing capacity. Specifically, they suggest an increased need to monitor the activities of others. From such increased attentional demands, they predict decrements in other cognitive activities. Cohen and Spacapan ( 1 978) discuss an attentional i nterpretation of the aftereffects of stress. They postulate a cognitive fatigue explanation to account for the finding that crowded subjects were subsequently less l ike ly to help than noncrowded subjects. Evans ( 1 979) discusses the relationship between arousal , attention and task performance. He explains the effect of overarousal o n complex tasks as resulting from high arousal causing a focusing of attention. Attention is focused such that it is al located only to the more important information cues in the task. Complex tasks involve a greater number of relevant cues than simple tasks , and under conditions of high arousal it is more l ike ly that one of the salient cues wi l l be ignored thus creating a decrement in task performance. In simple tasks the focusing on more important information is less l ike ly to result in the lack of attention to salient cues, s ince it is l ike ly that there wi l l be less of these. The two factor model of crowding similarly d raws on attentional processes. Worchel and Yohai ( 1 979) suggest that fol lowing the attribution of arousal to the closeness and presence of others (crowding) , attention wil l be focused on alleviating this presumably aversive state. This attentional demand wi l l result i n both performance decrements and interpersonal hosti lity. 54 In sum , the two factor model assumes the plasticity of attributions to be di rected by the focus of attention . Crowding is assumed to result in attentional demands, which may induce some negative consequences and may further set the stage for attributional manipulation. Information Langer and Saegert ( 1 977) reduced the stress associated with crowding by presenting pre-exposure information which created the expectation of arousal. This i nformation ameliorated the aversiveness of crowding for those who were exposed to it. The information one has prior to the experience of crowding can exert a powerful effect on the actual experience. In Worchel and Yohai's (1 979) study, informing subjects that environmental facto rs wou ld arouse them did not reduce thei r experience of crowding when they suffered no personal space invasion. In fact, this information tended to make subjects fee l somewhat more crowded. However, when spatial violations were invo lved , th is alte rnative explanation for arousal significantly reduced the experience of crowding . Informational manipulation was also used to good effect in Fisher and 8aum's (1 980) study. In the context of anticipatory crowding they assessed the effects of two control relevant messages with a different informational focus. The foci were either situational ("othe rs wi l l be close and may bump into you") o r emotional, 55 ("you may be uncomfortable because others wi l l be close to you"). They predicted situational ly focused i nformation wou ld be more effective in reduci ng the impact of high density because of the salience of external threats in many crowded settings. This prediction was confirmed, as was the second hypothesis which predicted that emotional ly focused information wou ld el icit reduced label l ing of the context as crowded because emotional ly based messages were expected to shift the locus of attribution for discomfort away from the external envi ronment. Situationally re levant information led to an environmental ly based attribution for the discomfort (crowding). Wener and Kaminoff ( 1 983) manipu lated information by way of signs in a fie ld setting. Providing accurate information resu lted i n sign ificantly reduced perceptions of crowding , discomfort, confusion and anger. In sum, the evidence reviewed suggests that information has an important role to p lay in the experience of crowding. Whi le Worchel and Yohai (1 979) provided bogus information about an alternative source of potential arousal with the effect of reducing perceptions of crowding , other studies have similarly reduced the negative aspects of crowding by the provision of accurate information about the nature of the crowding experience. This high lights the possibil ity that Worchel and Yohai may have uti l ised an independent variable which they i nterpret within a particu lar framework, whi le the same manipulation may be open to quite a different interpretation. I n other words the informational studies provide evidence to suggest that subjects respond directly to information , without the need to 56 i nvoke attributional interpretations. Salience Fiske and Taylor ( 1 984) define salience as a stimulus property that attracts an observer's attention. Pryor and Kriss ( 1 977) also link sal ience to attention in not ing that when something is sal ient i t receives a disproportionate amount of attention relative to its context. Fu rther, they relate salience to attribution by proposing that people, or envi ronmental entities that are salient, wi l l receive more causal ascriptions or attributions. Taylor, Crocker, Fiske, Sprinzen , and Winkler ( 1 979) suggest that salient social stimu li also have a disproportionately large impact on the judgement process (even though such stimul i may be logically uninformative or even i rrelevant) . They suggest that , generally, the form of the salience effect is that people attribute causality to the stimul i that engulf their attention . Taylor and Fiske ( 1 978) discuss salience as a "top of the head" phenomenon. They suggest people do not view all the evidence that may bear on a particular problem. Instead they frequently use the i nformation which is most salient o r readi ly avai lable to them, that is , that which is most easily brought to mind. They also maintain that salience of i nformation has a di rect bearing on causal attri butions . . . . . "we believe that the causal attributions people make . . . . are often shaped by seemingly trivial but salient information" (p.253). This view is later summarised (Fiske & Taylor, 1 984) by their conclusion that , generally speaking , 57 causal attributions fol low the focus of attention. PERFORMANCE DEFICITS AND INTERPERSONAL HOSTILITY Worchel and Yohai ( 1 979) hypothesise that the experience of crowding wi l l result in performance deficits on complex tasks and also increased interpersonal hosti lity. Fol lowing the attribution of arousal source to the closeness of others (crowding) , attention wil l be focused on al leviating this negative state. This attentional demand wi l l be disruptive to task performance because the individual wi l l mobi l ise efforts to reduce the negative state. Interpersonal re lations are predicted to suffer sim i larly because the focus on reducing the experience of crowding increases the sal ience of others in the environment. These others are thus seen as d i rectly causing the individual to feel crowded and, i n some cases , frustrating attempts to reduce crowding. Worchel ( 1 978) cites evidence (Evans, 1 975; McClelland, 1 974) which suggests that crowding does i nterfere with the performance of complex tasks, whereas it has little deleterious effect on more simple tasks. This generalised view of crowding-induced decrements on complex but not simple tasks is endorsed by Walden and Forsyth ( 1 981 ) . Worchel ( 1 978) explains the reasoning behind the prediction of increased i nterpersonal hosti lity as being a resu lt of others frustrating individual attempts to 58 restore personal space. This frustration is thought to lead to an increase in hosti lity and possibly aggression . Support for this prediction is mixed , with both supporting evidence (Hutt & Vaizey, 1 966; Evans, 1 979 ; Worchel & Yohai , 1 979) and nonsupporting evidence (Loo, 1 972; Stokols, Rai l , P"inner, & Schopler, 1 973 ; Aiel lo, DeRisi , Epstein , & Karl in , 1 977). This chapter has examined detai ls, background and research evidence re levant to the elements of the two factor model of crowding. The notion that spatial invasion is involved in the experience of crowding has received considerable support in the literature. The two factor model holds that spatial invasion wi l l result in increased arousal , and the literature provides some support for that relationship. The model maintains that accompanying arousal is the need to explain the source of arousal. This assumption has been criticised. Explanations of arousal wi l l be sought via the attributional search , with the possibi l ity of misattribution high lighted by the model. The misattribution of arousal has been questioned as being open to the more parsimonious interpretation that subjects respond di rectly to the information provided. This interpretation wi l l be developed more fully in the next section which wi l l discuss the rationale for Study One. RATIONALE FOR STUDY ONE Worchel and Yohai's ( 1 979) attributional model suggests the arousal which fo l lows a spatial invasion wi l l be misattributed to an alternative source of arousal (provided a plausible alternative is present) rather than being correctly attributed 59 to the actual source. Further, it is suggested that in the absence of such an alternative , performance deficits and i nterpersonal hosti lity wi l l be apparent. These are regarded as the negative aspects of experimental crowding. One un ique aspect of the model is the role of attributions, specifically, the apparent ease with which misattribution may be induced. The model predicts that given a plausible , alternative source of arousal , subjects wi l l make the attribution to that sou rce rather than to the fact of being exposed to a spatial invasion. The basis for this prediction is open to criticism. I n the Worchel and Yohai (1 979) study subjects in the NOISE AROUSING condit ion reported lower perceptions of crowding when compared with two other conditions, a no information (CONTROL) condition and a NOISE RELAXING condition . The criticism stems from the fact that subjects in the noise relaxing condition were l ikely provided with an explanation which was high ly implausible ! The implausibi lity of this explanation comes about because subjects i n the CLOSE (crowded) condition were , in all l ikelihood, aroused yet the "plausible alternative" with which they were provided informed them they should be fee ling relaxed. The point being that this information is discordant with the reality of their experience, and is therefore not a p lausible alternative to which they could reasonably attribute thei r arousal . In its simplest form this criticism implies that Worchel and Yohai 's subjects may have opted for the bogus noise alternative because it was the on ly salient explanation for their arousal. Subjects were made acutely aware that the noise might influence their arousal level while they remained unaware that a spatial invasion cou ld have the same effect. The specific mechanisms through which an alternative explanation for one's arousal become established remain unclear. One question which arises from such consideration is: Why do subjects opt for a plausible alternative to which they may attribute arousal, rather than the actual source? The model predicts that subjects wi l l attribute their arousal to plausible alternative sources (as if i t is safe to assume that subjects wi l l be bl ind to the actual source) , but offers no explanation as to w hy this should occur. Theoretically, this question is of interest since the answer is likely to have a di rect bearing on the basic tenets of the model . 60 One reason why misattribution may occur stems from the. fact that people are typically u naware of their spatial behaviour (Holahan , 1 982) . Pau lus, Annis, Seta, Schkade, and Mathews (1 976) in discussing the lack of consistent effect of density on affect, simi larly raise the possibil ity that density may result in psychological effects "without involving subjective awareness". This impl ies that spatial behaviour is a subtle aspect of interpersonal relationships, with people being unaware of the i nfluence of spatial factors. This point was not lost on the "father" of proxemics. Hall 's ( 1 959 , 1 966) book tit les, (The silen t l an guage, and T he hi dden dimen sion) reflect both the obvious publ ic elements of spatial behaviour and also the more private concealed aspects. If it is the case that people are by and large unaware of the importance of spatial factors o n behaviou r then it seems reasonable to assume that when an 61 alternative source of arousal is present, this wi l l take precedence over spatial sources as these are typically unavailable to subjects' awareness. The view being expressed here is that alternative sources of arousal are only accepted because they are made more sal ient than the actual aro·usal source. Another way of expressing this point is to suggest that it is not the p lausibi l ity of the alternative that has been examined, but rather the sal ience of the alternative. In brief, misattribution results from a process which has the focus of attention being di rected towards the alternative , and therefore away from the actual source of arousal. The finding that misattribution resu lts from the manipulation of the salience of possible sou rces of arousal wou ld represent a significant contribution to the literature. One aim, therefore , of the first study was to examine the role of salience of information about arousal sources. This contribution wou ld add to our understanding of the experience of crowding itse lf. Such knowledge wou ld also serve as a theoretical contribution i n helping to answer a question which is currently unanswered by the two factor model . Finally, i nformation of th is nature wou ld make a useful contri bution to the existing l iterature on causal attribution. To examine the effects of the salience of arousal sources, Study One employed fou r information conditions (CONTROL, NOISE, DISTANCE, NOISE + DISTANCE). The fi rst two conditions, control and noise salient, served to repl icate the two essential conditions of the Worche l and Yohai ( 1 979) study. In accord with the replication , i t was predicted that crowded subjects from the noise 62 sal ient condition wou ld fee l less crowded than their counterparts from the contro l condition. In order to assess more closely the need for an alternative sou rce of arousal the distance condition was run. The i nterest here was to look at the effect of provid ing subjects with a possible source of arousal which was n ot an alternative . The distance information was intended to make salient the spatial factors which were operati ng in the situation. The mode l suggests on ly that plausib le alternative sources of arousal wil l resu lt in reduction of the perceptions of c rowding. It was hoped to show that a non-alternative source of arousal , p resented plausibly and with equal salience, cou ld also result in reduced perceptions of crowding. Such a demonstration wou ld add support to the notion that it is sal ience rather than "alternative ness" which is important , and also that (g iven the condition of plausibi l ity) the content of informational messages is less important than the form. Support for this suggestion comes from Langer, B lank, and Chanowitz ( 1 978) who showed that subjects are wil li ng to use information provided, as long as it fits an appropriate "script". Langer et al. contend that pseudo-thinking is more the ru le than the exception for practical ly al l verbal and nonverbal behaviour. P rovided the structure (form) of the message is congruent with one's past experience, it may occasion behaviour mindless of relevant detai ls. The poi nt being that subjects in the Worchel and Yohai ( 1 979) study may have processed i nformation about the arousal properties of the "noise" quite mindlessly and uncritically, accepting it at face value. 63 Another aim of Study One was to examine the effect of reducing the salience of the "plausib le alternative". This was done by including a dual sal ience condition with two information foci , (NOISE + DISTANCE). Thus , subjects were informed that both bogus noise and the spatial factors operating cou ld result in arousal. This condition al lowed the opportunity to examine whether providing subjects with two salient information sources wou ld produce an additive effect. If this were so, then providing both sou rces of information should result in less perceptions of crowding than either source of information alone. 64 C HAPTE R FOU R METHOD OVERVIEW AND DESIGN The experiment was a 2 (distance) x 4 ( information) factorial design . Subjects were run in g roups of five , with al location to groups being on the basis of avai labi lity. Once constituted, g roups were then randomly assigned to one of the eight conditions. SUBJECTS The subjects were 320 female vo lunteers solicited from introductory level university papers (Psychology, Education , Nursing and Business). Female subjects were chosen because they were avai lable in g reater numbers and also because they wou ld provide a more rigorous test of the model. Subjects were run i n g roups of five , with group members being previously unacquainted. The nonacquaintance of subjects was previously used by Worchel and Yohai ( 1 979) . Other research (Cohen , Siaden , & Bennett, 1 975 ; Rotton, 1 987) has suggested that nonacquainted subjects are l ike ly to experience more crowding than subjects who are friendly. 65 G roups of five were used as this replicated the size of the g roups employed by Worchel and Yohai ( 1 979) . With in the experimental crowding literature the size of g roups has varied between three (Burger, Oakman , & Bullard , 1 983) and 1 0 (Nagar & Pandey , 1 987) . Numerous studies have used groups of fou r (e.g . , Karl in & Epstein , 1 979) . . A number of studies have had groups of five (e.g . , Klein & Harris , 1 979) , whi le the most popular number employed in small g roup crowding research is six (for example, Walden & Forsyth, 1 981 ) . The students were approached i n class where they were informed of the requirements of the study, and in l ine with Tesch's ( 1 977) recommendation they were also advised of the possible benefits of participation. Specifically they were informed that the study was looki ng at g roup performance and would take approximately 45 minutes. They were also told that the study wou ld provide them with the opportunity to experience the role of the research subject and they would see how laboratory experiments are conducted with human subjects. Those volunteering completed a consent form on which they supplied their names and telephone numbers with the understanding that they wou ld be contacted at a later date to arrange a suitable time to complete the experimental procedure. The volunteers were also told to expect to receive by post a summary of the results on the completion of the study. EXPERIMENTAL MATERIALS The experimen tal room was a small (2 .75 m x 2.47 m x 2 .42 m high) internal 66 room , without windows. On the door was a sign , "Small Group Research Lab". The room was furnished with five armless chai rs, arranged in a closed circle in the center of the room . In the center of the circle and suspended from the cei l ing was a microphone. The microphone was suspended 63 em below the cei l ing and was clearly visible when subjects entered the room. In one corner of the room was a large speaker clearly label led "NOISE TRANSMITTER". On an adjacent shelf was a noise generator labe lled "NOISE GENERATOR". A small tape recorder was also positioned on the shelf. The experimen tal pre- in struction s (see Appendix A) were pre-recorded and for each group one of fou r sets of pre-recorded instructions were used. The four sets of pre-instructions corresponded to the fou r information condit ions which were NO EXPLANATION, DISTANCE EXPLANATI ON, NOISE EXPLA N ATION, and D U AL (distance plus noise) EXPLANATI ON . The no expl an ation condition served as a contro l condition . The subjects were thanked for volunteeri ng for the study which , they were told, was to examine group performance and g roup interaction. Subjects in the distan ce expl an ation condition were told the same, but were also to ld that the study wou ld be examining the effects of i nterpersonal distance on behaviour. They were informed that the distances involved may cause them to feel stressed and uncomfortable. The n oise expl an ation condition was the same as the no explanation condition . Subjects were thanked and told that the study was examining group performance and interaction. Additionally, subjects in the noise explanation condition were informed that the study was looking at the effects of sublimi nal noise on behaviour. Subjects were 67 led to expect an inaudible noise and were informed that it may cause them to feel stressed and uncomfortable. Fi nal ly, the dual exp lan at ion instructions began in the same manner as the no explanation condition , but in addition subjects were i nformated that the study was interested in examining the effects of both interpersonal distance and subliminal noise on behaviour. Five i dent ificat ion ba dges ( label led "A" to "E") were issued, one to each of the subjects. These identification badges were used to help the subjects identify individuals later when they were completing the ratings included i n the post­ experimental questionnai re. A group word task, identical to that used by Worchel and Yohai ( 1 979), involving the extraction of words from a master word was used as a measure of g roup performance. Subjects were required to derive as many words as possible from the master word, OBSERVATIONALLY. A group discu ssion/decision task, Johnny Rocco (Ward, 1 970) and the accompanying decision scale (Love-Punishment scale) , was used. This task was the same as employed by Worchel and Yohai ( 1 979) . The task i nvolved reading a case study about a young offender, Johnny, and then making a decision about the most preferred manner of treating Johnny. The post-experimental qu est ionna ire consisted of a 20 item questionnaire (see Appendix B) . These questions were based directly on the questionnaire used by 68 Worchel and Yohai ( 1 979) , which examined reactions to the experiment, the room and fel low group members. Responses to these questions formed the key dependent measures of arousal and crowding. Affect measuremen t employed the Multiple Affect Adjective Check List (MAACL: Zuckerman , Lubin , & Robins, 1 965) . The MAACL elicits self report ratings of Depression , Anxiety and Hosti lity. It comprises 1 32 adjectives which are arranged alphabetical ly . Of these , 21 items assess Anxiety, 40 assess Depression and 25 items assess Hosti lity. There are 23 buffer items which are unscored. The Today form of the MAACL was used as this i nstructs subjects to respond in terms of their immediate feelings. In the review by Kelly (see Buros, 1 972), the re liabi l ity of the MAACL (odd-even and plus-minus) i s reported as rang ing between . 1 7 and .92, with the median being . 72. The checklist therefore possesses reasonable i nternal consistency. On a more cautionary note, Kelly (see Buros, 1 972) is critical of the fact that i n spite of no item bei ng scored on more than one scale, the intercorrelations among the three scales are very high . The reported i ntercorrelations are i n fact as h igh as their rel iabi lities. The high i ntercorrelations among the three affect scores may reflect true i ntercorrelations between anxiety, hosti l ity and depreSSion, or may reflect a lack of discriminant validity. 69 In considering the validity of the instrument, Megargee (see Buras, 1 972) notes that much of the research dealing with the validity of the MAACL has consisted of administering the instrument under conditions l ike ly to e licit the relevant affective states and comparing the scores with those obtained under normal conditions. The MAACL has been used in a number of settings. Students have been tested before and after examinations, mi litary personnel have been tested duri ng basic training and actors have been tested prior to going on stage. Hypnotically i nduced mood states, sensitivity training , stress interviews, tranqu i l l ising medication and induced relaxation have been used to alter mood, and concomitant changes i n MAACL scores have been noted. Megargee concluded that , by and large , the results of these studies had been positive . Kelly ( in Buras, 1 972) also highl ights the problem of acquiescent response set. Kelly cites a study by Si l ler and Chipman ( 1 965) which reported correlations of .05 to .48 between MAACL anxiety scores and various measures of acquiescence . Of even more concern , according to Megargee ( in Buros, 1 972) , is Herran's (1 969) report that a set to respond with too many or too few adjectives contributes a significant proportion of the variance on the MAACL. The MAACL has previously been used in crowding research. Zeedyk-Ryan and Smith ( 1 983) , for example, found increased hosti l ity i n subjects who experienced increased social density. 70 PROCEDURE Due to the replicative aspect of the current study and the desi rabi l ity of avoiding confounds, the procedure for this study was mode lled as closely as possible on Worchel and Yohai ( 1 979) . Subjects were telephoned to arrange a convenient time. This time had to be suitable to the other members of the group. For this reason the al location of subjects was not random but rather more haphazard . This arrangement was often made several days in advance of the actual meeting time. In order to increase attendance, all subjects were telephoned on the eve of the appointed t ime to remind them of the time of the meeting. This precaution was taken to avoid on ly four of the five subjects attending. Where this did happen the fou r subjects were dismissed and another t ime was rescheduled. On arrival , al l subjects were met by the experimenter and introduced to the female experimental assistant. At th is point the procedure was conducted solely by the assistant who was blind to the experimental hypotheses. This precaution was taken to avoid the possibi l ity of experimenter b ias. Subjects were shown i nto the experimental room and each given an identification badge to pin on . The experimental assistant had previously been provided with a prearranged randomly al located schedule of experimental conditions. These conditions applied to the two i ndependent variables, D ISTANCE and INFORMATION. The 71 distance variable was either cl ose or f a r, while the information variable was operational ised by reference to one of the four prerecorded sets of taped instructions. In the f a r condition the front legs of adjacent chairs were spaced 51 cm apart. In the close condition the front legs of each chair were touch ing those of the adjacent chairs. In this close condition the assistant d rew back one of the chai rs to al low fou r of the subjects to be seated, and once the fifth subject was seated her chair was moved in to complete the circle. The knees of subjects in the close condition were thus touchi ng the knees of the subjects either side of them . Once seated the fol lowing instructions were read to al l subjects: "Do not move your chai rs because the session is being recorded, and the chairs have been careful ly placed to ensure optimum recording quality. This is i mportant because later we wi l l be analYSing voice levels." Fol lowing the request to retai n the seating positions, subjects were played one of the fou r sets of pre-instruction information (Appendix A). The fou r information conditions were NO EXPLANATION, D ISTANCE EXPLANATION, NOISE EXPLANATI O N , and DUAL (distance plus noise) EXPLANATION. Fol lowing these pre-instructions subjects were asked to begin the fi rst task, which was to derive as many smaller words as possible from the master word (OBSERVATIONALLY). Subjects were asked to work together to produce a 72 sing le list, and the task was l imited to 1 0 minutes. One subject acted as recorder and was provided with a clipboard, paper and pen. After checking to ensure that all subjects understood the instructions, the assistant left the room return ing 1 0 minutes later. On returning , the experimental assistant co l lected the word list and introduced the next task. This task involved both individual and group decisions concerning the best treatment for a juveni le (Ward , 1 970). Subjects were each given a case history and a copy of the love/punishment scale. This scale lists seven alternative treatments which range from extending love and understanding at one extreme to a punitive option at the other extreme. Subjects were asked to read the case history and then indicate which of the avai lable options they favoured by circling one of the numbers corresponding to each option . Following the col lection of the individual decisions, subjects were asked to discuss the case as a g roup for five m inutes and arrive at a g roup consensus. The experimental assistant then left the room for a period of five minutes, whereupon she returned and recorded the group decision. The final task in the experimental room was to complete a copy of the MAACL. Fo l lowing this the subjects were taken to an adjacent room containing desks and chairs and asked to complete the 20 item post-experimental questionnaire. On completion of the post-experimental questionnaire , the experimental assistant thanked the subjects and left the room , to be replaced by the experimenter who 73 then debriefed the subjects and answered any questions. Al l subjects were then thanked for their participation , and asked to respect the confidentiality which the study requ i red. Many of the research subjects were living in the same hoste ls and th is precaution therefore seemed necessary. ETH ICAL CONSIDERATIONS AND DEBRIEFING The pri ncipal deception in th is study was that subjects were not informed that they would be seated (as half of them were) at an unusually close distance to others . This deception amounted to tel l ing subjects that the study was i nterested i n g roup performance when in fact the key interest was in g roup performance under specific envi ronmental conditions. This deception was necessary because to provide complete i nformed consent would have seriously compromised the i nternal validity of the study. Thus, partial informed consent was employed, with the p rovision that any obvious expression of discomfort displayed by subjects was to be immediate ly attended to by the experimental assistant. I n extreme cases it was thought that the group might have to be disbanded if a subject was i n any way distressed by the closeness involved i n the experimental procedure. In practice this eventuality never occurred. In add ition , subjects were fully debriefed. Debriefing procedures may be considered (e.g. , Tesch, 1 977) as little more than an article of professional faith. The extent to which such procedures may vi lify the deceptions, or function 74 remedially to somehow undo any i l l effects which may have resulted, are questionable. However, the procedures employed shou ld address the issues of ethics, methodology and education. Accordingly, the debriefings considered the issue of deception . Its use was explained and justified. All aspects of the methodology were also considered and fi nal ly the key aims of the research were explained. Subjects were encouraged to raise any concerns or ask any questions. They were also asked, in the case of subjects in the close condition , about the extent to which they felt uncomfortable in the situation with the i r knees touching the knees of two other strangers for the duration of the experiment. Final ly, the subjects were informed that at the conclusion of the experiment and when the data had been analysed they wou ld be sent a copy of these results i n summary form (see Appendix C). This was designed to maximise the educational value of participation i n this research. Fol lowing this, subjects were thanked and dismissed. UN IT OF ANALYSIS One issue which has a beari ng on the resu lts of any crowding research is that of unit of analysis. In studying the phenomenon of crowding , is i t best to examine the i ndividual's experience within the crowd, or i s i t better to examine the effect of the experience on the g roup as a who le? In short, should the i ndividual o r the g roup be the unit of analysis? If the g roup is used as the unit of analysis this i nvo lves summing the data from the individual members of the group and dividing by the number of g roup members. Thus, if the group is the unit of analysis the data to be analysed are a smaller number of group means. An individual analysis simply invo lves using each individual g roup member's data. 75 One of the problems i n studying crowding is that such research is very subject demanding , especially when the group is employed as the unit of analysis. As previously mentioned, there are difficulties in recruiting and schedul ing g roups of subjects , particularly if large numbers of subjects are required. One way of overcoming this wou ld be to run a small number of groups and use the individual subject as the unit of analysis. The efficiency afforded by individual analysis has previously been employed (Aiel lo et aI. , 1 977 ; Epstei n & Karlin, 1 975) , but on ly after tests for the equality of groups within treatments had been performed and yie lded no significant differences. Such tests provide some support for the view that the assumption of independence had not been defied. Also , i n these studies, data were col lected i ndividually after leaving the experimental room ; thus, there was no interaction during measurement. Jain ( 1 987) included only one group (comprised of ten members) in each of four experimental conditions and analysed results on the basis of the individual as the unit of analysis . Subjects were instructed not to talk to each other, thus reducing the chance of i nteraction and presumably also attempting to preserve their independence. Zeedyk-Ryan and Smith ( 1 983) conducted an individual analysis without taking either of the above precautions. 76 A more conservative, and i ndeed more popu lar, approach is to use the g roup as the un it of analysis. Most experimental studies of crowding employ th is approach. Epstein and Baum (1 978) maintain that whenever the possibi l ity of g roup members i nfluencing one another exists the group should be the primary un it of analysis. The g roup as the unit of analysis was used i n the current study as, u n l ike the studies which have used an individual analysis, subjects were requ i red to engage in i nteractive tasks which i ncreased the possibi lity of non­ i ndependence between individuals. It would have seemed unreasonable to have subjects engage in dependent activity , and then to have treated the data as if the assumption of independence had not been defied. I n addition, the previous research by Worche l and Yohai ( 1 979) used the group as the unit of analysis. The possibi l ity of employing noninteractive tasks was also considered , but d ismissed on the grounds that this wou ld have served to enhance the artificiality of the situation. 77 C HAPTE R FIVE RESU LTS For each of the key dependent variables a 4 (information salience) x 2 (distance) analysis of variance was calculated. Planned comparisons (Rosenthal & Rosnow, 1 985) were used to examine specific predictions. Data were analysed using the Statistical Package for the Social Sciences (Norusis, 1 985). Decisions concerning statistical significance were guided by the convention of p< .05. Summary ANOVA tables are included in Appendix D. AROUSAL The two factor theory of crowding predicts that crowding wi l l result in subjects being more aroused than their noncrowded counterparts. It was therefore predicted that crowded subjects would manifest higher leve ls of arousal than noncrowded subjects. Worchel and Yohai ( 1 979) measured arousal as being subjects' responses to the question of how "relaxed or i l l-at-ease" they felt. The cu rrent study used the same measure. The mean ratings for arousal are provided in Table 1 . Crowded subjects (those in the close condition) were not significantly more aroused than their uncrowded counterparts, F (1 ,56) = 1 .9 1 , ns. Surprisingly, this 78 study was unable to show that crowded subjects were any more aroused than uncrowded subjects. TABLE 1 MEANS AND STANDARD DEVIATIONS FOR RATINGS OF AROUSAL Distance Information Salience Contro l Noise Distance Noise & Distance Close 3.90 4.58 3.65 4. 1 3 1 .67 0 .83 1 .70 1 .21 Far 4. 1 8 4. 1 8 3.25 3.58 0.64 0.92 1 .04 0.68 Note : 1 = Relaxed , 1 0 = I I I at ease. N = 8 for al l cel ls. CROWDING To assess perceptions of crowding, subjects were asked both how crowded and how confined they felt. Again , this parallels the measures used by Worchel and Yohai ( 1 979). Table 2 clearly shows that subjects seated at the close interpersonal distance felt more crowded than those subjects seated at the far interpersonal distance. There was a highly significant main effect for the interpersonal distance manipulation, F ( 1 ,56) = 65. 1 5 , P < .001 . (Note that where "p = " is used this result has come from the Anova table, Appendix 0, and where 79 "p <" is used this signifi es that the result of a specific comparison has been extracted from an appropriate table of values. The one exception is when the p value is taken from the Anova table and reads "p = .000" . I n this instance the resu lt is changed to read "p < .001 " ) . The means collapsed across i nformation salie nce conditions were (Close = 6 .32 , Far = 3.43) . TABLE 2 MEANS AND STANDARD DEVIATIONS FOR RATINGS OF CROWDING Distance I nformation Salience Control Noise Distance Close 7 . 1 3 6.65 5.75 1 .35 1 . 1 9 2 .01 Far 2.95 3.90 3.98 0.67 0.56 1 .24 Note : 1 = Not at a l l crowded, 1 0 = Very crowded Noise & Distance 5 .75 2.04 2.90 0 .74 The fol lowing resu lts consider on ly subjects from the close conditions as the remain ing predictions all concern the relative effectiveness of various forms of i nformation in reducing the impact of crowding. These predictions do not, therefore, apply to the noncrowded subjects. It was predicted on the basis of the two factor model that the noise salient i nformation condition wou ld report less crowding than the control condition. This 80 prediction was examined by a planned comparison between these two conditions, but fai led to reach significance , F < 1 . One of the principal predictions of th is study was that making the effects of distance manipu lations known to subjects wou ld have the effect of reducing the extent to Which they felt crowded. A p lanned comparison examined the difference between the contro l and distance salient i nformation condit ions. Subjects in the close distance sal ient i nformat ion condition reported fee l ing less crowded (mean = 5.75) than subjects i n the close contro l condition (mean = 7. 1 3). This difference was in the expected direction but fai led to reach significance , F ( 1 ,56) = 3.69, P < . 1 0 . The expectation that distance salient information , being more plausible than the (bogus) noise salient i nformation , wou ld resu lt i n less crowding was unsupported F (1 ,56) = 1 .58, ns. Final ly, it can be seen from Table 2 that combin ing both noise and distance i nformation failed to decrease the experience of crowding compared to either condition alone. TABLE 3 MEANS AND STANDARD DEVIATIONS FOR RATINGS OF CONFINEMENT Distance I nformation Salience Control Noise Distance C lose 6.35 5.83 4.70 1 .25 1 .24 1 .92 Far 4.23 4.48 4. 1 5 0.92 1 .49 1 .23 Note : 1 = Not at al l confined, 1 0 = Very confi ned Noise & Distance 5.68 2.07 3.65 0.82 81 The means for the ratings for confinement are provided in Table 3 . In general , these resu lts parallel those obtained for crowding. Subjects i n the close condit ion (mean = 5.64) reported feel ing s ignificantly more confined than subjects in the far condition (mean = 4. 1 2) . The ANOVA for confi nement showed a significant mai n effect for the distance manipu lation , F ( 1 ,56) = 1 5.74, P < .001 . Considering only the crowded (close) subjects, the re lative reduction i n confinement for the noise salient condition when compared to the contro l condit ion fai led to reach significance, F < 1 . This resu lt mi rrors the result for crowding. Taken together, these two results high light the fai lure to repl icate one of the fundamental aspects of the Worchel and Yohai ( 1 979) study. However, u n l ike the crowding measure , the relative reduction in confinement experienced by subjects in the distance salient information condition when contrasted with the contro l condition was significant, F (1 ,56) = 4.68, P < .05. Although the distance salient i nformation condition did result in lower ratings of confinement than any other condition , the difference between this and the noise salient i nformation condition was not sign ificant, F ( 1 ,56) = 2. 1 8, ns. As with the crowding measure , the dual salience (crowding and noise) i nformation condition fai led to produce any significant differences when contrasted with the other i nformation conditions. PERFORMANCE The performance measure was the same as that used by Worchel and Yohai 82 ( 1 979) , that is, the number of words derived from the master word. The two factor theory clearly predicts performance deficits for subjects in crowded conditions. Contrary to this prediction , subjects i n the close conditions performed no differently ( mean = 98.41 words) when compared to subjects in the far conditions (mean = 97.31 words). Means for the measure of performance are provided in Table 4. Distance Close Far TABLE 4 MEANS AND STANDARD DEVIATIONS FOR THE NUMBER OF WORDS DERIVED FROM THE MASTER WORD Information Salience Contro l Noise Distance Distance & Noise 1 02 96 1 04 92 22.98 1 6.53 8 .61 1 1 . 1 1 1 03 85 98 1 03 8.65 1 2. 1 2 1 1 .48 1 6.02 Note : means rounded to the nearest number of words. Information provided to subjects in the noise conditio n seemed to have an effect on the performance of these subjects when compared to those subjects i n both the control and distance conditions. The comparisons supported this with noise condition subjects showing a performance deficit when compared to contro l condition subjects, F ( 1 ,56) = 4.96, P < .05. This performance deficit also was 83 apparent when noise condition subjects were compared to the distance condition subjects , F (1 ,56) = 4.07, p< .05. INTERPERSONAL RELATIONS The two factor model predicts that the experience of crowding wi l l resu lt i n increased hosti l ity . Five separate dependent measures were investigated i n order to test this prediction : punitiveness (Johnny Rocco), ratings of friendl iness and agg ression toward fe l low group members, the extent to which other g roup members were l iked , and MAACL hosti lity scores. Surprisingly, subjects who were crowded (close condition) were no more punitive (mean = 2.80) in thei r treatment of Johnny Rocco than were uncrowded (far condition) subjects (mean = 2.60). Although in the expected di rection , the mai n effect for distance fai led to reach significance, F ( 1 ,56) = 2.38, P = . 1 3. However, the main effect for information was of borderl ine significance, F ( 1 ,56) = 2.69, P = .06. On closer examination , using specific contrasts for the crowded subjects only, a significant difference emerged in the comparison between close distance (mean = 2.55) and close control subjects (mean = 3.00), F ( 1 ,56) = 4.31 , P < .05. Informing subjects about their crowding experience in advance resu lted i n them treating Johnny Rocco less punitively than the control subjects who were not provided with the same information. The results for ratings of friendliness of other subjects also fai led to support the 84 mode l . Crowded subjects rated their fel low group members to be no less friendly (mean = 7.59) than the uncrowded subjects (mean = 7.53). S imi larly , no differences were evident on the measure of perceived agg ressiveness of other group members. Crowded subjects rated their fel low g roup members to be no more aggressive (mean = 6.97) than their uncrowded counterparts (mean = 7.06). The lack of support for the two factor model was further confirmed by a consistent lack of differences between crowded and u ncrowded subjects on the final two measures of interpersonal re lations. On the measure of l iking for fel low g roup members, crowded subjects indicated only marginal ly less l iking for other members of t heir group when compared to uncrowded subjects, F < 1 . The respective means were 7.07 for crowded subjects and 7.26 for their uncrowded counterparts. On the Hosti l i ty scale of the MAACL, crowded subjects were only marg inal ly more hosti le than the uncrowded subjects, but the difference was not significant, F ( 1 ,56) = 1 .6 1 , P = .2 1 . , Col lectively, these five measures have fai led to establ ish any differences in the way in which crowded subjects responded to others when compared to u ncrowded subjects. 85 OTHER MEASURES The effectiveness of the i nterpersonal distance manipu lation was supported by the finding that crowded subjects regarded themselves as less comfortable (mean = 5.03) than uncrowded subjects (mean = 4.28), F ( 1 ,56) = 6.24, p = .02. Not on ly did the interpersonal distance manipu lation result i n subjects feel ing crowded, but they also felt less comfortable (as may be expected if subjects were tru ly crowded). Sim i larly, in the case of the other two affective scales (Depression and Anxiety) crowded subjects scored more highly than uncrowded subjects. Crowded subjects reported increased levels of depression (mean = 1 5.05) when compared to uncrowded subjects (mean = 1 3.56) , F ( 1 ,56) = 5.35, p =.02. The increases on the Anxiety scale of the MAACL also suggested that crowded subjects reported fee ling more anxious (mean = 7 . 1 0) than their uncrowded counterparts (mean = 6.38) . This difference was of only marg inal significance, F ( 1 ,56) = 3.63, p = .06 Finally, there were two results which were of i nterest in terms of the impact of different information on the experience of crowding. Fi rstly , subjects were asked to rate their abi lity to concentrate during the experiment. For crowded subjects, in both conditions where subjects were i nformed about the bogus noise (noise mean = 6.72 , noise + distance mean = 6.60) , abi lity to concentrate was apparently impai red in comparison to subjects who were told the focus of the study was to examine the effect of interpersonal distance on behaviour (distance mean = 7.72) . This specific comparison was significant , F ( 1 ,56) = 4.85, p < . 05. 86 In addition, an overal l main effect was apparent for information, F (3,56) = 3.06, P = . 04, suggesting that this particular effect was more widespread and influenced the uncrowded subjects also. Secondly, subjects were asked to rate the extent to which they felt distracted duri ng the experiment. For crowded subjects those in the noise condition reported being more distracted (mean = 4.22) than either the contro l condition (mean = 3.02) or the distance condition (mean = 3.20) . The difference between these means was significant, F ( 1 ,56) = 7.40, P < .01 . I nforming subjects about the possibi l ity of subl iminal noise in the experimental room increased the extent to which they fe lt distracted while in that envi ronment. THE RELATIONSH IP BETWEEN AROUSAL AND CROWDING Subjects in the close condition reported i ncreased levels of crowding compared to subjects in the far condition . However, the reduced i nterpersonal distance i n the close condition was not accompanied by the predicted increase i n levels of arousal. These results raised questions about the relationship between arousal and crowding. Did this result, for example, mean that crowding was independent of arousal? To answer this question a series of analyses were conducted to c larify this relationship . Firstly, the correlation between these two variables was calculated, r = 0.40, P = . 00 1 . Clearly, the two are not unrelated. Secondly, an ANCOVA was conducted on the variable of crowding with arousal as a co­ variate. This resulted in a significant co-variation between arousal and crowding , 87 F ( 1 ,55) = 26.25, P < .001 . (See Appendix E for ANCOVA summary table) . However, with arousal control led the ANCOVA revealed one additional significant result : an interaction between distance and information, F (3,55) = 3.37, p = . 025. INDIV IDUAL ANALYSIS As mentioned previously, the g roup was the accepted unit of analysis for this study. However, out of interest, some individual analyses were conducted. Only selected dependent variables were included and the summary ANOVA tables for arousal , crowding and confinement are included in Appendix F. The individual analyses boast larger degrees of freedom i n accordance with the larger N's which individual analysis affords. This increased power more than offset the increased variabil ity and resu lted in i ncreased levels of significance. Examples of this can be found in the summary tables provided in Appendix E. Considering the dependent variable of arousal , the effect of the distance manipu lation had a significance leve l of p = . 1 7 with group as the u nit of analysis. This level became p = . 1 0 for the individual analysis. This resu lt is typical of the kind of increase in statistical sensitivity afforded by the i ndividual analysis. However, as previously argued, it was not considered appropriate to complete al l the analyses this way. 88 CHAPTER SIX DISCUSSION Contrary to predictions, crowded subjects did not report significant i ncreases in arousal as a resu lt of the reduced interpersonal distance which they experienced in the experiment. This resu lt is of particu lar interest i n view of the theoretical expectations and the already equivocal results surrounding this issue (Freedman , 1 975; Karl in & Epste in , 1 979; Smith & Knowles , 1 979 ; Webb, Worchel , Riechers , & Wayne, 1 987). The controversy surroundi ng the role of arousal includes a number of diverse views. These views include those (e.g . , Freedman) who have been unable to provide experimental support for the invo lvement of arousal in the experience of crowding , as wel l as those (e.g . , Karl in & Epstein ) who maintain that crowded subjects are almost invariably more aroused than uncrowded subjects. The middle ground in this debate is occupied by Smith and Knowles ( 1 979) whose field study pOinted to the role which cogn itive i nterpretations may occupy in mediating the nature of reactions to spatial invasion. They conclude that when a spatial invasion seems appropriate for the conditions, stress and arousal may not occur. S imply tel l ing subjects that the research was " investigating group behaviour under a variety of conditi ons" may have been sufficient to legitimize spatial vio lation in the present study. The two facto r model is predicated on the notion that arousal is an i nherent e lement of the experience of croWding. Resu lts , such as those derived from the 89 current study, which show increased levels of reported crowding without sign ificant i ncreases i n arousal strike at the very foundation of the two factor theory . These results run contrary to one of the fundamental assumptions on which the two factor theory i s based. The current study - is not alone in its fai lure to document crowding-i nduced arousal. The earl ier work of Freedman ( 1 975) and also the more recent studies by Patterson , Roth , and Schenk (1 979) , and Webb et al . ( 1 987) have also found that crowding may occur without arousal necessarily being i nvo lved. The fai lure of the current study to induce arousal i n crowded subjects stands as an interesting finding which adds to the existing literature on the re latio nship between crowding and arousal. This result should not be interpreted to mean that there is no re lationship between arousal and crowding as there was i n fact a significant positive correlation between the two. I n addition, crowding was seen to covary sign ificantly with arousal when examined with an analysis of covariance. With arousal contro lled, a s ignificant interaction between the i ndependent variables (distance and information) emerged. This may be i nterpreted to mean that f contro l (no information) subjects were more extreme in their ratings of bei ng crowded o r not crowded, but these extremes were reduced for subjects who were provided with i nformation . These results are i nteresting in demonstrating the emergence of th is i nteraction when arousal was control led. This suggests that there is some un rel iabil ity and error variance associated with the measure of arousal. This highl ights the need to employ a more reliable measure of arousal in futu re . 90 A fu rther qualification concerns the issue of measurement. If arousal is not measured in a psychometrically sound fashion , then the validity of resu lts may be threatened. The question being considered is to what extent can we be sure that asking how "re laxed or i l l at ease" people feel is a usefu l measure of their arousal. The answer to this question is, regrettably, that we do not know at this stage. This reflects on the crowding research which, up to the present, has neg lected important psychometric issues. The current study used the same measure of arousal as used by Worchel and Yohai ( 1 979), with considerable care taken to replicate two of the Worchel and Yohai experimental conditions. It shou ld be noted, however, that if the measure of arousal used was un re liable then fai lure to replicate the previous Worchel and Yohai results should not be seen as too surprising. Unre liable measures are l ikely to produce un rel iable results. Another possible explanation for the nonarousal of crowded subjects relates to the sex of subjects. It has previously been demonstrated (Evans & Howard, 1 973) that females tolerate small interaction distances better than males, although Hayduk ( 1 983) reports inconsistent resu lts for the effect of sex on spatial variables. Kalb and Keating ( 1 981 ) , for example, found that females felt more crowded than males i n a field sett ing. This inconsistency is disputed by Nicosia, Hyman , Karl in , Epstein , and Aiel lo ( 1 979) who argue that there are con sistent sex differences. They conclude that when compared with noncrowded counterparts, men generally react more negatively than women when crowded. Reasons for th is include the suggestion of Epstein and Karl in ( 1 975) that the 9 1 experience i s stressful for both men and women, but women are able to share this distress with each other while cu ltural norms prevent men from engaging in such open expression. I n Stokols' ( 1 976) view, the difference reflects the cu ltural emphasis on male territoriality and aggressiveness. Gifford (1 987) s imi larly concludes that i n laboratory studies men respond more negative ly than women. The sexual composition of groups in experimental laboratory studies are perhaps best characterised by the impressive nature of the variety involved. A number of studies have used male subjects on ly (Schaeffer & Patterson , 1 980 ; Nager & Pandey, 1 987). Others have used same sexed groups (McCal lum , Rusbu lt, Hong, Walden , & Schopler, 1 979 ; Nicosia et aL , 1 979 ; Worchel & Yohai , 1 979) , with male and female subjects being run in separate groups. Pau lus and Mathews ( 1 980) employed mixed sex groups, with male and female subjects being included i n the same group. A further strategy has been to leave the sexual composition of the groups "uncontrol led" (Klein & Harris , 1 979 ; Pau lus et aL, 1 976) . In the case of Kuykendall and Keating ( 1 984) the sex of subject was not even mentioned i n the description of group composition . In the Worchel and Yohai (1 979) study, men and women were d ivided i nto same sex groups. They reported no differences due to sex of subject , as was also the case i n the McCal lum et aL , ( 1 979) study. Worchel and Yohai suggest that previous findings where differences between males and females have been apparent may have resu lted from a fai lure to "severe ly" vio late the personal space of the women i nvolved. However, it would be difficult to leve l this criticism 92 at the present study which took care to use the same seating arrangements used by Worchel and Yohai, where subjects' personal spaces were clearly violated. Whi le the sex of subject may be invoked as an explanatory factor in consideri ng the u nexpected nonarousal of crowded subjects i n the present study, it is also clear that further research is required on the nature of sex differences before these issues can be settled. One final consideration regarding the fai lure to induce high levels of arousal in subjects at close i nterpersonal distances stems from the suggested (Morasch et aI. , 1 979) corre lat ion between task fai lure and arousal. They argue that fai lure at an i mportant task wi l l increase arousal and also perceptions of crowding. In the present study, subjects experienced no fai lure and if Morasch et al. are correct shou ld not have experienced arousal. This possibi l ity, however, wou ld not be helpful in explai n ing the increased levels of arousal reported by the subjects in Worchel and Yohai's ( 1 979) study. Despite the fai lure to find a mai n effect for arousal, subjects in the close i nterpersonal distance condition sti l l reported feel ing more crowded and confined than subjects i n the far condition. In addition , these subjects reported feel ing less comfortable, as would be expected in an experimental study designed to make subjects experience crowding. There are two major implications which stem from these resu lts. Fi rstly, the resu lts may be seen as an endorsement of the effectiveness of the 93 crowding manipu lation . Seating subjects at close distance resulted i n subjects reporting more crowding , confinement and discomfort. This suggests that the distance manipu lation worked , and that subjects did i ndeed feel crowded in the experimental setting. Secondly, this finding can be seen as clear evidence again st the two factor model which suggests the experience of crowding fo l lows an attributional sequence which is activated by increased arousal. Nonaroused subjects, if the model is correct, would be u n likely to engage in the attributional search as the activating cue for the search would be absent. In the present case , the motivating force which d rives the attributional search was apparently absent. I n s imple terms, the model suggests that nonaroused subjects shou ld n ot report feel ing crowded. The present study shows that nonaroused subjects did, in fact, report feel ing crowded and confined. This finding has i nteresting implications for the mode l . Specifically, it suggests that it is possible to experience crowding without being aroused. Arousal, therefore , does not appear to be a n ecessary prerequisite for the experience of crowding. This result therefore represents a challenge to the two factor model demonstrating that the experience of crowding may not be arousal dependent, as the two factor model implies. A further qualification is that although the predicted increases i n levels of arousal fai led to reach significance , arousal was positively correlated with the measure of crowding. This may imply that although not significantly aroused, subjects cou ld have been sufficiently aroused to in itiate the attributional sequence. 94 It is important to note at this stage that further examination of the theoretical predictions become academic due to the fai lure to document arousal in the crowded subjects. With the apparent fai lure of the spatial invasions to i nduce feelings of increased arousal, specific predictions based -on the two factor mode l lose much of their meaning . If the attributional process suggested was not operating there would be no reason to expect d ifferences between the i nformation condit ions. With few exceptions, this is precisely what was found for crowding , confi nement, performance and the measures of i nterpersonal relations. The model predicts that subjects exposed to noise salient information wou ld experience reduced crowding. This predict ion was unsupported. Simi larly, the expected difference between crowded contro l subjects and crowded subjects in the distance sal ient condition fai led to reach significance. The prediction that distance salient i nformation , bei ng more plausible than the bogus noise sal ient i nformation , would resu lt i n reduced crowding was also u nsupported. On an equally unspectacu lar note, the dual information condit ion also fai led to decrease the experience of crowding when compared to other conditions. In general , the resu lts for confinement paral lel those obtained for crowding . There was a very clear cut main effect for the distance manipu lation , supporting the effectiveness of the manipu lation and also highl ighting the possibi l ity that crowding (and the comparable feel ing of confi nement) may be experienced without the associated high levels of arousal. However, with on ly one exception the i nformation conditions had no impact on reports of confinement. The 95 exception was that subjects in the distance salient information condition experienced a significant reduction in confi nement compared to crowded contro ls . One i nterpretation of th is is that i nforming subjects about the actual nature of the crowding experience and the purpose of having them seated in that fashion served to reduce the extent to which they subsequently felt confined in the situation. Taken together, the results for confinement and crowding suggest that providing subjects with simple, plausib le i nformation about the situation in which they found themselves had significant effects on behaviour. This i nterpretation is supported by studies which found that accurate situational information reduced crowding­ re lated stress. Baum, Fisher, and So loman (1 981 ) found that for subjects who were not fami l iar with the sett ing al l types of information was of benefit , whereas subjects who were fami liar with the setting derived particular benefit, in terms of reducing crowding related stress, from information which was more accurate. Fisher and Baum ( 1 980) , in an anticipatory crowding context , found that messages were effective in rel ieving discomfort to the extent that they were accurate i n their descriptions of the anticipated setting. I n the same vein , Wener and Kaminoff ( 1 983) manipulated accurate i nformation by way of signs posted i n a crowded lobby with consequent reductions in users' perceptions o f crowding , discomfort , anger and confusion. Langer and Saegert ( 1 977) found that supermarket shoppers had the aversive ness of their crowded environment amel iorated when provided with i nformation about the effects of crowding. 96 It seems possib le , therefore , that informing people about their re lationship with spatial variables may have beneficial effects (reduced perceptions of crowding). The fact that the distance salient i nformation had this effect whi le the noise sal ient i nformation did not, may reflect subjects' judgments about the tangibi l ity and plausibi l ity of the i nformation presented to them. With regard to performance, the model predicts deficits for crowded subjects. Contrary to th is prediction , close subjects performed no differently when compared to far subjects. In considering the influence of information salience on task performance it seems that subjects i n the noise sal ient condition suffered a performance deficit when compared to control and distance sal ient subjects. This decrement in performance might i ndicate that i nforming subjects about bogus events can serve to interfere with cogn itive tasks, especially when compared to the i nnocuous effect of providing accurate i nformation about the experimental environment . Of special i nterest is the fact that interference produced by the noise salient i nformation i n this study runs at complete odds with the effects reported by the study conducted by Worchel and Yohai ( 1 979). I n their study, close subjects who were provided with the noise arousing explanation outperformed close subjects who were not provided with any explanation . Such contradictory results represent an interpretive chal lenge. At the very least it would seem prudent to consider the claims of the two factor model with renewed caution in the l ight of the resu lts which have emerged from the present study. It is possible to have subjects i n crowded conditions without 97 sign ificantly increasing arousal levels, and without detrimentally influencing thei r performance on cognitive tasks. A further conclusion to emerge from the current study is that the sort of information which subjects are given may have a negative effect on subsequent performance. The two factor model makes predictions about the influence of crowding on the manner in which people relate to one another. Specifically, the model predicts i nterpersonal hosti lity among crowded subjects, in contrast to uncrowded subjects. Five separate dependent measures were taken to examine th is prediction . The pattern of results was fairly consistent across the measures, with a general fai lure to establish any differences in the way in which crowded subjects responded to each other when compared to thei r uncrowded counterparts. The crowded subjects were no more punitive in their treatment of Johnny Rocco, reported no less friendliness towards their fel low subjects, detected no more aggression on the part of their fel low subjects, reported l iking thei r fe l low group members as much , and rated themselves as feel ing no more hosti le (on the MAACL) than uncrowded subjects. Just two of the affective measures produced differential results for the crowded subjects compared to the uncrowded subjects. Crowded subjects were both more depressed and marg inally more anxious than uncrowded subjects. However, the general lack of support for the predictions of the two factor model is impressive in its consistency. This suggests the model has been unable to capture the crucial e lements i nvolved in the experience of crowding . 98 This fai lure to find the predicted interpersonal hosti lity may have been due to the relative nonarousal of crowded subjects. If this were the case it would suggest that arousal may be important i n mediating the negative concomitants of crowding. However, Worchel and Cooper (1 983) discuss situations where spatial violations may not always lead to the experience of crowding. They i nclude situations such as the spatial i nvasions which occur at sports events and rock concerts, where people apparently fai l to report feel ing crowded. They suggest this may be due to people attributing their arousal to the source which is most evident to them, namely the excitement of the game or concert. The present findings high light two other possible explanations for fai lu re to report crowding at events where crowds typically gather. Fi rstly, it may be that spatial vio lation is simply not arousing for people who attend such events. This nonarousal is possible, as the results of the present study testify. Secondly, it may be that people do become aroused but that no misattribution occurs. The arousal in this case may be due to the excitement of the game or concert, and people correctly attribute the source of thei r arousal to that . People may be aware of the arousal they are experiencing (especial ly if the game is exciting) , but the presence and closeness of others is not arousing because it is expected and considered reasonable withi n this context. Sports events and rock concerts, to take but two examples, legitimise spatial vio lation and under these conditions the negative aspects of crowding are not experienced. This particular explanation receives some support from the results of the present study as crowded subjects were nonaroused and those who were 99 informed of the distance manipu lation experienced less confi nement as a result of the information they had been given . The suggestion here is that this info rmation somehow served to legitimise the close conditions. On a broader level , the lack of negative elements which the current research has been able to document suggests that the model should perhaps be modified to account for such disparate results. Further, there now exists some additio nal evidence for the view that crowding is not the universal ly negative experience that some authors have believed it to be. In many respects, the experience of the crowded subjects i n the current research was i ndistinguishable from the experience of the uncrowded subjects. They responded on the majority of the measures no differently to the uncrowded subjects. The effect of i nformation on the measures of i nterpersonal hosti lity was l imited with less punishment, on the scale deal ing with John ny Rocco, being recommended by those subjects in the close condition who were provided with i nformation about the nature of the experience i n advance when compared with u ninformed controls. On a purely exploratory level there were two additional findings of i nterest both reflecting the differential impact of information on subjects' behaviour whi le i n crowded conditions. For those subjects who were i nformed of the bogus noise (noise and noise + distance conditions), abi l ity to concentrate was impai red when contrasted with those subjects who were g iven information about the nature of 1\1. . , I ' -- - 1"T""'\ .... . � 1 00 the distance manipulat ion. This comparative reduct ion i n abi l ity to concentrate suggests that providing i nformation which is designed to ameliorate the negative effects of crowding may have actually been counterproductive in achieving th is aim. The other aspect of th is information which may be important is that i t was in acc u rate and therefore may have i nterfered with subjects' abi l ity to concentrate . This suggestion is supported by the resu lts which showed that noise informat ion i ncreased distraction. Specifical ly, it was found that i nforming crowded subjects of the possibil ity of subl iminal noise in the experimental room increased their reported level of distraction when compared to contro l o r distance sal ient conditions. This disruptive effect of the bogus noise information is of special i nterest because it is the exact opposite of the effect predicted by the two factor model . To conclude, the present results suggest the experience of crowding need not be dependent on feel ings of arousal. In this respect, these results fai l to support one of the basic assumptions of the attributional arousal model . However, it seems unreasonable to argue that the general lack of differences necessari ly demonstrates a fai l u re of the model . These results may have been quite different had subjects shown the predicted levels of arousal. Unfortunately , the question of the extent to which i nformation salience can affect causal attribut ions in the context of crowding is one which remains unanswered by the present resu lts. Neverthe less, the fact that subjects reported feel ings of crowding and confinement without associated increased arousal is an i nterest ing one. At a theoretical level this finding requires examinat ion and i nteg ration , and at an empirical level requires further consideration. What seems to be necessary for future research to consider is a way of reconci l ing these unexpected results within some meaningfu l theoretical framework. This task wi l l be attempted in the next chapter. 1 0 1 1 02 CHAPTER SEVEN AN I NFORMATION EXPECTANCY MODEL O F CROWDI NG The two factor attributional model of crowding has received some support i n the literature (Worchel & Teddl ie, 1 976 ; Worchel & Yohai , 1 979 ; Worchel & Brown , 1 984) . In addition , the model has received broad coverage in numerous introductory social psychology textbooks (e.g. , Baron & Byrne, 1 987; Forsyth , 1 987; and Penner, 1 986) . Despite receiving some research support and wide acknowledgement in i ntroductory texts, the model fai ls to explain two key findi ngs which emerged from the fi rst study. Fi rstly, the model (which is based on the idea that arousal precedes crowding), makes no provision for the experience of crowding w ithout arousal. This is the very situation described in the fi rst study. Secondly, the model makes predictions concerning the effects of provid ing subjects with i nformat ion about alternative sources of arousal , but is unable to explain why, for example , providi ng accurate i nformation about spatial variables reduced the impact o f the experience of crowding. In short , i t seems as though the fi rst study highl ighted inadequacies i n the power of the two factor model to explain and accurate ly predict human responses to the experience of crowding. A new model is cal led for in an attempt to deal with these inadequacies. The proposed (information expectancy) model incorporates one essential 1 03 element of the two factor model . Namely, the assumption that spatial i nvasion is an i ntegral part of the experience of crowding. The notion that spatial i nvasion mediates crowding is common to a number of crowding theories (e.g . , Epstein & Karl in , 1 975; Proshansky et aI . , 1 970 ; Freedman , 1 975) . - The relationship between spatial i nvasion and crowding was initial ly mentioned in Hal l's (1 959, 1 963, 1 966) work. The numerous attribution models, i ncluding the two factor model , also regard crowding as being in itiated by an invasion of personal space. The second e lement i n the proposed model is a cognitive one in which information re levant to the spatial i nvasion is appraised. It is proposed that information appraisal wh ich occurs prior to the spatial invasion wi l l serve to shape expectations. Based on the finding that providing accu rate information about spatial variables (thereby setting realistic expectations) can reduce the impact of crowding , it is further proposed that confi rmation of spatial expectations wi l l also reduce the impact of crowding. It is suggested , however, that disconfirmed spatial expectations wi l l serve to i ncrease the impact of crowding . The theoretical relationship between i nformation and appraisal was discussed within the perspective of cogn itive control by Averi l l ( 1 973) . Information ( referred to as "information gain") is , accordi ng to Averi l l , concerned with cognitive preparation for an event, and relates to the predictabi lity and anticipation of that event. Appraisal involves both interpretation and evaluation of events which are l inked to expectations. Schmidt and Keating ( 1 979) maintain that i nformation functions to i nitiate cognitive preparation and to generate accurate expectancies. 1 04 Expectations are seen as important because they provide the individual with the opportun ity to prepare for the specific context in which the spatial variables wi l l occur. Such contextual considerations have been shown to have an important bearing on personal space requ irements (Gifford, 1 982) , · but to date the context i n which crowding occurs remains largely unexplored. It is the contention of the information expectancy mode l of crowding that when a context is made explicit, certain expectations about what is spatially appropriate become apparent. This contention wi l l be experimentally examined in the next study with the aim of clarifying the ro le of information and expectations as a fi rst step towards li nking experimental studies of crowding to similar experiences which occur naturally in the context of a social setting. Support for the ro le of expectations comes from a number of sources. Di rect support comes from the existing literature while i ndirect support may be derived from the fact that a number of key crowding studies may be reinterpreted to fit with the fundamental propositions of the i nformation expectancy model. In additio n to these two l ines of support (to be discussed shortly) , the role of expectations are also implicated in reports (Tuan , 1 977; cited in Worchel and Cooper, 1 983) that people who attend crowded events do not report feeling crowded. This apparent contradiction may simply reflect expectations about spatial aspects of such events being confirmed. Spectators may expect (on the basis of previous attendance) to watch their favourite sport standing shou lder to shou lder with other fans. Furthermore, they may feel quite comfortable with that level of density because it is expected i n such situations. 1 05 Further support for the role of expectations having a bearing on the experience of crowding comes from a study by Langer and Saegert ( 1 977) in which shoppers in a crowded supermarket were provided with information about what they cou ld expect under such conditions. This i ntervention attenuated the negative effects of crowding. The authors conclude that knowledge of potential relationships with the environment wou ld seem to be beneficial. These benefits included ease of finding items, the extent to which other customers were perceived to be interfering , level of comfort and perceived levels of crowding. Baum et al . (1 981 ) reviewed the literature on pre-exposure i nformation . They exami ned a variety of different contexts and concluded that accurate expectations of what one may feel , ' or what wi l l happen , may reduce stress. These contexts were considerably broader than the usual ones in which spatial concerns are of i nterest. They i ncluded : preoperative information improving recovery from surgery ; preparatory information reducing distress i nvolved i n painful medical procedures; and pre-exposure i nformation reducing stress associated with crowding and noise. Across all these settings the most usefu l i nformation is that which leads to expectations which are accurate, that is , expectations which are confirmed by the experience subsequent to the setting of expectations. Baum et al. suggest the reason accurate expectations may prove usefu l is that they i ncrease the predictability of the situation and this may assist in copi ng with the situation . Accurate expectations may also provide standards against which social comparison may occur. 1 06 Gochman and Keating (1 980) were also i nterested in the relationship between expectations and crowding. They argue that disconfirmed expectations and unattained goals wi l l result in increased arousal which wi l l in turn be misattributed to density-related sources , even if the disconfi rmation of expectations and nonattainment of goals is not due to density-related factors. They examined the effects of disconfirmed expectations on attribut ions to crowding. Their resu lts showed that disconfirmed expectations about nonspatial aspects of the envi ronment can increase attributions to crowding. Expectancy disconfirmation occurred both i n laboratory and fie ld settings, employing both disconfirmation of t ime and performance expectancies, with result ing i ncreases i n attributions to crowding. Disconfi rmation resulted in subjects saying they fe lt more hindered by the number of people in the room. Actual perceptions of crowding were not significantly increased by nonspatial expectancy disconfi rmation . K le in and Harris ( 1 979) note that informing subjects of crowded condit ions has received little attention. They examined the effect of anticipating a crowd and subsequent confi rmation or disconfirmation of that expectation. This study fo l lowed the traditio n of anticipatory crowding studies (e.g . , Baum & Greenberg , 1 975) , and employed a forewarning which was indirect. When sign ing up for the experiment, subjects either signed on a sheet which had either two or five l ines for subjects' names. Klein and Harris' results suggest that expectancy confirmation enhanced performance while those subjects who were warned about a crowd which fai led to materialise performed most poorly. 1 07 I nd i rect support for the i nformation expectancy model may be taken from the Worchel and Yohai ( 1 979) study which crossed i nformation with distance in a two factor design . Information presented to subjects either led them to expect arousal o r relaxation while in crowded conditions. Control groups received no such information. The arousal expectation condition reported a reduced level of perceived crowding. The authors interpreted this finding as evidence for the misattribution of arousal , but it is also possible to invoke an expectancy confirmation interpretation . Subjects in the arousal information condition were led to expect to fee l "stressed and uncomfortable", which (even given the findings of the fi rst study) is an expectation more l ikely to be confi rmed by the experience of crowdi ng than subjects who were led to expect they wou ld feel " relaxed and calmed". There is, after all , some evidence suggesting that crowded subjects may feel aroused (stressed and uncomfortable). As yet, there is no evidence demonstrating that crowded conditions are re laxing and calming. The point being that subjects in the arousing information condition are l ike ly to have had their expectations about how they wou ld feel in those conditions confirmed. In contrast , those subjects i n the relaxi ng information condition were l ikely to have experienced expectancy disconfi rmation. Therefore , there is some support for the idea that spatial information is used in the establishment of expectancies and that personal space invasion is a necessary prerequisite for the experience of crowding. 1 08 The cornerstone of the i nformation expectancy model is that : expectan cy confirmat ion may redu ce the impact of crowd in g, w hil e disconfirmation may serve to in crease the impact of crowding. PREDICTIONS OF THE INFORMATION EXPECTANCY MODEL The fi rst assumption of the model is that any i nformation avai lable or provided about the situation wi l l be appraised by the individual. The i nformation wi l l serve to shape i ndividuals' expectations. These expectations may be classified as either spatial or non-spatial. Spatial expectations refer to expectations about distancing behaviour i n particu lar situations. Information re levant to spatial expectations may, for example, be helpfu l in answeri ng an i ndividual's concerns about whether to expect close or far interpersonal distances. Non-spatial expectations wi l l derive from information which is i rrelevant to concerns about interpersonal distancing. Non-spatial expectations are of less importance to the model which attempts to understand the processes involved i n the experience of crowding. The major concern is with those elements which have a di rect and obvious bearing on crowding. Non-spatial expectations may wel l have a bearing but the nature of this must be considered as secondary, at least unti l the role of spatial expectations has been clarified. The second major assumption is that invasion of personal space is a necessary 1 09 prerequisite for the experience of crowding. The model assumes that interpersonal distance wou ld need to be reduced before crowding would be experienced. This sequence of events is flexible. The model is not concerned with the order of events so much as the process ; that is, spatial i nvasion may or may not precede information appraisal. For example, it could be that under some ci rcumstances the spatial invasion occurs as the fi rst stage in the process, as i n the case where a person i s thrust quite unexpectedly into a crowded situation . Normally , however, i t wou ld be expected that some information wou ld precede the spatial i nvasion. The spatial invasion is the physical aspect of crowding and is an essential component in the experience of crowding. It is the element about which the cognitive aspects of the model are centred. It is now possible to derive some predictions from the model based on the assumptions that information wi l l be appraised and that an i nvasion of personal space is important for the experience of crowding. The first prediction relates to the effects which both confirmation and disconfirmation wi l l have on perceptions of crowding for spatial and nonspatial expectations. It is p redicted that if spatial expectations are confirmed, th is wi l l have the effect of reducing the impact of crowding . Such reduction refers to perceptions of crowding bei ng lessened, subjects feeling less aroused and performance deficits being less l ikely. Confi rmation of spatial expectations wou ld parallel the situation which occurred in Study One, where accurate i nformation about spatial variables reduced the impact of crowding. 1 1 0 Note that "accurate" means the information provided led subjects to expect that interpersonal distance was a variable of experimental interest and this distance may cause them to fee l uncomfortable. In other words, the accuracy of the information was verified by the nature of the subsequent experience which wou ld have thereby confi rmed the in itial expectancies. In this way, accurate information about the crowded situation wi l l lead to confi rmation of spatial expectations. Further, it is predicted that di sconfi rm ation of spatial expectancies may serve to exacerbate negative aspects of the experience of crowding. Thus, perceptions of crowding may be increased, and increased discomfort and performance deficits may also occur. Just as accurate information is l inked with spatial confi rmation , the converse is also the case. Inaccurate spatial information wi l l lead to disconfi rmation of spatial expectations. S imply stated , false information wi l l create expectations which wi l l be at odds with the reality of the subsequent crowding experience. A distinction needs to be made between positive and negative disconfi rmation. Negative spatial disconfi rmation refers to experiencing l ess interpersonal distance than expected , whereas positive spatial disconfirmation refers to experiencing greater interpersonal distance than expected. The i nformation expectancy model is concerned with the experience of crowding and therefore experiences where i nterpersonal distances are at a minimum wi l l be of most interest. One of the implications of the d istinction between positive and negative spatial disconfi rmation is that positive disconfirmations are l ikely to be of little 1 1 1 i nterest , as i n these situations individuals wi l l actually have more space avai lable to them than they anticipated. Such situations are un l ikely to be perceived as "crowded". The mode l can also address the question of what influence nonspatial variables have on the experience of crowding. In this regard Walden and Forsyth ( 1 981 ) h ighl ighted a relatively unexplored issue which concerned the effect of a second sou rce of stress which is concurrently imposed on those who are faced with "excessive i nterpersonal proximity". These authors attempted to establ ish an "additive stress" model , but thei r resu lts failed to support such a model. Note, however, that the sou rces were stress-re lated autonomic effects which were purported to be drug (placebo) i nduced. I n this respect the second stressor was intended to be symptoms (increased heart rate, respiration and other anxiety related symptoms) not un l ike those sometimes associated with the experience of crowding. Specifical ly, it was predicted that nonspatial expectations wi l l i nf luence the experience of crowdi ng such that disconfi rmation of nonspatial expectations wi l l exacerbate the negative aspects of crowding . Simi larly, i t was predicted that confi rming nonspatial expectations wi l l operate to amel iorate the negative aspects of crowding. These predict ions derive from the nucleus of the information expectancy model . The fol lowing study examined the extent to which these predictions received 1 1 2 empi rical support. But fi rst, the predictions of the model wi l l be further clarified by explicitly stating some formal hypotheses. HYPOTHESES Appraisal The i nformation expectancy model is based on the premise that subjects wi l l process avai lable informatio n and th is wi l l resu lt i n the formulation of expectations about both spatial and nonspatial concerns. The fol lowing hypothesis examines the validity of this premise. It was predicted that : 1 . 1 subjects whose spatial expectations were confi rmed would report anticipating less i nterpersonal distance than subjects whose spatial expectations were disconfi rmed. Crowding A fundamental assumption of the i nformation expectancy mode l is that subjects appraise i nformation . It was predicted that if this information was accurate this would result i n lowered perceptions of crowding. Specifical ly, it was predicted : 2 . 1 that subjects whose spatial expectations were confi rmed wou ld report reduced perceptions of crowding when compared to (no information) control subjects who were not i nduced to have any spatial expectations. 1 1 3 S imi larly, it was predicted that inaccurate information which lead to expectations subsequently disconfirmed by close interpersonal distancing, wou ld serve to exacerbate o r i ncrease perceptions of crowding. It was therefore hypothesised : 2 . 2 that disconfirmed spatial expectations wou ld increase perceptions of crowding for those subjects compared to (no information) control subjects. It was anticipated that disconfi rmed nonspatial expectancies wou ld serve to exacerbate the experience of crowding. The role of nonspatial expectancies was examined by the hypothesis : 2 . 3 that confi rmed nonspatial expectations wou ld resu lt i n subjects experiencing less crowding than subjects who experienced nonspatial disconfi rmation . Arousal The i nformatio n expectancy model does not assume that arousal is a necessary prerequisite for the experience of crowding. Thus, if subjects are essentially nonaroused whi le in crowded conditions (as was the case in the previous study), then the confi rmation or otherwise of their spatial and nonspatial expectations should make litt le difference to levels of arousal . It was therefore hypothesised : 3. 1 that no differences i n reported levels of arousal wou ld be evident between subjects whose spatial expectations were confirmed and those whose spatial expectations were disconfi rmed. Simi larly: 1 1 4 3.2 that no differences i n reported levels of arousal wou ld be seen between subjects who experienced nonspatial confirmation and those who experienced nonspatial disconfirmation. Annoyance It was anticipated that the annoying elements of being i n a crowded situation cou ld be reduced with the provision of accurate information about both spatial and nonspatial concerns. In this regard it was predicted : 4 . 1 that subjects whose spatial expectations were confirmed wou ld report less annoyance at being crowded than (no information) control subjects. Simi larly, it cou ld be argued that inaccurate spatial information wou ld i ncrease subjects' annoyance levels u nder crowded conditions when compared to control subjects. It was therefore predicted : 4.2 that subjects whose spatial expectations were disconfirmed wou ld report more annoyance at being crowded than (no i nformation) control subjects. The effect of disconfi rmed nonspatial expectations was examined by the next hypothesis. It was expected : 1 1 5 4 .3 that disconfi rmed nonspatial expectations wou ld increase subjects annoyance at being crowded when compared to subjects whose nonspatial expectations were confirmed. Interpersonal Relationships On a wide range of measures, Study One fai led to provide any support for a key prediction of the two factor model . The crowded conditions in the fi rst experiment did not negatively influence interpersonal re lationships. The mode l attempts to explain why some research has found negative consequences (Karlin , Katz, Epstein , & Woolfolk, 1 979) whi le others (Freedman , 1 979) have fai led to do so. The current attempt to explain these differences is embodied in the prediction that where spatial expectations are confi rmed, crowding wi l l not have a negative i nfluence on i nterpersonal relationships. The information expectancy model predicts that many of the negative concomitants associated with crowding may be ameliorated by providing people with accurate information re levant to their spatial concerns. It was expected that on a series of measures: 1 1 6 5 . 1 spatial and nonspatial disconfi rmation wou ld resu lt in i ncreased punitiveness, perceptions of others aggressiveness, and decreased liking for others . Affect Previous research has shown that negative feel ings may be associated with spatial density (Smith & Knowles, 1 979) and also social density (Evans , 1 975) . Study One has also documented increases in anxiety and depression i n crowded subjects. It is the contention of the information expectancy model that if subjects are provided with accurate spatial information prior to the experience of crowding , then the impact of that experience on mood may be reduced. Specifically, it was predicted that : 6 . 1 subjects whose spatial expectations were confi rmed would report less negative emotion than subjects whose spatial expectations were disconfi rmed. Also that: 6 .2 subjects whose spatial expectations were confi rmed would report less negative emotion than (no information) control subjects. Performance Performance decrements associated with crowding have been reported by a 1 1 7 number of studies (e.g . , Evans , 1 978; Paulus et aI. , 1 976) . The effect is especially notable when performance is measured on a complex task (Nagar & Pandey, 1 987). However, a number of studies have fai led to find performance decrements on complex tasks under conditions of h igh density (Freedman et aI. , 1 97 1 ; Marshall & Hesl in , 1 975) . The current study examined the influence of confirmed and disconfi rmed nonspatial expectations on task performance, combined with confirmation and disconfirmation of spatial expectations. Nonspatial expectancies were operationalised by way of performance on a set of simple anagrams, with the expectation that these were a practice set. The subsequent anagrams were , for the nonspatial confi rmation condition , easy (as anticipated). However, for subjects i n the nonspatial disconfirmation condition , the subsequent anagrams were more difficult. Verification of the effectiveness of the manipulation was examined by the hypothesis which anticipated that : 7 . 1 subjects given easy anagrams wou ld outperform those subjects g iven difficult anagrams. The present study examined the i nfluence of confi rmed and disconfirmed spatial and nonspatial expectations on task performance. It was therefore predicted that : 7.2 subjects whose spatial expectations were confirmed would outperform those subjects whose spatial expectations were disconfi rmed. Also, that : 7.3 subjects whose nonspatial expectations were confi rmed wou ld outperform subjects whose nonspatial expectations were disconfi rmed. Sex 1 1 8 Much of the previous research examin ing sex differences under crowded conditions has provided equivocal resu lts. However, in reviewing th is l iterature , Nicosia et a l . ( 1 979) have concluded that men react more negatively than women when crowded. In l ine with this general conclusion it was expected that: 8.1 women wou ld report being less crowded than men. Simi larly, that : 8.2 women wou ld report being less annoyed than men. This chapter has i ntroduced the i nformation expectancy model . The model has been discussed i n terms of predictions which have further been b roken down i nto specific hypotheses, to be examined by the next study. The final section of this chapter examines measurement issues, specifically with regard to the concepts of arousal and crowding. 1 1 9 MEASUREMENT OF AROUSAL AND CROWDING An auxi l iary aim of the second study was to improve on the rather limited manner i n which the pivotal concepts of arousal and crowding were measured in the fi rst study. Arousal Previous crowding research has defined and measured the notion of arousal in a variety of ways. As was apparent from Chapter One, the exact meaning and usage of the term " arousal" in the crowding literature unfortunately is rather muddied. Some writers have made a significant contribution in this respect by using the terms arousal and stress i nterchangeably (e.g . , Worchel & Yohai , 1 979). Aiel lo et al. (1 977) also l ink the two concepts by referring to stress-related arousal. I ncreased physiological responses i n crowded conditions are described as "stress reactions". Others are more straightforward ; for example, Karli n and Epstein ( 1 979) simply refer to arousal as physio logical arousal , which they operationalise in terms of skin conductance, pulse rate and blood pressure. I n addit ion t o the concepts of arousal and stress, other more peripheral notions are sometimes used. These terms include "crowding-re lated arousal", "crowding stress" and also "spatial discomfort" , although the last of these is more frequently used in conjunction with other negative aspects of crowding. 1 20 I n general , the literature suffers from a paucity of conceptual definitions of these notions and researchers offer litt le i n the way of clarification of terms, except by way of outl in ing the style in which they have chosen to measure "arousal". These styles of measurement are fascinat ing ly varied. A number of studies, (for example, Aiel lo et aI . , 1 977; Epstei n & Karlin , 1 975 ; and Patterson et aI . , 1 979) have used cogn itive tasks as indices of arousal . Presumably, this practice is based on the re lationship between arousal and task performance which has emerged from social faci litation research. Epstein and Karl in employed two cogn itive tasks together to measure arousal. Aiello et al . also used two cognitive tasks to assess the cognitive correlates of arousal. Awareness of bodi ly changes in arousal (e.g. , heart rate and sweating palms) was employed as the measure of arousal by Webb et al. ( 1 987) . Worchel and Brown ( 1 984) asked subjects about the extent to which they had found particu lar types of fi lms arousing. Evans ( 1 979) adopted a more comprehensive approach and took physiological measures (pu lse rate and blood pressure) in conjunction with performance and observational measures. Physiological measurement can be problematic i n crowded experimental conditions as was the case for Worchel and Yohai (1 979) who abandoned the col lection of palmar sweat measures as a result of subjects' complaints and logistical difficu lties. It is i nteresting to note that Worchel and Yohai argue that the theoretical position they adopt with respect to arousal is that it does not re ly on physiolog ical measurement. They argue that such measurement, while of i nterest, is not absolute ly necessary. The two factor model deals with how people 1 21 handle perceived arousal and for this reason the focus of measurement should be on individuals' perceptions of arousal. This argument contributes little to the question of how best to measure perceived arousal , which remains problematic. Schaeffer and Patterson ( 1 980) adopted three separate measures for arousal. They included a behavioural observation (frequency of self- manipu lative behaviours) , a performance measure , and a self-report check list of arousal symptoms. The results of this study are interesting as the performance measure and the self-report measures were significantly correlated (r = -.69) . The frequency of self-manipulative behaviour was poorly associated with both the se lf-report measure and the performance measure. Schaeffer and Patterson note that on the basis of the correlations among these arousal measures, and from past research supporti ng the validity of the self- report and performance measures, it is l ike ly that these two measures did reflect arousal. The task of measuring arousal for the second study was assigned to two self­ report measures which had both been previously used. The first (an in dex of somati c arousal) was in itially reported in the l iterature i n 1 977 by Aie llo et al . The index of somatic arousal comprises five items asking the subject the extent to which they experienced sweating palms, rapid breathing, rapid heartbeat, tense muscles and nervous tension. The reported intercorrelations for this index were (mean r = .23, median r = .23). The study by Nicosia et al. ( 1 979) used the same index of somatic arousal (representing the mean of the same five items). These five items were reported to be "high ly" i ntercorre lated (r =.35) . The third 1 22 study to report the use of the somatic arousal i ndex was the Patterson et al . (1 979) study. Final ly , the somatic arousal index was uti l ised by Schaeffer and Patterson ( 1 980) , although they refer to it as a checklist of symptoms of nervous tension. The i ndex of somatic arousal has received some use, no doubt encouraged by the acceptable alpha coefficients suggesting reasonable i nternal consistency. Given that the somatic arousal index may possess some reliabi lity , the other important question concerns its validity. Aiello et a l . ( 1 977) note that during crowding , subjects reported a greater degree of somatic arousal . This wou ld support the construct validity of the instrument whi le the fact that Aie l lo et al . report that somatic arousal measures paralleled the resu lts for skin conductance wou ld be suggestive of the convergent validity of the index. In this respect it was found that interpersonal distance preference mediated both skin conductance and reported somatic arousal. Simply stated , this meant that subjects who preferred greater distances had greater somatic arousal levels and skin conductance leve ls whi le i n crowded conditions. N icosia et al. ( 1 979) found that crowded men reported more somatic arousal than did noncrowded men , whi le men i n a crowded but no touch condition were i ntermediate i n their reported somatic stress. Further, these results parallel resu lts for skin conductance , annoyance and expressed bother by spatial i ntrusion. Finally, Schaeffer and Patterson ( 1 980) fou nd that increased intimacy (operational ised by di rect gaze) resulted in i ncreased levels of somatic arousal. 1 23 I n sum , it appears as though the somatic arousal i ndex possesses sufficient support for its psychometric properties to warrant continued use. As a measure it has good internal consistency and several l ines of support for its validity. The second arousal measure (an a rousal scal e) , was based on a scale developed by Mehrabian and Russe l l ( 1 974). The scale consisted of seven items. Each item was presented as two bipolar adjectives which subjects were asked to respond to (e .g . , "How dul l or jittery did you fee l during the experiment?") . The six o riginal items were retai ned , with the on ly change being the "sluggish / frenzied" item being altered to "sluggish / alert" . The term "frenzied" seemed inappropriate for the setting in which it was to be used. The additional item was the " relaxed / i l l-at-ease" item which had previously been used as the sole index of arousal in the fi rst study. Based origi nally on i ntuitive descriptors , the arousal scale has undergone considerable factor analytic deve lopment. Mehrabian and Russel l conceptualise arousal as a feel ing state which varies along a si ng le dimension rangi ng from sleep to frantic excitement. They cite research by Thayer ( 1 967, 1 970) supporting the view that physiological i ndices are high ly correlated with verbal self report measures. They conclude that arousal is a feeling state that is most di rectly assessed by verbal report. Within the crowding literature, this arousal scale has been used by Paulus and Mathews ( 1 980) who fou nd that arousal i ncreased (along with discomfort and unpleasantness) as subjects spent more time in a crowded experimental room. Simi larly, E l l iot and Cohen ( 1 98 1 ) employed the Mehrabian and Russell ( 1 974) 1 24 scale and found that crowded subjects were more aroused than uncrowded contro l subjects. Additionally, they found that of the close , moderate and far i nterpersonal distances uti lised , the close interpersonal distance resu lted in the most arousal. Crowding The measurement of crowding has a history at least as varied as that associated with the notion of arousal. Due to the nature of the construct Choi , Mirjafari , and Weaver ( 1 976) argue that crowding cannot be measured directly but must be i nferred from indirect measures. Such inferences are usual ly taken from self­ report measures which have i ncluded a number of different elements i ncluding the setting , the room , the i ndividual and the extent of discomfort engendered by the experience. Setting-re lated measures i nclude the Baum and Greenberg (1 975) study of anticipatory crowding i n which subjects were asked to respond on a seven point scale examining the degree to which the setting appeared crowded. Greenberg and F i restone ( 1 977) asked a simi lar question , but additional ly asked subjects if they fe lt crowded. Evans ( 1 979) asked subjects whether they perceived the situation as crowded. This question was used as a manipu lation check. The results indicated that subjects i n high density conditions perceived thei r situation as considerably more crowded than thei r low density counterparts. I n th is respect the notion of crowding seems to be remarkably robust. It seems to matter little 1 25 whether questions are d irected at the setti ng , the room or how the i ndividual feels. Crowded subjects almost without exception report a more crowded setting, a more crowded room or feel ing more crowded than uncrowded subjects. A large number of studies have examined various perceptions of crowded rooms. These perceptions range from "how crowded was the experimental room ?" (Hel ler et aI . , 1 977) , to measures which consider various aspects of the crowded envi ronment . Epstei n and Karl i n ( 1 975) , for example, examined percept ions of the experimental room in terms of its lack of privacy ; how crowded, confined and small it fe lt ; and the degree to which subjects felt as though they were competing for space. Klein and Harris ( 1 979) adopted a s imi lar envi ronmental focus i n asking for room ratings i n terms of prettiness, i l lumination levels, stuffiness, size, adequacy, cheerfu lness, dampness, pleasantness, how much physical discomfort was i nduced, and how crowded it appeared. More recent studies have also included room rat ings (Kuykendall & Keating , 1 984 ; Nagar & Pandey, 1 987). The th i rd area in which crowding measurement is often taken concerns not perceptions of the envi ronment, but the impact of that environment on the i ndividual in terms of how that person felt whi le i n a crowded situation . Kalb and Keating (1 981 ) suggest that there are distinct differences between asking people how crowded they fee l and asking them to rate how crowded they perceive a setting to be. They asked both questions (using a between subjects design) of customers in a busy book shop. The two measures were factor analysed and found to be conceptually distinct. The factor loadings for the fee l ing item were 1 26 things l ike behavioural constraint , restriction , confinement and stressfulness. The envi ronmental rating , on the other hand, loaded with high density and negative affect. The most popular single item used to assess feel ings of crowded ness is the question , " How crowded did you feel?". Other i ndividual responses have i ncluded the extent to which subjects have felt tense, calm and friendly (Patterson et aI . , 1 979). Other common areas to be considered by researchers are the questions of confi nement (e.g . , Muel ler, 1 984) , and crampedness (e.g . , N icosia et aI . , 1 979) . The notion of confinement is conceptually l inked to other issues concerning the extent to which subjects experience discomfort whi le crowded, and the extent to which others i n the crowded situation are perceived as bothersome to the i ndividual . These concerns derive di rectly from the behavioural constraint perspective which defi nes crowding in terms of its constrai ning properties. Baum and Greenberg (1 975) assessed physical discomfort by asking subjects how much physical discomfort they had been caused while in the crowded room. Aiel lo et al . ( 1 977) developed a spatial discomfort index which comprised seven items looking at aspects of crowding such as l imited privacy, others being too close, and being in competition for space. Hel ler et al. ( 1 977) asked about the extent to which the presence of others i nterfered with performance. Simi larly, Gochman and Keating ( 1 980) asked subjects about the degree to which they were hindered by the number of people in the room. The i nterference value 1 27 associated with other subjects was also examined by McCallum et al. (1 979). For the purposes of the second study it was decided to use two measures of crowding which had been previously used in experimental crowding research. The fi rst of these was an in dex of spat ial intrusion developed by Nicosia et al . ( 1 979) . This i ndex consisted of three items asking subjects how bothered they were by territorial invasion , by the number of people i n the room , and by close i nterpersonal distances. The authors report an inter-item coefficient (r = . 4S) which they describe as suggesting the items were high ly intercorre lated. The reported results for this i ndex are in accordance with theoretical predictions about crowding , thus providing some supporting i nformation about the construct validity of the i ndex. Specifical ly, the authors report that male subjects were more bothered by spatial i ntrusion than females. Also subjects who were i n the crowded touching condit ion were more bothered by spatial intrusion than those subjects who were in the no touch crowded condition . The second measure , an in dex of in div idual crOW din g, was based on the study by Worchel and Teddlie ( 1 976). These authors asked four questions to measure subjects' experience of crowding. These questions asked subjects how comfortable, confined, and i I I-at-ease they fe lt , and also how crowded they found the experimental room. These four items yielded simi lar results and were therefore combined to form one index of crowding. The validity of this index was supported by the result that reduced interaction distance was associated with i ncreased reports of crowding on this index. Walden and Forsyth ( 1 981 ) used the 1 28 same i ndex, the only change bei ng that the item asking about crowding was altered to focus on how crowded the subject felt during the research procedure. The results of th is study suggest that three of the four items were usefu l in differentiating between subjects who were seated at close as opposed to far i nterpersonal distances. The i l l-at-ease item was the item which appeared to contribute nothing. For this reason, as wel l as the fact that this item had been i ncluded in the arousal measure , this item was dropped from the current index. It was replaced by an item asking how cramped subjects felt whi le in the experimental room. This item , in modified form ("how cramped was the experimental room ?") was taken from Nicosia et al. ( 1 979) who reported it as being high ly successfu l in differentiating subjects' reports between crowded and uncrowded conditions. The change in focus from room to how the i ndividual felt was i n keeping with the earlier defi nition of crowding as being a subjective experience. 1 29 C HAPTER EIGHT STU DY TWO: METHOD OVERVIEW AND DESIGN The general aim of Study Two was to examine the predictions derived from the information expectancy model. The experiment was a 3 x 2 facto rial design in which SPATIAL EXPECTATION was crossed with NONSPATIAL EXPECTATION. The two independent variables were operatio nalised by way of inform at ion (for spatial expectation) and task difficul ty (for nonspatial expectation). All subjects i n th is experiment experienced close i nterpersonal distance. The focus of the study was to examine processes which occur whi le crowded, so for that reason it seemed unnecessary to i nclude a "far" condit ion. This procedure of running subjects on ly at close i nterpersonal distances has previously been employed in crowding research (Patterson et aI . , 1 979) . The advantage of th is approach is that it reduces the al ready large number of subjects required for th is type of research. The main disadvantage is that possible i nteractions between interpersonal distance and the other i ndependent variables are not possible to detect. Comparisons between close and far i nterpersonal distance g roups are also not possib le, although Study One clearly showed that the distance manipu lation was successfu l . 1 30 Spatial expectation was manipu lated by i nforming subjects of the experimental aim which was to examine the effect of distance on behaviour, and leading them to expect to be seated at either an unusually close distance, or at a distance that wou ld usually be regarded as normal. The contro l subjects received no i nformation , and therefore should have had no particu lar expectations about the spatial aspects of the experiment. Nonspatial expectation was manipu lated by leadi ng all subjects to expect a series of easy tasks (anagrams) and either subsequently doing very well on such tasks (as expected) , or (unexpectedly) doing very poorly on such tasks. This type of manipu lation has previously been used in crowding research by Gochman and Keating ( 1 980) . SUBJECTS The subjects were 240 male and female fi rst and second year psychology students who volunteered for the study. Subjects received course credit for participation . Their participation meant they were requi red to attend one less laboratory class. A certain level of attendance was requ i red to successful ly complete these courses. Subjects were run in same sex groups of five previously u nacquainted subjects. Three quarters of the subjects were female ( 1 80), with one quarter being male 1 31 (60) . This ratio of three to one fai rly represented the composition of the classes from which the subjects were drawn. The decision to use male subjects was made so as to allow an examination of sex differences under these particu lar laboratory conditions. Although no sex differences were noted in the Worchel and Yohai (1 979) study, it was sti l l regarded as possible that sex may have been an explanatory factor for the cu rious results from Study One, which was run with groups of female subjects on ly. At the point when subjects were solicited they were told that the study was looking at g roup performance and wou ld take approximately 50 minutes. The educational benefits of participation were mentioned, after which volunteers were asked to complete a consent form. This form included name and te lephone number, and subjects were told to expect to be contacted later to arrange a date and time to complete the procedu re . Al l volunteers were told to expect to receive a summary of the resu lts fol lowing the conclusion of the study. Subjects were allocated to groups on the basis of avai labi lity and nonacquaintance. Once constituted , g roups were then randomly assigned to one of the s ix experimental conditions . 1 32 EXPER IMENTAL MATERIALS The pre-expe riment al room was a large room adjacent to the experimental room where subjects were issued with in itial instructions · and completed the fi rst set of anagrams. At the conclusion of the experiment subjects returned to this room for completion of post-experimental questionnaires and debriefings. The expe rimental room was the same smal l room as used in the fi rst study. The room was labe l led "Small G roup Research Lab" , and was furnished with five armless chairs . The chai rs were arranged in a closed ci rcle in the center of the room . I n the center of the circle a microphone was suspended from the cei l ing. Experi m en t al p re-in structions (see Appendix G) were al l pre-recorded. These three sets of i nstructions had an identical begi nning which introduced the practice anag rams and led subjects to expect the second set to be simi lar to the practice set. The pre-instructions contro l led the timing of the practice anagrams. The th ree sets of instructions corresponded to the three i nformation conditions. This i nformation aimed to manipu late subjects' spatial expectations. Subjects i n the CONTROL condition were given no information about what to expect i n the experimental room. SPATIAL CONFIRMATION subjects were informed that the study was i nterested in the effect of distance on behaviou r and they should therefore expect to be seated at an unusually close distance to others. Subjects in the SPATIAL DISCONFIRMATION condition were also told the study wou ld be examin i ng the effect of distance on behaviour, but they shou ld expect to be 1 33 seated at a distance which wou ld usually be considered normal. In addition to the three sets of pre-instructions there was also a set of pre­ recorded instructions which standardised the conditions under which the test anagrams were completed. The instructions indicated that there wou ld be 1 0 anagrams and they would be al lowed 30 seconds to work on each one before they were to begi n the next one. These instructions ensured the timing was the same for all g roups. These instructions are included in Appendix G . The pr actice anagr ams were provided to subjects i n the form o f a "Practice Anagrams Booklet" with an accompanying answer sheet (see Appendix H) . These anagrams were modelled on anagrams used by Feather and Simon ( 1 971 ) . There was, however, a need to deve lop both an easy and a difficu lt set. Two sets of 1 0 anagrams were developed and pi lot tested with 1 0 col leagues. The mean number correct for these two sets was 8. 1 for the easy set and 3.5 for the difficu lt set. The practice anagrams comprised three very easy anagrams which Feather and Simon had also used to induce the expectation of success at the task. The test an agr ams (see Appendix H) were either easy or difficult. The practice anagrams were different from those appearing in the easy set of test anagrams. The same iden tification ba dges used i n Study One were used again . These badges were labelled "A" to "E" and were helpfu l in assisting subjects identify individuals when completing the post-experimental questionnai re. 1 34 A group word task simi lar to the task used in Study One was intended to provide a measure of g roup performance. The task involved extracting smaller words from a master word ( INDUSTRIOUSLY). As a g roup, subjects worked together with the aim of deriving as many words as possible within the prescribed time limit. The grou p discussion/decision task was the same as that used in Study One. The John ny Rocco task requires subjects to read the detai ls of the case study about a young offender and then decide on the most appropriate treatment. The range of t reatment avai lable is l imited to seven options on the Love-Punishment scale. Th is task has previously been seen as an index of interpersonal relationsh ips. A pos t-ex perimen tal questionn ai re (see Appendix I ) consisting of 37 items s imi lar to the question naire used in the previous study measured subjects' responses to the experimental room , and to their fel low group members. One aim in deve loping this questionnai re was to improve the psychometric properties of the measurement of the concepts of crowding and arousal. An implication of Study One was that the measurement of these concepts was based on a response to a s ingle item and was therefore psychometrically questionable in terms of re l iabi lity and validity. Single item measures are notoriously unstable (Anastasi , 1 988). Other th ings being equal , the longer a test the more reliable it wi l l be. The reason for this is that larger samples of behaviour wi l l produce more adequate and consistent measures. However, in Study One the decision to use 1 35 these si ng le item measures was made out of the need to keep essential aspects of the study the same as the Worchel and Yohai ( 1 979) study. On the basis of a review of the measurement practices previously employed in crowding research, i t was decided to develop fou r scales to attempt to more accurately measure the key concepts most relevant to the experimental hypotheses. These fou r scales were CROWDING, ANNOYANCE, AROUSAL and SOMATIC. The CROWDING scale comprised four items ("How crowded! confined! cramped! comfortable did you feel?") . This scale attempted to measure the extent to which subjects were crowded in the experimental room. The ANNOYANCE scale attempted to measure the extent to which , under the experimental conditions, the presence and closeness of other people was a bother to subjects. The three items used were as fol lows: "How bothered were you by the number of people in the room! invasion of your own personal space ! the closeness of other subjects?". Both the AROUSAL and SOMATIC scales were designed to measure perceived arousal at a general level (AROUSAL) and also at the level of more specific awareness of physio logical symptoms (SOMATIC). The AROUSAL scale comprised seven items: "How relaxed! aroused! alert! jittery! stimu lated! excited! sleepy did you feel?". The SOMATIC scale was made up of five items: "To what 1 36 extent did you experience rapid heartbeat! nervous tension/ rapid breath/ muscle tension/ sweating palms?". Affect was measured by the MAACL-R. This revision (Zuckerman & Lub in , 1 985) of the MAACL attempted to overcome a number of psychometric weaknesses in the original scale. These relate to the fact that scales were h ighly correlated, and further that the scales suffered from an acquiescent response set. More specifically, the subscales of the MAACL typically correlated too h igh ly with each other to infer that the individual scales had good discriminant validity. These suspicions about discriminant validity were high lighted by the report that the subscales typically correlated between . 7 and .9 on any given · occasion (Zuckerman & Lubin , 1 965). The MAACL-R subscales are moderate ly correlated , with the manual (Zuckerman & Lubin , 1 985) reporting corre lations among the three negative affect scales fal l ing between .4 and .6. Whi le this represents a considerable improvement over the correlat ions among the old scales, the authors maintain that the three subscales are representative of the larger dysphoria scale. The dysphoria scale is the sum of the raw scores of the three negative affect scales. The reduced interscale correlations are at least suggestive of greater discriminant validity. The problem of acquiescent response set meant that MAACL scales corre lated sign ificantly with the number of items checked. In the case of the MAACL-R this 1 37 response set has been control led th rough a standardisation process which has reduced the correlations for the negative affect scales from .66 -.68 to .37 -.43. In reviewing the MAACL-R Templer (1 987) refers to it as being brief and easy to administer , with good reliabi l ity , re lative independence of response set, and commendable construct validity. PROCEDURE Subjects were contacted by telephone to arrange a suitable time. Such times were usual ly set days in advance of the actual meeting and a second call on the eve of the arranged time was made. This procedure was simi lar to that employed in Study One where it was found that a second "reminder" cal l increased the l ike l ihood of subjects' attendance. On arrival , all subjects were met by the experimenter and introduced to the female experimental assistant. This woman was not the same person who conducted the fi rst experiment , but was a senior graduate student experienced i n experimental procedure. She was bl ind to the experimental hypotheses. The experimental assistant ushered subjects into the pre-experimental room whe re they were g iven an identification badge to pin on. They were then seated and i nformed that the instructions for that part of the study had been 1 38 prerecorded. At this point the appropriate instructions were played. The three possible sets of instructions (contro l , spatial confi rmation , spatial disconfi rmation) , a l l had identical beginnings (see Appendix G) . The instructions introduced the practice anagrams and subjects were provided with a copy of the practice anagram booklet. They were requested to read the instructions on the cover of the booklet and asked to complete the three items. The t iming of the practice items was contro l led by the prerecorded taped instructions which al low 30 seconds for the completion of each item. At the conclusion of the third item the taped instructions varied the information which was provided for subjects before they proceeded to the next phase of the experiment. Contro l subjects were not given any information. Subjects in the spatial confi rmation condition were told the aim of the experiment (to study the effect of i nterpersonal distance on behaviour) , and they should expect to sit at an unusually close distance to the other subjects. The spatial disconfi rmation information also informed subjects of the aim of the experiment and told them to expect to sit at a normal distance to other subjects The experimental assistant then collected the practice anag ram booklets and answer sheets, and showed subjects i nto the experimental room which was across the hal l and adjacent to the pre-experimental room. The experimental room (the same as used in Study One) contained five armless chairs , the front legs of which touched those of adjacent chai rs. The experimental assistant broke this circle of chairs by drawing back one of the 1 39 chairs and i nviting four subjects to be seated. The fifth subject was also seated after which the fifth chair was pushed in to complete the circle. Once seated , al l subjects were asked not to remove their chairs as the session was being recorded and the chairs had been specially p laced to ensu re optimum recording quality. They were further told that this was important as voice levels wou ld later be analysed. At this point the experimental assistant pointed to the m icrophone which was suspended to the cei l ing , equidistant from each of the five chairs . Fol lowing the request to retai n the seating positions, subjects were introduced to the anagram test task by way of tape-recorded instructions. They were provided with an anagram test booklet, answer sheet and penci l . Half the groups received a booklet containing easy anagrams whi le the remain ing groups received a test booklet contai ning difficult anagrams. Those subjects who received easy anagrams comprised the nonspatial confirmation condition , while those who received difficu lt anagrams experienced nonspatial disconfi rmation. At the completion of the first item , the experimental assistant left the room and retu rned on the completion of the last of the 1 0 anagrams. After col lecting the materials used for the anagram task the group word task was i ntroduced. Subjects were instructed to work together to produce a single l ist within the 1 0 minute time l imit. One subject acted as a recorder for the group and was handed a clipboard , paper and pen. After checking to ensure that al l subjects understood the task the experimental assistant left the room . She returned after exactly 1 0 minutes and collected the list. 1 40 The Johnny Rocco case was next introduced by providi ng each subject with a copy of the case study and also the Love / Punishment scale. Groups were then asked to d iscuss the case and try to reach a group decision as to which of the seven treatments they wou ld recommend for Johnny. Five minutes was allocated for this task. Again , the experimental assistant left the room for the duration of the task, returning at the conclusion to col lect the materials. Final ly , subjects were handed a copy of the MAACL-R and asked to complete the checklist. The directions for the completion of the checklist were read out and they were asked to identify themselves by the identification letter which corresponded to the badge they had worn . When al l subjects had finished, the checklists were col lected and it was explained that the experiment was over, but that one fi nal task remained. Subjects were taken back to the pre-experimental room , seated at desks and asked to complete a copy of the post-experimental questionnai reo Fol lowing the completion of the post-experimental questionnaire, the experimental assistant col lected the materials and thanked subjects for their participation . At this point , she departed and the experimenter debriefed the subjects and answered any remaining questions before dismissing them. Subjects were asked to respect the confidentiality which the study requi red. 1 41 ETH ICAL CONSIDERATIONS AND DEBRIEFING Some deception accompanied each of the independent variables. I n the case of spatial expectation, subjects entered the experimental room with either an accurate expectation of spatial arrangements, an inaccurate expectation , or no expectation . Thus, one third of the subjects were correct ly informed, whi le one th ird were m isinformed , with the final third being un informed. Those subjects either misinformed or un informed were subject to experimental deception . Nonspatial expectation was manipu lated by exposure to either easy or difficu lt anagrams. In the case of the easy anagrams these were consistent with what wou ld have been expected fol lowing the trial anagrams, with no deception involved. Subjects unexpectedly exposed to a set of difficult anagrams had pu rposefu l ly been deceived about the nature of the task they could expect. This misinformation was seen as necessary in order to exami ne the effect of disconfirmed nonspatial expectations on the experience of crowding. A third form of deception invo lved the ruse which a l l subjects received as a "rationale" for retai ning their al located seating posit ions. They were misinformed that the close seating was necessary as the sessions were recorded and voice levels were monitored for later analysis. This ruse was necessary to encourage subjects to remain in the close seating formation . The cover story was, however, based on a half truth. It was correct that the sessions were recorded, but voice leve ls were unimportant. The sessions were recorded so the experimental 1 42 assistant cou ld monitor the sessions when she was not present and also the recordings were taken so they cou ld be re-examined for any casual comments which subjects may have made about the nature of the seating arrangements. Any overt signs of distress or discomfort shown by the subjects was to be attended to by the experimental assistant who was close to the subjects at al l t imes. When the procedu re required the experimental assistant to be out of the room she had been instructed to wait in an adjacent room where the voices of the subjects were being recorded, and also played into the room for the experimental assistant to monitor. As previously, the experimental assistant had been instructed to disband the group if the procedure became obviously distressi ng for any subject. This precaution proved to be unnecessary. In accordance with Tesch's (1 977) recommendations, the debriefing procedure addressed the issues of ethics, methodology and education. The debriefing revealed a l l deceptions to which that particu lar g roup had been exposed. The use of the deception was explained and justified. The main aims of the study were explained and essential aspects of the methodology were outlined. Subjects were asked if they had any questions and these were answered fu l ly. In the case of some groups, the questions indicated a level of real curiosity about certain aspects of the study, whi le in other groups there were no questions at al l . After all questions had been dealt with satisfactorily, subjects were rem inded that fol lowing the conclusion of the experiment when the data had been analysed, 1 43 they would be posted a copy of the resu lts in summary form. These notes (see Appendix J) outli ned the aims, p rocedure and results of the study. It was hoped in this way to maximise the educational value of participation in the experiment by emphasising how their individual contributions make research possible and the value of such research . CHAPTER N I N E STU DY TWO : R ESU LTS PSYCHOMETRIC PROPERTIES OF THE SCALES MEASURING AROUSAL AND CROWDING 1 44 The concepts of arousal and crowding were each measured by two scales. Arousal was measured by a seven item AROUSAL scale and a five item SOMATIC scale . Crowding was measured by a four item CROW D E D scale , and a three item A N N OYANCE scale. One aim of the second study was to employ measurement instruments which improved on the single item measures used in Study One. Accordingly, each of these scales was subjected to an examination of the inter-item rel iabi l i ty (alpha) coefficient, as an index of reliabi lity. The alpha coefficient represents a measure of the extent to which the scale is i nternal ly consistent. Internal consistency estimates of reliabil ity are based on the average intercorrelations among the items on the scale. If the items intercorre late then the scale wil l have some internal consistency , and is more l ikely to be measuring the same construct. Another issue concerned the question of whether this re liabi lity analysis should be based on individual o r g roup data. The analysis of the results for the first study was conducted on the basis of using the group as the un it of analysis. The 1 45 reason was that analysis of variance assumes independence of subjects. The nature of experimental crowding research is l ikely to defy this assumption , and for that reason the unit of analysis for the fi rst study was the group. For the same reason the fo l lowing re liabi lity analysis was based- on the group as the unit of analysis. Fol lowing each alpha coefficient a second (bracketed) coefficient wi l l appear. Th is second alpha denotes the alpha based on the individual as the unit of analysis. As wi l l become clear, the differences between the two sets of figures are min imal. Perce ived arousal was represented by two scales , AROUSAL and SOMATIC. The arousal scale consisted of the sum of seven items. It was assumed that the items (examin ing , for example, awareness of excitement , alertness and stimu lation) wou ld represent subjects' general level of awareness of their arousal. A reasonable level of internal consistency was recorded for the arousal scale with alpha = .64 (.68). The somatic scale comprised five items asking subjects about specific awareness of physiological symptoms such as heartbeat, rapidity of b reathing and sweating palms. A high level of internal consistency was recorded, with alpha = .83 ( .77). Two scales were used to measure crowding , a CROWDED scale designed to measu re the extent to which people felt crowded, and an ANNOYANCE scale which looked at the extent to which others were seen as annoying i n the experimental sett ing. The crowded scale was made up of fou r items asking subjects about crowding , confi nement, crampedness and comfort. The alpha coefficient was high at .85 (.83). The annoyance scale comprised three items asking about the extent to which others were of bother in the experimental setti ng. For this scale, the i nternal consistency was again high , alpha = .90 (.8 1 ) . 1 46 In summary, the leve ls of internal consistency for these scales was sufficient to justify the ir use. The alpha levels were more than encouraging and the subsequent analysis was able to proceed with some confidence i n the measures. MAN IPULATION CHECKS Manipulation checks were concerned with several issues. Firstly , the effectiveness of the distance manipulation. Secondly, the effectiveness of the information manipulation in leading subjects to expect to sit at close or normal distances. Thirdly, the issue of the relative difficu lty of both the practice and test anagrams. DISTANCE MAN IPULATION The standard practice in examining the effectiveness of distance manipu lations is 1 47 to compare the ratings of how crowded subjects felt between close distances and far d istances. In this study there were no far interpersonal distance g roups. It was assumed, on the basis of extensive and conclusive previous research, that the i nterpersonal distance manipulation would be effective in inducing fee l ings of crowding i n subjects. While this assumption seemed reasonable, it also seemed prudent to examine subjects' responses which may have a bearing on this assumption . In this regard , subjects were asked whether their personal space had been invaded or not. Such invasion has previously been seen as an important dimensio n in the experience of crowding , and the i nformation expectancy model simi larly rel ied on spatial invasion as being important to the experience of crowding. Of the 240 subjects, 97 (40.4%) indicated that thei r personal space had not been invaded. The remain ing 1 43 (59 .6%) reported that thei r personal space had been invaded. This resu lt was a surprise in view of definitions which are predicated on the notion of a personal space i nvasion being a prerequisite for the experience of crowding. Of further interest is the fact that, when asked to quantify the extent of the personal space invasion, only a small number reported this as a severe invasion. This information is detai led in Table 5 be low. Table 5 shows that very few subjects fe lt as if their personal space had been severely i nvaded. Over half the 59 .6% of subjects who reported a spatial invasion, rated the extent of the personal space invasion as being either 7, 8 or 9, that is, of minor magnitude. TABLE 5 FREQUENCIES FOR EXTENT OF PERSONAL SPACE INVASION TOTAL VALUE o 1 2 3 4 5 6 7 8 9 FREQUENCY 4 1 1 7 1 3 8 1 2 1 2 28 22 26 1 43 Note : 0 = severely invaded, 9 = sl ightly invaded. PERCENT 1 .7 4 .6 2.9 5.4 3.3 5.0 5.0 1 1 .7 9 .2 1 0.8 59.6 INFORMATION MAN IPULATION AND SUBJECTS' APPRA ISAL The three i nformation conditions were designed to i nduce subjects either to 1 48 expect to be seated at an unusual ly close distance, or a normal distance , or in the case of the control g roup no particular expectations about seati ng arrangements were conveyed. As a manipu lation check subjects were asked how close they had expected to be seated to other subjects. On a scale for which higher values indicated an expectation of g reater closeness, the condition which had the highest mean value (5.6) was the condition which had led subjects 1 49 to expect to be seated at an u nusually close distance (spatial confi rmation) . The next highest expectation for closeness was recorded for the condition in which subjects were provided with i naccurate spatial i nformation (4.54) . Subjects in th is condition (spatial disconfi rmation) were told to expect to be seated at a normal distance from other subjects. Final ly, the control condition recorded the lowest mean value (3.74) suggesting that they were less l ikely to have expected to be seated at close interpersonal distances. On the omnibus ANOVA (see Appendix K for all summary ANOVA tables for Study Two) the main effect for i nformation was highly significant, F (2, 36) = 1 1 .03, p< .001 . This showed that subjects i n the spatial confi rmation condition expected to be seated at significantly closer distances than the control g roup lending support to the effectiveness of the manipulation. Further, on a specific contrast the difference between spatial confi rmation and spatial disconfi rmation conditions was also significant, F (1 ,36) = 7. 1 0 , P < .025. This resu lt confirms hypothesis 1 . 1 which predicted that subjects would use avai lable information about spatial concerns in the development of spatial expectations. It therefore appears that the i nformation manipu lation had a significant impact on subjects spatial expectations. D IFFICUL TV OF ANAGRAMS I t was important to show that practice anagrams were simi larly easy for subjects 1 50 i n al l condit ions, and that the test anagrams were in fact difficult and easy for the respective conditions. Table 6 sets out the group means for the average number of practice anagrams correct. TABLE 6 MEANS AND STANDARD DEVIATIONS FOR NUMBER OF PRACTICE ANAGRAMS CORRECT PER CONDITION INFORMATION DIFFICULTY EASY DIFFICULT CONTROL 2.50 2.70 0.32 0.30 SPATIAL CONFIRM. 2 .60 2.57 0 . 1 5 0.42 SPATIAL D ISCONFIRM. 2.62 2.75 0.36 0.28 Note : each mean is based on group means, with each of these cells representi ng eight such g roups. These means look very simi lar. Two things are clear from Table 6 . Fi rstly, the practice anag rams were easy, as can be seen from the magnitude of the means (maximum mean value = 3). Furthermore, 71 .2% (1 71 subjects) got three out of three anagrams correct. A further 21 .3% (51 subjects) were successful with two anagrams, whi le 6.3% (1 5 subjects) were successfu l with only one anagram. The remaining 1 .2% (three subjects) fai led on all three anagrams. The fact that over 90% of subjects were correct on either two or three of the three practice anagrams testifies to thei r ease. 1 51 Secondly , the simi larity among the means suggested that no differences existed among the conditions. In fact, an overall ANOVA indicated ( F< 1 ) that this was the case. Not only were the practice anagrams easy, but they were of simi lar ease for subjects in all conditions. Note that for the analysis of variance of most of the dependent variables for Study Two sex was included in the analysis as a contro l factor (Keppe l , 1 982). The two exceptions to this were for both practice and test anagrams where this procedure was not necessary. I n the case of the test anagrams it was important to demonstrate that those subjects provided with easy anagrams did perform in accordance with the ease of the task, and similarly that the difficult anagrams were of sufficient difficulty. Table 7 s hows that with difficu lt anagrams on average most subjects on ly got about two out of a possible 1 0 correct. I n the case of the subjects g iven easy anagrams this figure was closer to bei ng eight out of 1 0. An ANOVA revealed a highly significant main effect, F ( 1 , 42) = 61 0.88, P < .001 , for the differences between easy and difficu lt anagrams. This result confirms hypothesis 7. 1 which suggested that subjects g iven easy anagrams wou ld outperform those subjects g iven difficult anagrams. These resu lts indicate that subjects' nonspatial expectations were confirmed in the case of the easy anagrams and disconfi rmed in the case of the difficult anagrams. TABLE 7 MEANS AND STANDARD DEVIATIONS FOR NUMBER OF TEST ANAGRAMS CORRECT PER CONDITION INFORMATION DIFFICUL TV CONTROL SPATIAL CONFIRM. SPATIAL DISCONF IRM. EASY DIFFICULT 7.92 2 . 1 0 0.57 0.67 7.67 1 .97 0.90 0.85 7.35 1 .67 0.95 0 .91 Note : each mean is based on group means, with each of these cel ls representi ng eight such g roups. In summary, there appeared to be some doubt about the effectiveness of the distance manipulation (although this was examined indirectly) , whi le both the 1 52 information and difficu lty of anagrams manipu lations were effective. The fol lowing section wi l l examine specific hypotheses beginning with those related to the issue of sex of subject . SEX I t was predicted that women wou ld report fee ling less crowded than men (hypothesis 8. 1 ) , and that women would report feel ing less annoyed than men (hypothesis 8.2) . As expected, women on the CROWDED scale (mean = 1 4.34) reported feel ing less crowded than men (mean = 1 6 .03) . This difference was significant, F ( 1 , 36) = 5.07, P = .031 . Simi larly, on the ANNOYANCE scale women reported feel ing less annoyed (mean = 5.80) than their male counterparts (mean = 8.25). Again , the difference between the means was significant, F ( 1 , 36) = 5.46, p = .025. 1 53 I n view of the clear support for these two hypotheses it was decided to examine the resu lts for both sexes separately for a l l hypotheses which related to either crowding or annoyance. CROWDING I t was predicted that subjects whose spatial expectations were confi rmed would report reduced perceptions of crowding compared to control subjects (hypothesis 2 . 1 ) . For women, the means shown in Table 8 were in the expected direction but the p lanned comparison revealed no significant effect, F ( 1 ,36) = 2.30 , P < .20. I nteresti ngly, for male subjects the means were in the opposite di rect ion to that predicted. This suggests that for males accurate i nformation failed to reduce subjects' perceptions of crowding. If anything the means were suggestive of increased crowding for male subjects who were exposed to accurate information. TABLE 8 MEANS AND STANDARD DEVIATIONS FOR RATINGS OF CROWDING AS A FUNCTION OF SEX AND INFORMATION INFORMATION SEX WOMEN MEN CONTROL 1 4.88 1 5. 50 2.25 0.99 SPATIAL CONFIRM. 1 3.48 1 8.00 2.72 3. 1 1 SPATIAL DISCONFIRM. 1 4.65 1 4.60 1 .31 1 .84 1 54 Note : for women each mean was based on 1 2 g roups, and for men each mean was based on four g roups. The second crowding-related prediction was that disconfi rmed spatial expectations would i ncrease perceptions of crowding for subjects when contrasted with contro l subjects (hypothesis 2 .2) . I t is apparent from Table 8 that there is no support for this hypothesis. For both male and female subjects the means were in the opposite direction to that predicted. Hypothesis 2.3 examined the effect of disconfi rmed nonspatial expectations on perceptions of crowding . It was anticipated that disconfirmed nonspatial expectations (difficult anagrams) wou ld serve to increase subjects' perceptions of crowding compared to subjects whose nonspatial expectations were confirmed. TABLE 9 MEANS AND STANDARD DEV IATIONS FOR RATINGS OF CROWDING AS A FUNCTION OF SEX AND TASK DIFFICULTY DIFFICUL TV SEX EASY D IFFICULT WOMEN MEN 1 3.58 2.29 1 5. 1 0 1 .90 1 4.70 2.31 1 7.37 1 .65 Note : means for women are based on 1 8 groups and those for men on six g roups. As can be seen from Table 9 the disconfi rming of nonspatial expectations did 1 55 i ncrease subjects' perceptions of crowding . For women the means were in the predicted di rection and on the basis of a planned comparison proved to be significant, F ( 1 ,36) = 4.08, P < . 05. For men the means were in the predicted direction and the difference between the means also was significant, F ( 1 ,36) = 4. 1 9 , P < .05. ANNOYANCE . Each of the three hypotheses related to crowding were paralle led by three hypotheses related to the extent to which subjects were annoyed or bothered. 1 56 These hypotheses examined the degree to which the experience of crowding and the hypothesised variables, which influence that experience , were seen by subjects to have an aversive e lement. Hypothesis 4. 1 predicted that spatial confirmation would reduce annoyance when compared to contro l subjects. As Table 1 0 indicates, annoyance leve ls were certai n ly not reduced for spatial confirmation conditions for either female or male subjects. TABLE 1 0 MEANS AND STANDARD DEVIATIONS FOR RATINGS OF ANNOYANCE AS A FUNCTION OF SEX AND INFORMATION INFORMATION SEX WOMEN MEN CONTROL 5. 1 0 6.55 1 . 1 7 4.60 SPATIAL CONFIRM. 5 .73 1 0.40 3. 1 2 5.66 SPATIAL DISCONFIRM. 6 .57 7.80 2 .43 0.28 Note : means for women are based on means of 1 2 groups whi le those for men are based on four g roups. Hypothesis 4.2 predicted that spatial disconfi rmation would resu lt in increased annoyance for subjects when contrasted with contro l subjects. As Table 1 0 1 57 shows, the predicted increases were apparent for both sexes. P lanned comparisons revealed the comparative increase for women , F ( 1 ,36) = 1 .3 1 , n .s . , and men , F (1 ,36) < 1 , was not sufficiently large to attain significance. The third hypothesis (4.3) re lating to annoyance considered the effect of disconfi rmed nonspatial expectations on perceptions of annoyance. It was expected that nonspatial disconfirmation (difficult anagrams) wou ld increase subjects' annoyance at being crowded compared to subjects whose nonspatial expectations were confirmed (easy anagrams). Table 1 1 detai ls the mean annoyance ratings for both women and men as a function of task difficu lty. TABLE 1 1 MEANS AND STANDARD DEVIATIONS FOR ANNOYANCE RATINGS AS A FUNCTION OF SEX AND TASK DIFFICULTY DIFFICULTY SEX WOMEN MEN EASY DIFFICULT 5.28 1 .73 6.32 3. 1 8 7.43 5.04 9.07 1 .98 Note : means for women are based on 1 8 groups and those for men are based on 6 groups. As is clear from Table 1 1 , nonspatial disconfi rmation did result in comparative 1 58 increases in subjects' ratings of annoyance. These i ncreases occurred for both sexes but in neither case was the i ncrease significant. For women, F ( 1 ,36) < 1 , and for men, F (1 ,36) < 1 . AROUSAL The hypotheses re lating to arousal predicted no sign ificant differences between subjects whose spatial expectations were confi rmed and subjects whose spatial expectations were disconfi rmed (hypothesis 3. 1 ) . S imi larly, it was anticipated that nonspatial disconfi rmation would not increase subjects' arousal in comparison with subjects whose nonspatial expectations had been confirmed (hypothesis 3.2). For each of these hypotheses two measures of arousal were taken (AROUSAL and SOMATIC). Table 1 2 presents the means for arousal , whi le Table 1 3 presents the means for the somatic measure of arousal. The uniformity of the means for both measures of arousal presented in Tables 1 2 and 1 3 suggest strong support for both hypotheses 3 . 1 and 3.2. On both arousal scales no significant main effects or interactions were noted. On the overal l ANOVA for arousal the main effect for information was not significant, F < 1 . On the overall ANOVA for somatic arousal the main effect for information also fai led to reach significance, F < 1 . TABLE 1 2 MEANS AND STANDARD DEVIATIONS FOR AROUSAL AS A FUNCTION OF INFORMATION AND TASK DIFFICUL TV I NFORMATION DIFFICULTY EASY DIFFICULT 1 59 MEAN TOTAL CONTROL 34.92 35.85 35.39 1 .56 2.50 SPATIAL CONFIRM. 35.92 33.92 34.92 1 .98 1 .45 SPATIAL DISCONF IRM. 36.20 35.00 35.60 2. 1 0 2.34 MEAN TOTAL 35.68 34.92 Note : means are all based on eight groups per cel l . Simi lar resu lts were obtained for task difficulty. On the ANOVA for arousal the main effect for task difficulty had little appreciable effect, F (1 , 36) = 1 .70, p =.20. Also on the overall ANOVA for somatic arousal the main effect for task difficu lty fai led to reach significance , F < 1 . Therefore , on both measures of arousal neither of the manipulated variables had any significant impact on subjects' reported levels of arousal. TABLE 1 3 MEANS AND STANDARD DEVIATIONS FOR SOMATIC AS A FUNCTION OF INFORMATION AND TASK DIFF ICUL TV I NFORMATION DIFFICULTY EASY DIFFICULT 1 60 MEAN TOTAL CONTROL 26.57 27.30 26.94 1 . 1 3 1 . 1 6 SPATIAL CONFIRM. 27.02 27.37 27.20 3.05 1 .63 SPATIAL DISCONFIRM. 27. 1 0 26.55 26.82 1 .63 1 .40 MEAN TOTAL 26.90 27.07 Note : means are al l based on eight g roups per cel l . I NTERPERSONAL RELATIONSH IPS Interpersonal relationships were i ndexed by ratings of punitiveness, perceived aggressiveness of others and l iking for others. The general expectation (hypothesis 5. 1 ). was that interpersonal re lationships wou ld suffer in the face of both spatial and nonspatial disconfi rmation . 1 61 PUN ITIVENESS Table 1 4 shows the respective means on the Love - Punishment scale which indexes punitiveness. H igher scores indicate g reater levels of punishment. It was expected that punitiveness would be increased as a result of both spatial and n onspatial disconfi rmation . The omnibus ANOVA showed no main effect for information, F < 1 , i ndicating that spatial disconfi rmation fai led to sign ificantly increase subjects punitiveness even though the means were in the expected direction . TABLE 1 4 MEANS AND STANDARD DEVIATIONS FOR THE LOVE-PUNISHMENT SCALE AS A FUNCTION OF INFORMATION AND TASK DIFFICULTY INFORMATION DIFFICULTY EASY DIFFICULT INFORMATION MEAN TOTAL CONTROL 2.75 3.25 3.00 0 .71 0 .71 SPATIAL CONFIRM. 2.38 3.38 2.88 0.52 0 .74 SPATIAL DISCONFIRM. 3.25 2.88 3.06 0.71 0.64 MEAN TOTAL - 2.79 3. 1 7 Note : means are al l based on eight groups per ce l l . 1 62 With respect to nonspatial disconfirmation , Table 1 4 shows that subjects provided with difficult anagrams were more punitive (mean = 3. 1 7) than subjects given easy anagrams (mean = 2.79). On the overall ANOVA this difference was significant, F ( 1 , 36) = 4. 1 9 , p = 0.048. There was also a significant interaction , F (2, 36) = 4.81 , P = 0.01 4, between information and task difficulty such that subjects g iven easy anagrams but disconfi rmed spatial expectations were more punitive than subjects who had easy anagrams but confirmed spatial expectations. Less punitive subjects were l ikely to be those who were given easy anag rams but disconfirmed spatial expectations, or difficult anagrams but confi rmed spatial expectations. AGGRESSIVENESS It was anticipated that both spatial and nonspatial disconfirmation wou ld resu lt in increased judgments for aggressiveness of other group members. The means for subjects' rati ngs of aggressiveness of others in their group are shown in Table 1 4. These means are based on the total aggressiveness rating of the other fou r members of each group; these five totals have then been averaged. The overal l ANOVA indicated no main effect for i nformation , F < 1 , suggesting that spatial disconfirmation did not resu lt in increased ratings of aggressiveness for fel low g roup members. A visual examination of the direction of the means shown in Table 1 5 confirms this conclusion. TABLE 1 5 MEANS AND STANDARD DEVIATIONS FOR AGGRESSIVENESS AS A FUNCTION OF TASK DIFF ICULTY AND INFORMATION INFORMATION DIFFICULTY EASY DIFFICULT 1 63 INFORMATION TOTAL MEAN CONTROL 25.30 27.97 26.64 2 .81 4 .71 SPATIAL CONFIRM. 24.02 26.42 25.22 2.63 5.09 SPATIAL D ISCONFIRM. 2 1 .60 27.67 24.64 3. 1 8 2.88 TOTAL MEAN 23.64 27.36 Note : means based on eight groups per cel l . The ANOVA, however, revealed a main effect for task difficu lty, F ( 1 , 36) = 7.99, p = 0.008. Subjects who had their nonspatial expectations disconfirmed (exposed to difficult anag rams) rated their fel low g roup members as being significantly more aggressive compared to those subjects given easy anag rams. LIKING FOR OTHERS It was anticipated that disconfirmation of both spatial and nonspatial expectations would resu lt in decreased l iking for other people in the group. There was no main effect for either information , F (2, 36) = 1 .86, P = 0 . 1 71 ; or for task d ifficu lty, F < 1 . These results indicate that the disconfi rmations which were manipulated in this study had no significant effect on subjects' ratings of the extent of their l iking for each other. 1 64 On the variable of l iking for others there was an unanticipated main effect for sex of subject, F ( 1 , 36) = 6 .61 , P = 0.01 4. This meant that women subjects l iked their fel low g roup members (mean = 6.37) more than the ·men subjects l iked their fe l low participants (mean = 5.78). In summary, on two of the three measures used, disconfi rmation of nonspatial expectations resu lted in some disruption to interpersonal re lationships. Disconfi rmed spatial expectations fai led to have a significant effect on i nterpersonal relationships. AFFECT The hypotheses relating to affect suggested that confi rmation of spatial expectations wou ld reduce negative affect. On each of the three main scales of the MAACL-R (anxiety, hosti l ity and depression) no main effects were evident. On the anxiety scale , whi le there was no main effect, there was a significant interaction , F(2, 36) = 6.28, P = 0.005. This interaction , between spatial information and sex of subject, suggested that providing men with accurate 1 65 information served to increase their anxiety while for women this same information reduced anxiety. PERFORMANCE The study had an exploratory i nterest in the effect of confi rmed and disconfi rmed spatial and nonspatial expectations on task performance. The task concerned involved the group compiling as large a l ist as possible of smaller words derived from a larger word. TABLE 1 6 MEANS AND STANDARD DEVIATIONS FOR GROUP PERFORMANCE AS A FUNCTION OF INFORMATION AND TASK D IFFICULTY INFORMATION DIFFICULTY EASY DIFFICULT INFORMATION TOTAL MEAN CONTROL 63.75 72.88 68.31 1 0.21 1 0.26 SPATIAL CONFIRM. 63.63 65.50 64.56 9 .05 5. 1 6 SPATIAL DISCONFIRM. 59.38 65.63 62.50 6 .68 9 .33 TOTAL MEAN . 62.25 68.00 Note : means based on eight g roups per ce l l . 1 66 On the overall ANOVA there was no main effect for i nformation but there was for task difficu lty, F ( 1 , 36) = 4.92, P = 0.033. Subjects whose nonspatial expectations were disconfi rmed and who were exposed to difficu lt anagrams (mean = 68.00) outperformed subjects whose nonspatial expectations were confi rmed (mean = 62.25). This result is the opposite of that predicted in hypothesis 7.3. Having accurate nonspatial information fai led to enhance performance in any way, with the better performance coming from the subjects whose nonspatial expectations were disconfi rmed. 1 67 C HAPTER TEN STU DY TWO DISCUSSION MEASUREMENT OF AROUSAL AND CROWDING In order to overcome some possible deficiencies in the measurement of arousal and crowding, this study employed two specific scales to measure each of these key concepts. The alpha coefficients for these scales indicate that these measures were internal ly consistent. The indices of i nter-item rel iabil ity compare favourably with previously published re liabi l ity measures for these scales. Of the two scales measuring arousal (SOMATIC & AROUSAL) on ly the fi rst had previously been subjected to a considerat ion of internal consistency. In comparison with previous research by Aiel lo et at . ( 1 977) and Nicosia et al . ( 1 979) the current study compared very wel l with alpha = .83 . In the case of the arousal scale, there were no previously published studies with which to compare. However, even relative to the alpha coefficient for the somatic scale, the alpha leve l of .64 for the arousal scale looks encourag ing , and is suggestive of good internal consistency. Of the two crowding scales on.ly the ANNOYANCE scale had any previously reported i nformation about its reliabi lity. The resu lts from the current study 1 68 compare very favourably (alpha = .90) with the previously published research by Nicosia et al. ( 1 979). The CROWDED scale has no previously published data with which to compare but in a relative sense looked impressive (alpha = .85). In sum, the four scales each demonstrated a high level of internal consistency. Anastasi (1 982) notes that good internal consistency impl ies the scale is homogeneous. Such homogeneity is important for unambiguous i nterpretation of scores. MAN IPULATION CHECKS Manipulation checks were conducted in three areas. The fi rst of these considered the effectiveness of the distance manipu lation. The second looked at the effect the differing information conditions had on subjects' spatial expectations. The th ird manipulation check was concerned with establ ishing that subjects performed differentially on the test anagrams after receiving practice anagrams of uniform ease. I nterpersonal Distance The standard procedure in assessing the effectiveness of distance manipu lations is to compare ratings of subjects seated at close distance with ratings of subjects seated at far distance. Without exception , the literature reveals that close 1 69 subjects report higher levels of crowding, on whatever measures of crowding have been taken , than far subjects. Reduced interpersonal distance resu lts in increased reports of crowding. The consistency of this finding was the reason no far i nterpersonal distance conditions were i ncluded in the current study. On the basis of a large body of supporting evidence it was assumed that a simi lar "close" manipu lation wou ld be effective. Whi le this assumption seemed reasonable it was only an assumption and it therefore seemed prudent to examine any information which may have a bearing on it. For the reason mentioned above there was no comparative i nformation avai lable to determine whether the subjects in this study experienced spatial invasion re lative to subjects seated at far distances. There was, however, some descriptive i nformation available in the form of two questions which subjects responded to in an absol ute sense. Subjects were asked whether or not their personal space had been invaded, and if so, to what extent. Only 60% of subjects indicated that their personal space had been i nvaded. The su rprising aspect of this is that , given the closeness of the other subjects (they sat with adjacent knees touching) , not more subjects felt as though their personal space had been i nvaded. This resu lt is noteworthy in view of the definitions of crowding which are predicated on the belief that personal space i nvasions are necessary for the experience of crowding. While this view has been challenged by the present result , it may be as wel l to remember that for the majority of subjects the operationalisation of crowding w as effective in inducing an invasion of personal space. That is, 60% of subjects said that they felt as though thei r personal space had been invaded. 1 70 Nevertheless 40% of subjects fai led to report an invasion of personal space under conditions which had been specifically designed to induce such invasions. However, this result should be interpreted with caution because it is unusual to measure spatial invasion in this fashion. The preferred methods being naturalistic observation or the stop-distance technique (Hayduk, 1 983). Neither of these methods were appropriate in the present situation , so a di rect question method was employed. This measure is reasonable provided results are interpreted with due caution. Another point to consider is the wording of the question. Subjects were asked about an "invasion" of personal space. While a sizable proportion of subjects said "no" to that question, they may have agreed to something less strongly worded. More subjects, for example, may have been prepared to endorse the notion of an "infringement" of their personal space. Subjects who acknowledged a spatial invasion were asked about the severity of the i nvasion. The results i ndicate that the vast majority of subjects experienci ng a spatial invasion reported this as being of only slight magnitude. However, there are no particular theoretical expectatfons about the severity of spatial i nvasions necessary for the experience of crowding to occur. In other words, the min imal invasions reported may well have been sufficient to induce feel ings of 1 71 crowding . Again , i f this i nformation is taken at face value, i t suggests that even minimal invasions of personal space may have profound consequences for individuals' experiences of crowding. Perhaps the most remarkable aspect is that not more subjects acknowledged the severity of the personal space invasion. Each subject sat in a confined seating arrangement with knees touching the knees of two other subjects who were unknown to the individual. Given the ethical constraints it is difficult to imagine a situation where subjects cou ld have their personal space more invaded. Yet curiously, subjects described this as being on ly a sl ight spatial invasion. By way of explanation i t may be helpful to consider the natu re of the experimental situation. In this study, subjects were provided with advance warning ( in two of the three information conditions) that distance between subjects was a variable to be manipu lated. And in one of these cases subjects were told to expect close interpersonal distances. Perhaps it was a response to the pre-exposure information which accounted for the restricted endorsements of the nature of the spatial invasions. Simi larly, i nforming subjects about the l ikel ihood of spatial manipulations may have been responsible for the large number of subjects who failed to report an invasion of their personal space. However, these conjectures are not supported by the fact that 64% of subjects in the control condition acknowledged a spatial invasion compared to 50% and 65% of subjects in the two conditions which were alerted to the importance of 1 72 distance as a variable. Anothe r reason for the l imited reports of spatial i nvasion may stem from the fact that although subjects sat very close, they were able to achieve the set tasks without any disruption or interference from others. I n short, the lack of behavioural i nterference may have reduced the extent to which subjects felt their space had been i nvaded. Certainly, the various behavioural constraint models of crowdi ng would argue for this possibi lity. At issue here is the question of whether subjects' reports of spatial invasions are dependent on the invasion itself or on the aversive consequences which may accompany the i nvasion. A further point worth considering is that these questions were asked within the context of an experimental setting. The unique natu re of the setti ng may have contributed to the extent to which subjects were prepared to expose themselves to unusual conditions, and then subsequently regard these conditions as not especial ly u nusual. This degree of experimental compl iance has previously been documented (Orne, 1 962), and wou ld suggest that there may be few limits to the extent to which subjects wi l l go in order to appease experimental demands. The absence of comparative information makes it impossible to draw fi rm conclusions abOut the effectiveness of the distance manipulation. The assumption that the distance manipu lation wou ld result in the personal space invasions necessary to i nduce feel ings of crowding has received reserved support . On the other hand, the results also suggest that perceptions of crowding 1 73 may not be contingent upon spatial invasion. I nformation Three i nformation conditions were included in the design in o rder to influence subjects' expectations about the distance they wou ld be seated from other subjects. Subjects were led to expect close seating, normal seating , or were provided with no information about seating. They were subsequent ly asked how close they had expected to be seated. As expected , those led to expect close interpersonal distances reported they had expected to be seated closer than either contro l subjects, or subjects who had been misinformed to expect a normal interpersonal distance. This confirmed hypothesis 1 . 1 which predicted subjects would use avai lable spatial information i n developi ng spatial expectations. It can therefore be concluded that this information had a significant impact on expectations about closeness to others. This suggests spatial information is l ikely to be appraised, and this appraisal wi l l have a bearing on subsequent spatial expectations. This result supports the i nformation expectancy model in consolidati ng the l ink between information and spatial expectations. Anag rams The manipu lation checks involvi ng anagrams were completed in o rder to demonstrate the ease of the practice anagrams and the difficu lty of the test anagrams. This was important as these differences represented the 1 74 operational isation of confi rmation and disconfi rmation of nonspatial expectations. Relevant to the ease of the practice anagrams was the fact that over 90% of the subjects g ot either two or three out of three correct. A majority of subjects (71 %) got al l three anagrams correct. Not on ly were the practice anagrams easy, but they were of un iform ease for subjects in all conditions. The resu lts indicate that the mean number of anagrams correct ranged from 2.5 to 2 .75 for the eight conditions, with an overal l mean of 2.62. There was no significant difference between these means which supports the conclusion that they were of simi lar ease for subjects in all conditions. In the case of the test anagrams, it was important to show that the easy anagrams were in fact easy and that the difficu lt anagrams were l ikewise difficu lt . Clear d ifferences emerged with subjects given easy anagrams outperforming those subjects provided with difficu lt anagrams. This result supports hypothesis 7. 1 . It seems safe to conclude that the manipu lation designed to confi rm and disconfirm subjects' nonspatial expectations was effective. SEX The issue of sex of subject has resu lted in equivocal resu lts within the crowding literature (Hayduk, 1 983). One theme which recently has received some support (Nicosia et aI . , 1 979) is that men respond more negatively to being crowded than women. Thus, it was predicted (hypothesis 8. 1 ) that women would report feel ing less crowded than men, and (hypothesis 8 .2) that they wou ld fee l less annoyed 1 75 than their male counterparts. The results supported these two predictions. On the CROWDED scale , women were significantly less crowded than men. Simi larly a sex difference emerged on the ANNOYANCE scale. These resu lts provide further support for the view that there are differences i n the way in which men and women respond to crowded situations. The resu lts lend support to Nicosia et al . 's ( 1 979) conclusion that men respond more negatively to being crowded. The issue of sex of subject was considered as an explanatory factor in the fai lure of the fi rst study to induce arousal in crowded subjects. This possibi lity arose because if women adapt better in crowded conditions, then a study based on women subjects may be open to the criticism that arousal was unl ike ly to be induced due to the adaptive nature of the subjects under those conditions. If th is had been the case then sex differences on the arousal measures should have been evident in the second study. Yet on both measures of arousal no main effect for sex was apparent showing that subjects of both sex were equal ly aroused, or nonaroused, by the experience of crowding. CROWDING AND ANNOYANCE One of the key predictions of the i nformation expectancy model was that confirmation of spatial expectations wou ld reduce the extent to which subjects 1 76 reported feel ing both crowded (hypothesis 2. 1 ) , and annoyed (hypothesis 4. 1 ) . As a result of the sex differences found on both crowding measures these results were considered separate ly for men and women. In the case of women subjects on the crowding measure there was a nonsignificant decrease i n crowding for those subjects whose spatial expectations were confi rmed. However, this marginal reduction in crowding was not supported by any reduction in the level of reported annoyance. For men, spatial confi rmation fai led to produce the predicted reductions in either crowding or annoyance. For subjects of both sexes, the provision of accurate i nformation fai led to reduce perceptions of crowding or annoyance. This may have been due to the ineffectiveness of the distance manipulation which may have fai led to induce strong feel ings of crowding and annoyance. Or it may have been a resu lt of the fact that subjects were able to function adequately and complete tasks without interference by the presence of the other subjects. On at least one of the tasks the cooperative requirements of the task would have i ncreased the desirabi l ity of having others present. Whatever the reason, this result represents a serious fai lure to support one of the main predictions of the i nformation expectancy model . I t was expected that disconfirmed spatial expectations ( inaccurate spatial information) wou ld result in increased perceptions of crowding (hypothesis 2.2) and annoyance (hypothesis 4.2) in comparison to subjects who were provided with no spatial i nformation. The results fai led to support the fi rst of these 1 77 hypotheses for both men and women. There was no increase in subjects' perceptions of crowding fol lowing inaccurate spatial information. The lack of support for the second of these hypotheses was almost as complete. On the annoyance scale the increases for both men and women were not significant. As for the provision of accurate i nformation , it would seem that inaccurate spatial information also fai led to have any real i nf luence on the way in which subjects subsequently perceived the extent to which they were crowded or annoyed. One of the cornerstones of the information expectancy model is that spatial ly relevant information wou ld be used i n establ ishing spatial expectations and confi rmation or disconfi rmation of these expectations wou ld influence the way i n which the individual perceived the experience of crowding. It seems that spatially relevant information is likely to be used in establishi ng spatial expectations but the accuracy of that information wi l l have little effect on the experience, in terms of seeing it as more crowded or more annoying. This may imply that expectations are rather more malleable t han was previously assumed. With in the context of experimental crowding , however, the uti lity of pre-exposure information now seems questionable. This fai lure to demonstrate the efficacy of accurate information has a number of impl ications for some models of crowding. Epstein and Karlin's (1 975) normative expectations model holds that environmental cognition wi l l operate to exacerbate the experience of crowding when there is a m ismatch between what is expected and what is subsequently experienced. Indeed, Study Two attempted to engineer 1 78 such a mismatch with the result that information and expectations mattered litt le i n i nfluencing the subsequent experience. An important consideration is the fact that these expectations were set i n the context of an experiment where wider normative expectations may operate. Altman's ( 1 975) privacy control model is perhaps less relevant given that all subjects knew that they would be participating with a smal l g roup of others. However, the privacy contro l model is based on the assumption that crowding wi l l be i ncreased by any loss of psycho logical contro l . One perspective on the current study is that it attempted to i ncrease subjects' predictive control by providing accurate pre-exposure i nformation . This measure proved to be ineffective in mediating subjects' perceptions of crowding. This suggests that the i nformation is not especial ly usefu l as a control mechanism or, alternatively, that control is less important when the crowding experience contains no obviously aversive aspects. One contradiction deserving comment is that accurate information in the fi rst study did result i n a reduction i n crowding , but this was not the case i n the second study. The pre-exposure information i n the fi rst study consisted of two aspects, i nformation that the variable of experimental i nterest was the distance between people and that they may experience negative effects as a consequence. Study Two provided accurate i nformation to subjects by tel l ing them that the study was looking at the effects of distance on behaviour (sim i lar to the information provided in the first study) . Un l ike the fi rst study, however, they were not informed of possible negative effects of sitting so close to others. One reason for this omission was the serious inconsistency it wou ld have 1 79 presented for subjects who were led to expect "normal" seating. Thus, with the benefit of hindsight it seems likely that the information about possible negative effects of crowding may have been important. It is also possible that anticipatory negative expectations if not confirmed by the experience lead to a cognitive reinterpretation of the experience. Subjects may th ink that being crowded is usually associated with negative consequences and the fai lure to experience these may lead them to believe they were not feel ing crowded, o r as crowded as they had been led to expect. On the basis of previous research it seemed as though the important dimension for pre-exposure i nformation was accuracy. Given this constraint, it would have seemed questionable to tell subjects that they cou ld expect to feel "stressed and uncomfortable" when , on the basis of the fi rst study, they had not reported feel ing stressed at al l . Also, the inclusion of such information is open to the criticism that it would serve as a demand characteristic. The final set of hypotheses relating to crowding and annoyance predicted that disconfirmed nonspatial expectations (difficult anagrams) wou ld increase perceptions of crowding (hypothesis 2.3) and increase the level of annoyance (hypothesis 4.3) for both men and women. Hypothesis 2.3 was supported with disconfirmed no"nspatial expectations resu lting in increased perceptions of crowding. Hypothesis 4.3 was not supported with disconfirmed nonspatial expectations fai l ing to increase annoyance. The resu lts complimented those achieved for ratings of crowding. It appears that providing subjects with a more 1 80 d ifficu lt task than they had been led to expect resu lted in increased perceptions of crowding. This finding is simi lar to the results reported by Gochman and Keating ( 1 980) who found that factors i ndependent of density can affect attributions to croWding. Thei r resu lts were obtained in fie ld and laboratory setti ngs where disconfi rmed expectations (positive and negative) affected attributions to crowding. Subjects were p repared to blame density re lated factors (crowding) for disconfi rmed performance expectations. Subjects whose performance was inconsistent with what they had expected attributed this unexpected resu lt to "the number of people in the room". While Gochman and Keating highl ight the nature of the relationship between nonspatial and spatial factors, it is a finding which is less di rect than the present one as it considered attributions to crowding rather than the extent to which subjects reported feeling crowded. Unexpectedly increasing task difficu lty in a crowded situation resulted i n increased perceptions o f crowding. This suggests nonspatial information may have a more potent role to play in the experience of crowding than that of spatial information . Inaccurate nonspatial information leading to a disconfi rmation of nonspatial expectations increased subjects' perceptions of crowding whi le inaccurate spatial information fai led to have a simi lar effect. In other words, an unexpectedly difficult task increased perceptions of crowding while unexpectedly close distances did not increase perceptions of crowding. 1 81 To summarise, it seems possible to steer spatial expectations via i nformation about the e nsuing spatial arrangements. However, the confirmation or disconfi rmation of these expectations has l ittle impact on subsequent perceptions of how crowded or annoyed subjects feel . Fai lure to find support for one of the key predictions of the i nformation expectancy model is seen as a serious flaw in the model . Within the context of experimental crowding these results raise questions about the importance of pre-exposure i nformation. The results for disconfi rmation of nonspatial expectations were more in l ine with predictions. Whi le disconfirmed nonspatial expectations i ncreased perceptions of crowding , perceptions of annoyance were u naffected. Providing subjects with a more difficult task than expected resulted i n subjects feel ing more crowded than subjects whose nonspatial expectations were confirmed. This suggests that perceptions of crowding are affected by factors not usually considered to be related to the notion of crowding. Disconfi rmation of nonspatial expectations should be added to the factors, such as size of room, number of people and distance between them, which have a beari ng on the experience of crowding. Although th is resu lt was predicted , i t was predicted in conjunction with simi lar effects for the disconfi rmation of spatial expectations. On its own , it suggests that i nformatio n which is not di rectly relevant to spatial concerns may be of importance in determin ing the extent to which subjects feel crowded i n experimental settings. Clearly, the activity o ne engages in whi le in crowded conditions is of importance. 1 82 AROUSAL Extending the view from Study One that arousal need not necessarily be i nvolved in the experience of crowding , it was predicted that confi rmation o r otherwise of both spatial and nonspatial expectations wou ld not i nfluence arousal levels. There is no way of knowing whether or not subjects were comparatively u naroused (as may have been established by i ncluding a far interpersonal distance condit ion). Hypothesis 3. 1 predicted that spatial disconfirmation would not increase arousal on either of the arousal measures. On both measures, th is hypothesis was supported. Simi larly disconfirmed nonspatial expectations showed no effect when compared to confi rmation of nonspatial expectations (hypothesis 3.2) . Therefore , it wou ld appear that neither of the manipu lated variables (information or task difficu lty) had any significant impact on subjects' reported levels of arousal. INTERPERSONAL RELATIONSHIPS The general expectation was that i nterpersonal re lationships wou ld suffer i n the face of disconfi rmed spatial and nonspatial expectations. Three i ndices of interpersonal relationships were taken , a punitiveness measure, a measure of aggressiveness of g roup members , and a measure of l iking for g roup members. On the Love-Punishment scale i t was found that spatial disconfi rmation fai led to significantly increase punitiveness. However, subjects whose nonspatial 1 83 expectations were disconfirmed were sign ificantly more punitive than subjects whose nonspatial expectations were confi rmed. Exposing crowded subjects to a series of difficult tasks after leading them to expect easy tasks increased their pun itiveness. This may reflect anger or disappoi ntment at having been misled as to the level of performance they could have expected. This i n turn may have manifested itself in advocating more punishment. There was also a significant interaction , the interpretation of which is problematic due to the lack of any clear theoretical relationship between the two independent variables. Subjects whose spatial expectations were disconfi rmed ( inaccurate i nformation) but whose nonspatial expectations were confi rmed (easy anagrams) were significantly more pun itive than subjects whose nonspatial expectations were confi rmed and who received either accurate spatial i nformation or were in the control group. Further, for the nonspatial disconfi rmation group those subjects whose spatial expectations were also disconfirmed were less punitive than subjects g iven difficu lt anagrams who received accurate spatial information or were in the control condition. For perceived aggressiveness of other group members the influence of spatial expectations was not significant. However, the disconfi rmation of nonspatial expectations resu lted in h igher ratings of the perceived level of aggression displayed by fel low group members than when nonspatial expectations were confi rmed. The th i rd index of interpersonal relations was reported liking of fe l low group members. For disconfirmed spatial and nonspatial expectations there was no main effect. Such disconfi rmations had no influence on the degree to which subjects reported l iking each other. Neither information nor difficulty of task influenced the extent to which the subjects l iked each other whi le in close 1 84 contact with each other. This is consistent with impressions which were gained at the debriefing sessions where subjects typically gave the impression that they had got along well with each other i n spite of the cramped nature of their interaction. Disconfirmed spatial expectations fai led to have the predicted negative effect on interpersonal relationships. However, disconfirmed nonspatial expectations, on two of the three measures, resulted in disrupted interpersonal relationships. This result is consistent with a pattern which emerged from the study. This is characterised by no effect for the influence of spatial expectations, but a consistent effect for thwarted nonspatial expectations. Disconfi rmed nonspatial expectations have increased the extent to which subjects felt crowded, and have disrupted the i nterpersonal relationships of crowded subjects. This result was expected on the basis of previous research which indicated that accurate information wou ld decrease the negative aspects associated with crowding. Inaccurate information was expected to have the opposite effect, which proved to be the case. This effect may operate such that inaccurate nonspatial i nformation effectively removes the neutral perceptions associated with others i n the situation and replaces them with more negative perceptions. In some way, which is far from clear, inaccurate information may have the effect of increasing the 1 85 interpersonal stress involved in crowded situations. For l iki ng of others there was an unexpected influence for sex of subject , with women subjects reporting g reater l iki ng for thei r fel low g roup members than men . Th is result , although not expected , is consistent with the previous findings that women subjects reported feel ing less crowded and less annoyed than their male counterparts. This finding also adds to the overall picture which has emerged and which typifies the experience of women in crowded conditions as less negative than the males. AFFECT It was expected that the MAACL-R measures wou ld be sensitive to the confirmation of spatial expectations. The results indicate that this was not the case, with no significant main effects being found. There was, however, an interaction between spatial i nformation and sex of subject on the measure of anxiety. This interaction was interpreted to mean that when men were provided with accurate spatial information this increased their anxiety, while the same information for women resulted in reduced anxiety. Whi le the overall result for affect is rather disappointing , the emergence of yet another sex difference is of interest , especially as this difference is one of a series which have come to l ight in th is study. Not only are women more adaptable i n crowded conditions but they apparently do not experience the same anxiety as men when informed about the experience. 1 86 PERFORMANCE Spatial disconfirmation groups did not perform as wel l as control g roups ; however, this difference was not significant. The disconfirmation of nonspatial expectations did result in those groups performing significantly better than g roups for whom nonspatial expectations were confi rmed. This is a curious resu lt, suggesting possible reactivity on the part of subjects who expected an easy task, on ly to be confronted with a difficu lt task. Fol lowing this, the performance task may have presented an opportunity to re-establish a sense of control or a degree of mastery over the experimental tasks. The superior performance wou ld suggest increased levels of effort. The result for performance stands in contrast to the general effect of nonspatial disconfi rmation, which has been to exacerbate crowding, annoyance and to disrupt interpersonal re lations. If these effects can be regarded as deleterious or negative , then improved performance on the group task i s the only positive outcome of nonspatial disconfirmation. SUGGESTIONS FOR FURTHER RESEARCH The present study has provided i nformation about the internal consistency of scales measuring arousal and crowding. Two suggestions arise out of this. Fi rstly , that future research should use measures which are reliable. Consistency 1 87 of measurement is important, and within the crowding literature has suffered some neg lect. Secondly, having reliable measures is important, but the next step should be the establ ishment of the validity of these measures. This task awaits future researchers. A number of suggestions for further research fol low from the fact that not more subjects reported spatial invasion. The re lationship between personal space and crowding needs further investigation. It has previously been assumed that the two are closely l i nked but the current fi ndings cast doubt on this. Whi le this assumption has not been toppled, i t has been shaken. There are measurement issues which have already been alluded to , which center around the wording of spatial "invasion/ infringement! adjustment". These words denote differing degrees of the spatial relationships between individuals. As suggested , it is possible subjects considered "invasion" to be an overstatement of what they actually experienced withi n the context of this study. It may also help explai n the present rather curious resu lt . As indicated, the information concerning the effect of i nterpersonal distance on personal space lacked comparative data. This must , in the l ight of the resu lts which emerged, be seen as a weakness of the study. It also, however, high lights the need for such comparative information in future studies. The most interesti ng question concerns the comparative differences between subjects at close as opposed to far interpersonal distances , in terms of the reported invasions of personal space. 1 88 This study has contributed to the question of sex differences by supporting Nicosia et al. 's ( 1 979) conclusion that men respond more negatively to crowded situations. These differences represent a chal lenge to future researchers as it wou ld seem that women are less negatively influenced by the closeness of members of the same sex. A suggestion for future research would be to examine these differences further, u ltimately with the aim of understandi ng the reasons for the d ifferences. The possibi lity of uti l ising noninteractive g roups of subjects under crowded condit ions should be considered. The advantage of such a strategy would be the reduction in numbers of subjects requ ired. The high demand on subject numbers makes experimental studies of crowding seem unattractive to prospective researchers. Smal ler numbers of noninteractive subjects may encourage further research in this area. The disadvantage of this move would be the l imits placed on the range of tasks that cou ld be undertaken under these conditions. Other suggestions for future research arise from the finding that inaccurate i nformation about a task which is to be undertaken in crowded conditions, led to i ncreased perceptions of the extent to which subjects felt crowded. This task was a simple anag ram task and it wou ld be instructive to examine the extent to which these resu lts may generalise to other tasks. The role of nonspatial information is an area where further research may be usefu lly di rected. The present study employed an inaccuracy which was almost the exact opposite of that to which subjects had been led to expect. Wi l l less extreme inaccuracies also have a 1 89 simi lar effect? Just how incorrect does nonspatial information need to be before it wi l l have the effect of increasing perceptions of crowding? Is it possible, for examp le , for subjects to tolerate minor inaccuracies i n the pre-task information they receive? Final ly, one area warranting further research is the extent to which accurate spatial i nformation may be more i nfluential when supported by information concern ing how subjects are l ike ly to be affected by the experience of close i nterpersonal distance. In Study One the combination of accurate spatial i nformation and i nformation about the possible negative effects on individuals reduced the impact of crowding. Accurate spatial information in the second study fai led to have the same effect. Some important questions cou ld well be pursued in an attempt to delineate the ro le which information about the possible negative effects has on those subsequent effects , and more widely on the experience of crowding itse lf. This chapter has reviewed and discussed the results of the second study. Measurement issues have been considered , the effectiveness of the manipu lations, and the key results for each of the dependent measures have also been discussed. Finally, a number of suggestions for future research have been put forward. The final section of this chapter wi l l address the implications which these results have for the information expectancy model. 1 90 IMPLICATIONS FOR THE INFORMATION EXPECTANCY MODEL The information expectancy model of crowding was founded on two assumptions. Fi rstly, that the experience of crowding is in itiated by an i nvasion of personal space, and secondly, that spatial information wou ld be appraised and formed into specific expectations. These assumptions led to two central predictions, that confi rmation of spatial expectations wou ld have an amel iorative effect while disconfi rmation wou ld exacerbate the impact of crowding ; and, finally, that nonspatial expectations wou ld (if confirmed or disconfirmed) act in a manner consistent with the role of spatial expectations. The in itiating function of personal space invasions received restrained support. The approach to determining the exact nature of personal space i nvasions was novel and for this reason the restrained degree of support for this idea must be viewed with ci rcumspection . At this stage it is possib ly best to acknowledge support for this aspect of the model , but also to signal that this support was not the wholesale endorsement which may have been expected on the basis of the previous literature. Some i nteresting possibi l ities for further research have emerged and these have al ready been alluded to. The notion of information appraisal leading to spatial expectations received strong support. Subjects do uti lise spatial information and, furthermore , this information is incorporated in spatial expectations. 1 9 1 The i nformation expectancy model makes predictions about the usefu lness or otherwise of spatial information . The resu lts i n this area were not supportive of the role of spatial information . Confi rmation or disconfi rmation of spatial expectations effectively had no impact on subjects perceptions of how crowded or annoyed they felt. This finding is theoretically d isappointing. This is also the most important area, as the information expectancy mode l is centered around the assertion that accurate spatial expectations are usefu l , and the current study casts serious doubt on this view. Final ly , the ro le of nonspatial expectations have been fi rm ly implicated in the experience of crowding. Disconfi rmed nonspatial expectations were predicted to enhance the degree to which subjects reported feeling crowded and this is exactly what happened. I n this respect, the importance of nonspatial information cannot be overstated, apparently being more important than spatial information in determin ing the extent to which subjects wi l l feel crowded. This finding is of particular i nterest as it suggests that nonspatial variables do interact with the experience of crowding . 1 92 CHAPTER ELEVEN SU MMARY AND CONCLUSIONS This thesis has attempted to answer some questions about the nature of human crowding. The experience of crowding is one with which we are al l fami l iar and , if populations continue to increase , may become increasi ngly fami l iar in the future. In spite of the ready avai lability of this experience we have relatively little knowledge of its nature or the variables which are important to it. This thesis has attempted to redress this balance. SUMMARY The brief h istory of the academic study of crowding was reviewed i n the fi rst chapter. The psychology of crowding was introduced as a development of the earl ier work in the sociological and animal literature. A legacy of th is work was the view that under crowded conditions negative consequences would fol low. This view has continued to underpin the psychological literatu re, with the prevai l ing notion that crowding is an aversive experience, even when extreme conditions are not present, and is to be avoided or reduced if at al l possible. The importance of this theme should not be ignored, as it is apparent in contemporary writing that the negative elements of the experience sti l l dominate the manner in which the experience is interpreted. The psychological study of crowding and its consequences occurs within the context of an interest i n the 1 93 behaviour of the individual in a group. This remains of fundamental interest for those concerned with social behaviour. The difficu lties in studying such a complex phenomenon as crowding are compounded by the myriad of conceptual and defin it ional problems which were also discussed in the fi rst chapter. The range of psychological models of crowding reflect the numerous perspectives which may impinge on this particu lar human" experience. One of these perspectives may be referred to as the spatial cognitive perspective , where varying factors associated with the experience of crowding are l inked to attributional processes. Three overlapping yet distinctive attributional models were reviewed in Chapter Two. I n Chapter Three the attributional arousal mode l was scrutinised closely. This model holds that a sequence of two factors (spatial i nvasion resu lting in arousal fo l lowed by an attributional search for the source of arousal) wi l l result in the experience of crowding. The attributional arousal model regards crowding as a negative experience and predicts performance deficits and deteriorated interpersonal relationships to be associated with the experience. A number of studies, based on the two factor model , were reviewed. These studies aimed to reduce reports of crowding by manipulati ng nonspatial factors. One study (Worchel & Yohai , 1 979) demonstrated that subjects could be induced to misattribute the sou rce of their arousal to an alternative (fictitious) sou rce. 1 94 Providing subjects with an alternative source of arousal resu lted in reductions i n the extent to which they reported fee l ing crowded. This demonstration provided the i mpetus for Study One as it provoked a number of interesting questions. The most fundamental of these questions was, why do crowded subjects opt for an alternative source to which they may attribute their arousal? A further source of i ntrigue was the question of why crowded subjects wou ld attribute arousal to an unverifiable alternative source i n preference to the actual source? One l ine of reason ing developed to answer these questions was based on the suggestion that people are typically unaware of their spatial behaviour. If this was so then i n an experimental setting , presenting an alternative source of arousal to subjects may have simply been effective due to the obvious salience of that potential sou rce. Study One examined th is salience hypothesis by presenting crowded subjects with a number of equally salient sources of arousal. Among these was an alternative as wel l as the actual source of arousal . The actual sou rce of arousal (spatial i nvasion) was presented with equal salience with the alternative sou rce to see if the alternative was sti l l the preferred choice. The results of Study One were both unexpected and provocative. The most important finding was the fact that crowded subjects fai led to report feel ing more aroused than uncrowded subjects. This result proved to be extremely disruptive as it made it impossible to subsequently examine the predictions which stem from the two factor model. The comparative nonarousal of crowded subjects suggested that the basic dynamics of the experience of crowding (as outl i ned by 1 95 the two factor model) were not seen to be functioni ng. The fai lure to document significant increases in arousal in crowded subjects was a particu larly damaging finding for the model which is predicated on the foundation that spatial i nvasions do cause arousal. A number of interpretations of this finding were discussed (Chapter Six) including the theoretical implications, measurement problems and the possible effects of using on ly female subjects. In spite of the nonarousal of crowded subjects, these subjects sti l l fe lt more crowded, confined and uncomfortable than their noncrowded counterparts. I f crowded subjects were not aroused and yet sti l l reported feel ing crowded, then this resu lt has especial ly critical consequences for the two factor model . These are firstly, that one of the fundamental predictions of the model fai led to receive any support from this study. Secondly, this result suggests the experience of crowding may occur without the associated high levels of arousal . Chapter Seven saw the i ntroduction of the information expectancy model , which was developed due to the inabi lity of the two factor model to explain the nonarousal of crowded subjects and the fact that accurate information about the natu re of the experience of crowding provided before the event had the power to reduce the impact of the experience. The cornerstone of the information expectancy model is the notion that confirmation of expectations wi l l serve to reduce the impact of crowding. A key distinction may be drawn between the two factor and information expectancy models. Supporti ng evidence for the two factor mode l depended on the provision of deceptive information about a nonexistent variable. It was intended to find supporting evidence for the information expectancy model by simply providing information about the actual variables involved. 1 96 Study Two aimed to clarify the role of both information and expectations, in an experimental context, with regard to the experience of crowding. Auxi l iary aims included col lecting information on the efficacy of two measurement scales for each of the key concepts of arousal and crowding. The psychometric properties of the four measurement scales were characterised by h igh indices of internal consistency. These results were encouragi ng and suggested that the four scales were high ly homogeneous. Furthermore, the measures of internal consistency compared favourably with previous resu lts, where these were avai lable. Concerning the role of spatial information the results were disappoint ing. Accurate spatial information fai led to have the predicted effect and th is result cal led into question the uti lity of accurate pre-exposure information . The disconfi rmation of nonspatial expectations resu lted in increases in reported levels of crowding , a resu lt which h igh lighted the importance of accurate information about the nature of the activity which was to occur whi le subjects were i n a crowded situation. A further result was the clear sex difference such that women reported feeling less crowded and less annoyed than men. This result contri butes to the literature on sex differences which has to date been equivocal , with a leaning toward the suggestion that women are more adaptable under conditions of restricted interpersonal distance, and do not report the negative aspects of crowding to the same extent as men. 1 97 Final ly, a number of suggestions for further research were advanced. These i ncluded the suggestions that fi rstly, the re lationship between the experience of crowding and personal space invasions be examined more closely, and secondly, that the impact which nonspatial information may have on the experience of crowding may previously have been underestimated and future research cou ld usefu l ly explore this. CONCLUS IONS Conclusions from this research may be drawn at a methodological or practical leve l , at a theoretical level and finally at a more general level . There exist a number of barriers to conducting experimental crowding research. The work is demanding of subject numbers , and access to large pools of vo lunteers is important. One way of overcoming this problem wou ld be to uti l ise non interactive groups. Organisational difficulties may also arise when arranging g roups of subjects to arrive at agreed times. Thus, at a practical level , there are some difficu lties i n conducting research of this type. Another problem concerns the doubts about whether we are actually studying 1 98 crowding by putting five people together in a room for an hour. This approach brings a high degree of artificiality to the study of crowding, with the added advantage of increasi ng the level of experimental control . These costs and benefits need to be considered carefu l ly before a decision is reached as to the most appropriate setting in which to study human crowding. Perhaps the term "small g roup crowding" may more accurately describe this type of experimental research. Conceptually, the area has drawn from a number of discipl ines which have an interest in human spatial re lationships. This diversity of inputs has resu lted in a degree of conceptual confusion. The defin it ions of basic constructs are less than clear. This situation has done nothing to enhance the state of knowledge in the area, and has resu lted in little agreement about important measurement issues. In some small way it is hoped that this thesis has contributed to the i nformation avai lable about possible measuring instruments for the concepts of crowding and arousal. Theoretical ly, this research has attempted to throw l ight on two models of crowding. The two factor attributional arousal model was initial ly examined, but was found unexpectedly to be wanting at a fundamental but crucial level . This led to the conclusion that arousal is not as essential to the experience of crowding as the two factor model suggests. In short, one of the main l inks in the sequence of events predicted by the two factor model was found to be missing, and yet subjects sti l l reported feel ing crowded. This lack of support for the two 1 99 factor mode l led to two conclusions. Firstly, that the model had previously been supported by only a handfu l of studies and the resu lts of Study One were i ncompatible with the basic predictions of the model. Secondly, that the resu lts of Study One could have simply been regarded as' anomalous and ignored or they cou ld have signalled a weakness in the two factor mode l , in which case an alternative interpretation was invited. The care taken to replicate the methodology of the key study supporting the two factor model counted agai nst the possibi l ity of the resu lts of Study One being a simple experimental anomaly. The i nabil ity of the two factor mode l to account for key aspects of the findings of Study One prompted the development of an information expectancy mode l . This model did not re ly on arousal-based mechanisms which had apparently not been present i n otherwise crowded subjects in the fi rst study. Rather, the mode l focused on the impact of i nformation and expectations o n the experience of crowding. The results of Study Two which examined hypotheses derived from the mode l were encouraging. Spatial i nformation did shape spatial expectations and subjects processed information relevant to nonspatial concerns with in the crowded situation . There was also qualified support for the idea that spatial i nvasions are l i nked to the experience of crowding. Whi le key aspects of the model were supported , includi ng the high l ighting of the importance of nonspatial i nformation , one fundamental prediction of the model fai led to receive support. Study Two was unable to demonstrate that accurate information (and 200 presumably also accurate spatial expectations) was of any assistance in reducing subjective reports of crowding. This fai lure must lead to the conclusion that whi le the resu lts were encouraging, the information expectancy model has some distance to travel before it is in a position to tel l the complete story about how people come to feel crowded. In the process of examining the predictions of these two models of crowding , this research has exposed evidence which runs contrary to assumptions which are fundamental to the models. I n Study One the arousal assumption was not supported and i n Study Two the util ity of accurate pre-exposure spatial information was found wanting. These findings emphasise the importance of examin ing assumptions careful ly. F inal ly , at a more general level , the contribution of this thesis must be considered in the l ight of the experience which we are al l so fami l iar with , yet know so little about. In its simplest form this research has raised the possibi lity of feel ing crowded without the need to involve arousal related explanations. Arousal may sti l l be present, but may operate independently of fee lings of crowding . This thesis has further contributed to our understanding of the many variables which may impact on the experience of crowding , by stressing the importance of both i nformation and subsequent expectations. Information about the activities that one wi l l engage in whi le in a crowded situation clearly may have an impact on the manner in which the experience of crowding is interpreted. Crowding is 201 not somethi ng which simply happens to peop le, but rather is an experience which people wil l actively thi nk about, develop expectations about and anticipate. Such activities are important because they have the power to influence the nature of the experience. Furthermore , these activities are high ly l ike ly to be stimu lated by the avai labil ity of information about the things people are expected to do while in the situation. In this regard this research has taken an important step i n attempting to deli neate the role of the context within which crowding occurs. 202 REFERENCES Aiel lo, J .R . , and Aiel lo, T.D. ( 1 974). Deve lopment of personal space : Proxemic behavior of children six to sixteen. Human Ecology, g, 1 77-1 89. Aiel lo , J . R. , DeRisi , D.T, Epstein , Y.M. , and Karl in , R.A. ( 1 977). Crowding and the role of interpersonal distance preference. Sociometry, 40 , 271 -281 . Aiel lo , J .R . , Thompson, D.E . , and Brodzinsky, D.M. ( 1 983). How funny is crowding anyway? Effects of room size, g roup size, and the introduction of humor. Basic and Applied Social Psychology, �, 1 93-207. Altman , I . (1 975). The envi ronment and social behaviour. Monterey, California: Brooks/ Cole. Altman , I. (1 978). Crowding : Historical and contemporary trends in crowding research. In A. Baum and Y.M. Epstein (Eds. ) , Human response to crowding. Hi l lsdale, N .J . : Erlbaum. Anastasi , A. ( 1 982) . Psychological testing (5th ed.) . New York: Macmi l lan . Anastasi , A. (1 988). Psychological testing (6th ed.). New York: Macmi l lan. Argyle , M. and Dean, J . (1 965). Eye contact, distance and affi l iation. Sociometry, 28, 289-304. Averi l l , J .R. ( 1 973) . Personal contro l over aversive stimu li and its relationship to stress. Psychological Bu l leti n , 80 , 286-303. Baldassare, M. ( 1 978). Human spatial behaviour. Annual Review of Sociology, �, 29-56. Baron , R.A. and Byrne, D.E. (1 987). Social Psychology: Understandi ng Human Interaction (5th ed. ) . Massachusetts: Al lyn and Bacon. 203 Baum , A, Fisher, J .D . , and Soloman , S.K. ( 1 981 ) . Type of information , fami liarity , and the reduction o f crowding stress. Journal of Personality and Social Psycho logy, 40, 1 1 -23. Baum , A, and Greenberg , C . 1 . (1 975). Waiting for a crowd: The behavioral and perceptual effects of anticipated crowding. Journal of Personality and Social Psycho logy, 32, 671 -679. Burger, J .M . , Oakman , J.A, and Bul lard , N .G. ( 1 983). Desi re for control and the perception of crowding. Personality and Social Psychology Bul letin , �, 475-479. Buros, O.K. ( 1 972). The Seventh Mental Measurements Yearbook. High land Park, New Jersey : Gryphon Press. Calhoun, J .B. ( 1 962). Popu lation density and social pathology. Scientific American, 206, 49-59. Calvert - Boyanowsky, J . , and Levanthal , H. ( 1 975). The role of information in attenuating behavioral responses to stress: A reinterpretation of the misattribution phenomenon. Journal of Personality and Social Psycho logy, 32 , 2 1 4-221 . Choi , S.C. , Mi rjafari , A. , and Weaver, H .B. (1 976). The concept of crowding : A critical review and proposal of an alternative approach. Environment and Behavior, §., 345-362 . Christian, J .J . , Flyger, V. , and Davis , D.E. (1 960). Factors i n the mass mortality of a herd of skia deer, Cervus nippon . Chesapeake Science , 1, 79-95. Cohen, J . L. , S iaden , B . , and Bennett, B. ( 1 975). The effects of situational variables on judgements of crowding. Sociometry, 38, 273-281 . Cohen , M .N . , Malpass, R .S . , and Klei n , H .G. ( 1 980). I ntroduction. I n M .N . Cohen , R .S . Malpass, and H .G. Klein (Eds. ) , Biosocial mechanisms of population regu lation . London: Yale University P ress. Cohen, S. ( 1 978) . Environmental load and the allocation of attention. In A Baum, J .E . Si nger, and S . Valins (Eds.) , Advances i n environmental psychology (Vol . 1 ) . Hi l lsdale , N .J . :Erlbaum. 204 Cohen , S. , and Spacapan, S. ( 1 978) . The aftereffects of stress: An attentional i nterpretation. Environmental Psychology and Nonverbal Behavior, �, 43- 57. Cox, V.C. , Pau lus, P .B . , and McCain, G. ( 1 984) . Prison crowding research : The relevance of p rison housing standards and a general approach regarding crowding phenomena. American Psychologist , 39, 1 1 48-1 1 60. D'Atri , D. (1 975). Physiological responses to crowding. Envi ronment and Behavior, I, 237-25 1 . Desor, J .A. ( 1 972). Toward a psycholog ical theory of crowding . Journal of Social Psychology and Personality, �, 79-83. El l iot, E .S. , and Cohen, J . L. (1 981 ) . Social faci l itation effects via i nterpersonal distance. The Journal of Social Psychology, 1 1 4 , 237-249. Epstei n , Y.M. ( 1 980) . Physiological effects of crowding on humans. In M .N . Cohen, R .S. Malpass, and H .G. Klein (Eds . ) , Biosocial mechanisms of population regu lation. Newhaven : Yale University Press. Epstein , Y.M. , and Baum, A. ( 1 978) . Crowding : Methods of study. I n A. Baum and Y.M. Epstein (Eds.) , Human response to crowding. Hi l lsdale , N .J . : Erlbaum. Epstein , Y.M. , and Karl in , R.A. (1 975). Effects of acute experimental crowding. Journal of Applied Social Psychology, §, 34-53. Evans , G.W. ( 1 978) . Human spatial behavior. The arousal model . In A. Baum and Y. Epstein (Eds.) , Human response to croWding. Hi l lsdale , N .J . : Erlbaum. Evans, G .W. ( 1 979). Behavioral and physiological consequences of crowding in humans. Journal of Applied Social Psycho logy, �, 27-46. Evans , G.W. , and Howard, H .R .B . ( 1 973). Personal space. Psychological Bul letin , 80, 334-344. 205 Faris , R. , and Dunham , H.W. ( 1 965) . Mental disorders in urban areas (2nd ed. ) . Chicago : Phoenix Books. Farr, R .M. (1 986). The social psychology of Wi l l iam McDougal. In C .F. Graumann and S . Moscovici (Eds . ) , Changing conceptions of crowd mind and behavior. New York: Springer-Verlag . . Farri ngton , D.P . , and Nuttal , C.P. (1 980) . Prison size , overcrowding, prison violence , and recidivism. Journal of Criminal Justice, ,§" 221 -231 . Feather, N.T. , and Simon, J .G. ( 1 971 ) . Causal attributions for success and fai lure in relation to expectations of success based upon selective or manipu lative control . Journal of Personality, 39, 527-541 . Festi nger, L. , Pepitone , A. , and Newcomb, T.M . ( 1 952). Some consequences of deindividuation i n a group. Journal of Abnormal and Social Psychology, 47, 382-289. Fisher, J .D . , and Baum , A. ( 1 980). Situational and arousal based messages and the reduction of crowding stress. Journal of Applied Social Psycho logy, 1Q, 1 9 1 -20 1 . Fiske , S .T. , and Taylor, S .E . ( 1 984) . Social cognition. Reading, M.A. : Addison­ Wesley. Forsyth , D.R. ( 1 987). Social Psychology. Belmont, C.A. : Wadsworth. Freedman , J .L. ( 1 972). The effects of popu lation density on humans. In J .Fawcett (Ed . ) , Psychological Perspectives On Population. New York: Basic Books. Freedman , J .L. ( 1 975) . Crowding and behavior. San Francisco : Freeman. Freedman , J .L. ( 1 979) . Reconci l ing apparent differences between the responses of humans and other animals to crowding. Psychological Review, 86, 80- 85. 206 Freedman, J . L. ( 1 980). H uman reactions to population density. In M .N . Cohen, R.S. Malpass , and H .G. Klein (Eds. ) , Biosocial mechanisms of population regu lation . London : Yale University Press. Freedman, J . L. , Klevansky, S . , and Ehrl ich, P .R. ( 1 971 ) . The effects of crowding on human task performance. Journal of Experimental Social Psychology, 1, 7-25. Gifford , R. ( 1 982). Projected interpersonal distance and orientation choices: Personality, sex and social situation . Social Psychology Quarterly, 45, 1 45-1 52 . Gifford , R . (1 987). Environmental psycho logy: Principles and practice. Newton, M.A. : Al lyn and Bacon . Gochman , I .R . , and Keating, J .P . (1 980). Misattributions to croWding : Blaming crowding for nondensity caused events. Journal of Nonverbal Behavior, 1, 1 57-1 75 . Greenberg , C . I . , and Fi restone, I .J . ( 1 977). Compensatory responses to crowding : Effects of personal space intrusion and privacy reduction. Journal of Personality and Social Psychology, 35, 637-644. Griffitt , W. , and Veitch, R. ( 1 971 ) . Hot and crowded : Innuences of popu lation density and temperature on interpersonal affective behavior. Journal of Personal ity and Social Psychology, .1Z, 92-98. Hal l , E .T. ( 1 959) . The si lent language. New York: Doubleday. Hal l , E .T. ( 1 963) . A system for the notation of proxemic behavior. American Anthropologist, 65, 1 003-1 026. Hal l , E .T. ( 1 966) . The hidden dimension. New York: Doubleday. Haydu k, L.A. ( 1 983). Personal space : Where we now stand. Psychological Bulletin , 94, 293-335. 207 He l ler, J .F . , Groff, B .D . , and Soloman, S.H. (1 977) . Toward an understanding of crowding : The ro le of physical interaction . Journal of Personality and Social Psycho logy, 35, 1 83-1 90. Ho lahan, C.J . ( 1 982). Envi ronmental psycho logy. New York: Random House. Hutt , C. , and Vaizey, M. ( 1 966). Differential effects of g roup density on social behaviour. Nature, 209, 1 37 1 - 1 372. I nsel , P .M. , and Lindgren, H .C. ( 1 978) . Too close for comfort. Eng lewood Cliffs, N . J . : Prentice-Hal l . Jain , U . ( 1 987). Effects of popu lation density and resources on the feelings of crowding and personal space. Journal of Social Psychology, 27, 331 -338 . Kalb , L.S . , and Keating, J .P . (1 981 ) . The measurement of perceived crowding . Personality and Social Psychology Bul leti n , Z , 650-654. Karlin , R.A. ( 1 980). Social effects of crowding in humans. In M.N. Cohen , RS. Malpass, and H .G. Klein (Eds. ) , Biosocial mechanisms of population regu lation . Newhaven : Yale University Press. Karlin , RA. , and Epstein, Y .M. ( 1 979). Acute crowding : A reliable method for inducing stress in humans. Research Communications in Psychology, Psychiatry and Behavior, �, 357-370. Karl in , RA. , Katz, S . , Epstein , Y.M. , and Woolfo lk, RL. ( 1 979) . The use of therapeutic interventions to reduce crowding related arousal: A prel iminary investigation. Envi ronmental Psychology and Nonverbal Behavior, �, 21 9- 227. Keppel , G. ( 1 982). Design and Analysis: A Researchers Handbook. Eng lewood Cliffs, N .J . : Prentice-Hal l . 208 Klei n , K., and Harris , B. ( 1 979). Disruptive effects of disconfirmed expectancies about crowding. Journal of Personality and Social Psycho logy, 37, 769- 777. Knowles, E.S. ( 1 980). An affi l iative conflict theory of personal and group spatial behavior. In P .B . Pau lus (Ed.) Psychology of group influence. Hi l lsdale , N .J . : Erlbaum . Kruse, L. ( 1 986). Conceptions of crowds and crowding. I n C.F. Graumann and S. Moscovici (Eds.) , Changing conceptions of crowd mind and behavior. New York: Springer-Verlag. Kuykendal l , D . , and Keating , J .P . ( 1 984) . Crowding and relations to uncontro l lable events. Population and Envi ronment, I, 246-259. Langer, E . , B lank, A. , and Chanowitz, B. ( 1 978) . The mindlessness of ostensibly thoughtfu l action : The role of placebic information in interpersonal interaction . Journal of Personality and Social Psychology, 36, 635-642. Langer, E. , and Saegert, S. ( 1 977) . Crowding and cognitive control . Journal of Personal ity and Social Psychology, 35 , 1 75-1 82. Le Bon , G. (1 903). The crowd. London : Unwin. Levitt , L. , and Levanthal, C. ( 1 978). Effect of density and environmental noise on perception of time, the situation, oneself and others. Perceptual and Motor Skil ls, 47, 999-1 009. Loo , c. ( 1 972) . The effects of spatial density on the social behavior of chi ldren. Journal of Applied Social Psychology, g, 372-381 . Manstead, A.S.R. , and Wagner, H .L. ( 1 981 ) . Arousal , cognition and emotion : An appraisal of two-factor theory. Current Psychological Reviews, 1, 35-54. Marshal l , J .E . , and Hesl in, R. ( 1 975). Boys and g i rls together: Sexual composition and the effect of density and g roup size on cohesiveness. Journal of Personality and Social Psychology, ,;tL, 952-961 . 209 McBride , G . , King , M.G. , and James, J.W. ( 1 965). Social proximity effects on galvanic skin responses in human adults. Journal of Psycho logy, 2.1, 1 53- 1 57. McCallum , A. , Rusbu lt, C.E., Hong , G .K. , Walden , T.A , and Schopler, J . ( 1 979). Effects of resource avai labi lity and importance of behavior on the experience of crowding . Journal of Personality and Social Psycho logy, 37, 1 304- 1 3 1 3. McDougal l , W. ( 1 921 ) . The group mind. Cambridge : Cambridge University Press. McFarlan , D . , (Ed . ) ( 1 988). The Guinness Book of Records 1 989. Middlesex: Guinness Publ ishing. McGrew, P .L. ( 1 970). Social and spatial density effects on spacing behavior in pre-school chi ldren. Journal of Personal ity and Social Psychology, 11, 1 97-205. Mehrabian, A , and Russel l , J .A (1 974) . An Approach To Environmental Psychology. Boston , Mass. : The MIT Press. Middlemist , A.D. , Knowles, E .S . , and Matter, C.F. ( 1 976). Personal space i nvasions i n the lavatory: Suggestive evidence for arousal. Journal of Personality and Social Psychology, 33, 541 -546. Milg ram, S. ( 1 970) . The experience of l iving in cities. Science, 1 67, 1 46 1 -1 468. Morasch , B . , Groner, N . , and Keating , J .P. ( 1 979). Type of activity and fai lure as mediators of perceived crowding. Personality and Social Psychology Bul leti n , § , 223-226. Moscovici , S. ( 1 986) . The discovery of the masses. In C.F. Graumann and S. Moscovici (Eds. ) , Changing conceptions of crowd m ind and behavior. New York: Springer-Verlag. Muel ler, C.W. (1 984). The effects of mood and type and timing of influence on the perception of crowding . The Journal of Psychology, 1 1 6 , 1 55-1 58. 2 1 0 Nagar, D. , and Pandey, J . (1 987) . Affect and performance on cognitive task as a function of crowding and noise. Journal of Applied Social Psycho logy, 12. 1 47-1 57. Neisser, U. ( 1 976) . Cognition and reality. San Fransisco : Freeman. Nicosia, G .J . , Hyman, D . , Karl in , R.A. , Epstein , Y.M. , and Aiel lo, J .R . (1 979). Effects of bodi ly contact on reactions to croWding. Journal of Applied Social Psycho logy, �, 508-523. Norusis, M.J . (1 985) . SPSS-X Advanced Statistics Guide. New York: McGraw­ Hi l l . Orne, M.T. (1 962). On the social psychology of the psychological experiment : With particu lar reference to demand characteristics and their impl ications. American Psychologist, 1l. 776-783. Patterson, M.L. ( 1 975). Personal space - time to bu�st the bubble? Man­ Environment Systems, §, 67. Patterson, M.L. (1 976). An arousal mode l of interpersonal intimacy. Psychological Review, 83, 235-245. Patterson, M.L. ( 1 978). Arousal change and cognitive labe ll ing : Persuing the mediators of intimacy exchange. Envi ronmental Psychology and Nonverbal Behavior, 3, 1 7-22. Patterson, M .L. , Roth, C .P. , and Schenk, C. ( 1 979) . Seating arrangement, activity, and sex differences in small g roup crowding. Personality and Social Psychology Bul letin , §, 1 00-1 03. Paulus, P.B. ( 1 980). Crowding. In P .B . Pau lus (Ed . ) , Psychology of group influence. Hi l lsdale , N .J. : Lawrence Erlbaum . Paulus, P.B , Annis, A.B. , Seta, J .J . , Schkade, J .K. , and Mathews, R.W. ( 1 976) . Density does affect task performance. Journal of Personality and Social Psycho logy, 34 , 248-253. Paulus, P .B . , and Mathews, R.W. ( 1 980). Crowding, attribution, and task performance. Basic and Applied Social Psycho logy, 1, 3-1 3. 21 1 Pau lus , P .B . , McCain , G . and Cox, V. ( 1 979) . Death rates, psychiatric commitment, b lood pressure and perceived crowding as a function of i nstitutional crowding . Environmental Psychology and Nonverbal Behavior, �, 1 07-1 1 6. Penner, L.A. ( 1 986). Social Psycho logy. St.Pau l , M .N . : West. Proshansky, H . , Ittleson , W.H . , and Rivlin , L.G. (Eds. ) . (1 970). Environmental psychology: Man and his physical setting. New York: Holt , Rinehart and Winston . Pryor, J . B. , and Kriss, M . (1 977). The cognitive dynamics of salience in the attribution process. Journal of Personality and Social Psychology, 35, 49- 55. Rosenthal , R., and Rosnow, R.L. ( 1 985) . Contrast analysis : Focussed comparisons in the analysis of variance. New York: Cambridge University P ress. Ross, L. , Rodin , J . , and Zimbardo, P.G. ( 1 969). Toward an attribution therapy : The reduction of fear through induced cognitive-emotional misattribution . " Journal of Personality and Social Psychology, �, 279-288. Rotton , J. ( 1 987). Hemmed in and hating it : Effects of shape of room on tolerance for crowding. Perceptual and Motor Ski l ls , 64, 285-286. Saegert, S. ( 1 978). High density envi ronments : Their personal and social consequences. In A. Baum and Y.M. Epstein (Eds. ) , Human response to crowding. Hi l lsdale , N .J . : Erlbaum. Schacter, S . ( 1 964) . The i nteraction of cognitive and physiological components of emotional state. I n L.Berkowitz (Ed . ) , Advances in Experimental Social Psychology (Vol . 1 ) . New York: Academic Press. Schacter, S . , and Singer, J .E. (1 962) . Cognitive , social , and physiolog ical determinants of emotional state. Psychological Review, 69 , 379-399. 21 2 Schaeffer, G .H . , and Patterson, M .L. ( 1 980). Intimacy, arousal , and small g roup crowding. Journal of Personality and Social Psychology, 38, 283-290. Schiffenbauer, A. , and Schiavo , R.S. ( 1 976). Physical distance and attraction : An intensification effect. Journal of Experimental Social Psychology, �, 274- 282. Schmid , C. ( 1 955). Completed and attempted suicides. American Sociological Review, 20, 273. Schmid, C. ( 1 969) . U rban crime areas : Part 1 . American Socio logical Review, 25, 527-542. Schmidt , D .E . , and Keating, J .P . (1 979). Human crowding and personal contro l : An integ ration o f the research. Psycho logical Bul leti n , 86 , 680-700. Seligman, M . P. ( 1 975). Helplessness: On depression, development and death. San Francisco : Freeman. Sherrod, D.R. ( 1 974). Crowding , perceived contro l , and behavioral aftereffects. Journal of Applied Social Psycho logy, �, 1 71 - 1 86. Smith , R.H . , and Knowles, E.S. ( 1 979) . Affective and cognitive mediators of reactions to spatial invasions. Journal of Experimental Social Psycho logy, 1§., 437-452. Sommer, R. ( 1 969) . Personal Space: The Behavioral Basis of Design. Englewood C liffs , N .J . : Prentice Hall . Stockdale , J .E . ( 1 978). Crowding : Determinants and effects. Advances in Experimental Social Psychology, 11, 1 97-247. Stokols , D. ( 1 972) . On the distinction between density and crowding : Some implications for future research. Psycho logical Review, 79 , 275-277. Stoko ls, D. (1 976) . The experience of crowding in primary and secondary environments. Envi ronment and Behavior, 8 , 49-86. Stoko ls, D . , Rai l , M . , Pinner, B. , and Schopler, J. ( 1 973) . Physical , social and personal determinants of the perception of crowding. Environment and Behavior, §, 87- 1 1 5. Storms, M .D. , and N isbett , R.E. (1 970). Insomnia and the attribution process. Journal of Personality and Social Psychology, 12, 31 9-328. Storms , M.D. , and Thomas, G.C. ( 1 977). Reactions to physical closeness. Journal of Personality and Social Psychology, 35, 41 2-41 8. 2 1 3 Strube, M.J . , and Werner, C. (1 982). Interpersonal distance and personal space: A conceptual and methodo logical note. Journal of Nonverbal Behavior, 2, 1 63-1 70. Taylor, S.E. , Crocker, J . , Fiske, S.T. , Sprinzen , M . , and Winkler, J .D. (1 979). The generalisabi lity of salience effects. Journal of Personality and Social Psychology, 37, 357-368. Taylor, S .E . , and Fiske , S .T. (1 978) . Salience , attention and attribution : Top of the head phenomena. In l.Berkowitz (Ed . ) , Advances in Experimental Social Psychology (Vol. 1 1 ) . New York: Academic Press. Templer, 0. 1 . (1 987). Review of Multiple Affect Adjective Check List - Revised. In D.J. Keyser, and R.C. Sweetland (Eds. ) , Test critiques compendium (Vol .4 , pp.323-326). Kansas City, M.O. : Test Corporation of America. Tesch , F.E. ( 1 977) . Debriefing research participants: Though this be method there be madness to it. Journal of Personality and Social Psychology, 35, 2 1 7-224. Valins, S . , and Nisbett , R .E. ( 1 972) . Attribution processes in the development and treatment of emotional disorders. In E . Jones, D. Kanouse, R. N isbet, and B .Weiner (Eds.) , Attribution: Perceivi ng the causes of behavior. Morristown , N .J . : General Learning Press. 21 4 Vali ns, S . , and Ray, A. ( 1 967) . Effects of cognitive desensitisation on avoidance behavior. Jou rnal of Personality and Social Psychology, I, 345-350 . Vine, I . ( 1 982). Crowding and stress 2 : A personal space approach. Current Psychological Reviews, g, 1 - 1 8. Walden, T.A. , and Forsyth , D .R . ( 1 981 ) . Close encounters of the stressfu l kind: Affective , physio log ical and behavioral reactions to the experience of crowding. Jou rnal of Nonverbal Behavior, Q, 46-64. Ward , C.D. ( 1 970) . Laboratory manual in experimental social psychology. New York: Holt, Ri nehart and Winston . Webb, B . , Worchel , S . , Riechers , L. , and Wayne , W. ( 1 987). The i nfluence of categorization on perceptions of crowding. Personality and Social Psycho logy Bul letin , 12, 539-546. Wener, R.E . , and Kaminoff, R .D. ( 1 983). Improving environmental information : Effects of signs on perceived crowding and behavio r. Envi ronment and Behavior, .1§, 3-20 . Wi l l is , F.N. ( 1 966). In itial speaking distance as a function o f the speakers' re lationship. Psychonomic Science, 9., 221 -222. Worchel , S. ( 1 978) . Reducing crowding without reducing space : Some applications of an attributional theory of crowdi ng. Journal of Population : Behavioral, Social and Environmental Issues, 1, 21 6-230. Worchel , S . , and Brown , E .H . (1 984) . The role of plausibil ity in influencing environmental attributions. Jou rnal of Experimental Social Psychology, 20 , 86-96. Worche l , S . , and Cooper, J . ( 1 983) . Understanding social psychology (3rd ed. ) . Homewood, ILL: Dorsey Press. Worchel , S . , and Teddl ie, C. ( 1 976). The experience of crowding: A two factor theory. Journal of Personality and Social Psychology, 34, 30-40. 21 5 Worche l , S . , and Yohai , S .M. ( 1 979) . The ro le of attribution in the experience of crowding. Journal of Experimental Social Psychology, .1§, 9 1 -1 04. Zajonc, R. (1 965). Social faci litation . Science, 1 49 , 269-274. Zeedyk-Ryan , J . and Smith , G .F. (1 983). The effects of crowding on hosti l ity , anxiety, and desire for social interaction . Journal of Social Psychology, 1 20 , 245-252. Zuckerman , M. ( 1 979). Attribution of success and fai lure revisited , o r: The motivational bias is alive and wel l in attribution theory. Journal of Personality, 47, 245-287. Zuckerman, M . , and Lub in , S. ( 1 985) . Manual for the Mu ltiple Affect Adjective Check List - Revised. San Diego : Educational and Industrial Testing Service. Zuckerman , M . , Lubin, S. , and Robins, S. (1 965) . Validation of the multiple affect adjective check list in cl in ical situations. Journal of Consulting and Cl in ical Psychology, 29 , 594. 21 6 APPENDIX A : EXPERIMENTAL PRE-INSTRUCTIONS (STUDY ONE) Fou r sets of experimental pre-instructions corresponding to the fou r information conditions were used. The subjects in the NO EXPLANATION condition were played the fo l lowing i nstructions : "Hel lo , thank you for volunteering for this experiment i n which we are i nterested in studying group performance and g roup interaction . . . The experimenter wi l l now explain the first task." Subjects i n the DISTANCE EXPLANATION condition heard the fol lowing i nstructio ns : "Hel lo , thank you for volunteering for this experiment i n which we are i nterested i n studying group performance and group interaction u nder varying i nterpersonal distances. In everyday life we are often exposed to distances between ourse lves and others that we are unconscious of but that may affect our behaviour. Our aim is to study the effects of i nterpersonal distance (such as the distance at which you are now seated) on performance. Previous studies have shown that this distance may cause you to fee l somewhat stressed and uncomfortable. The experimenter wi l l now explain the fi rst task." Subjects in the NOISE EXPLANATI ON condition were played the fo l lowing instructions: 2 1 7 "Hel lo , thank you for volunteering for this experiment in which we are interested in studying group performance and g roup interaction under varying types of subl iminal stimul i . In everyday life we are often exposed to stimu li that we are unconscious of but that may affect our behaviour. Our aim is to study the effects of sublim inal noise on performance. A subl im inal noise wil l be played into the room whi le you work on the tasks. This noise wi l l not be detectable to the naked ear. Previous studies have shown that this noise may cause you to feel somewhat stressed and uncomfortable. The experimenter wi l l now explain the fi rst task." The D U A L EXPLANATION condition subjects received the fol lowing instructions: "Hel lo , thank you for volunteering for this experiment in which we are interested in studying group performance and group interaction under varying types of subliminal stimuli and varying interpersonal distances. In 2 1 8 varying types of subl iminal stimul i and varying interpersonal distances . I n everyday life we are often exposed to distances between ourse lves and others , and stimul i that we are unconscious of but that may affect our behaviour. Our aim is to study the effects of both i nterpersonal distance (such as the distance at which you are now seated) and subliminal noise on performance. A subl imi nal noise wi l l be played into the room whi le you work on the tasks. This noise wi l l not be detectable to the naked ear. Previous studies have shown that this noise and this distance may cause you to fee l somewhat stressed and uncomfortable. The experimenter wi l l now explain the fi rst task." 21 9 APPEND IX 8: POST-EXPERIMENTAL QUESTIONNAIRE (STUDY ONE) INSTRUCTIONS The fo l lowing questions are designed to assess your reactions to the experiment. For each question p lace a cross (x) in the space which most accurately reflects your answer. Consider, for example, the following question : How warm do you expect the weather to be today? Extremely hot : X : : : : : : : : -- --------- Extremely cold The above response i ndicates an expectation of a very hot day. 1 . General ly do you prefer to work alone or i n a group? Alone_:_:_:_:_:_:_:_:_:_ In a g roup 2. How much did you l ike the people in your g roup? Disliked : : : : : : : : : Like very much -- -- -- -- -- 3. How bothered were you by the lighti ng in the experimental room? Not at al l _:_:_:_:_:_:_:_:_:_ Very bothered bothered 4. How confined did you feel during the experiment? Not at all _:_:_:_:_:_:_:_:_:_ Very confined confi ned 5. Du ri ng the experiment did you find the experimenter helpful or unhelpfu l? Very helpful _ :_:_ :_ :_:_:_:_:_:_ Very unhelpfu l 6 . Was there a leader in your group? YES/NO (Circle one response) If so, identify by letter A B C D E (Circle one response) 7. I n completing the tasks were you able to concentrate? U nable to _ :_ :_ :_ :_:_:_:_:_ :_ Able to concentrate concentrate 8 . How pleasant did you find the experimental room? Not at al l _:_ :_:_:_:_:_:_:_ :_ Very pleasant p leasant 9. During the experiment how bothered were you by noise? 220 Very bothered _:_:_:_:_:_:_:_:_:_ Not at all bothered 221 1 0 . How happy were you with the performance of your group? Very happy _:_:_:_:_:_:_:_:_:_ Not at al l happy 1 1 . How crowded did you feel during the experiment? Not at all _:_:_:_:_:_:_:_:_:_ Very crowded crowded 1 2. How bothered were you by the temperatu re in the experimental room? Very bothered _:_:_:_:_:_:_:_:_:_ Not at a l l bothered 1 3. Rate how aggressive each group member was (excluding yourself) Group Member A Very agg ressive_:_:_:_:_:_:_:_:_:_Not at al l aggressive B Very agg ressive_:_:_:_:_:_:_:_:_:_Not at al l aggressive e Very agg ressive_:_:_:_:_:_:_:_:_:_Not at all aggressive D Very agg ressive_:_:_:_:_:_:_:_:_:_Not at all aggressive E Very aggressive_:_:_:_:_:_:_:_:_:_Not at all aggressive 1 4. How comfortable did you feel during the experiment? Very . . . . . . . . . . . . . . . ---------- comfortable Very uncomfortable 1 5. How enjoyable was you r experience as a subject i n this experiment? Very _:_:_:_:_:_:_:_:_:_ Not at al l enjoyable enjoyable 1 6. How relaxed or i l l at ease did you feel duri ng the experiment? Relaxed _:_:_:_:_:_:_:_:_:_ II I at ease 1 7. How dist racted did you feel during the experiment? Not at all _:_:_:_:_:_:_:_:_:_ Very distracted distracted 1 8. Rate how friendly each g roup member was (excluding yourself) Group Member A Not at al l _:_:_:_:_:_:_:_:_:_ Very friend ly friendly B Not at al l _:_:_:_:_:_:_:_:_:_ Very friendly friendly C Not at all _:_:_:_:_:_:_:_:_:_ Very friendly friendly D Not at all _:_:_:_:_:_:_:_:_:_ Very friendly friendly E Not at all _:_:_:_:_:_:_:_:_:_ Very friendly friend ly 222 223 1 9 . Rate how nervous each group member was (excluding you rself) Group Member A Calm · . . . . . . . . Nervous ---------- B Calm · . . . . . . . . Nervous ---------- C Calm · . . . . . . . . Nervous ---------- D Calm · . . . . . . . . Nervous ---------- E Calm · . . . . . . . . Nervous ---------- 20. How wi l l ing wou ld you be to participate i n another, s imi lar experiment? Very wi l l ing _ : _ : _ : _ :_: _ : _ : _ : _ : _ Not at all wi l l ing 224 APPENDIX C : NOTES TO SUBJECTS (STUDY ONE) KEITH TUFFIN DEPARTMENT OF PSYCHOLOGY MASSEY UN IVERSITY Firstly I wou ld l ike to thank you again for volunteering as research subjects for the experiment which was conducted earl ier this year. You may recal l that fo l lowing the experiment you were debriefed. The major pu rpose of the debriefing was to reveal the deceptions i nvo lved and also to inform you of the questions which the research hoped to answer. You were also told that at the conclusion of the research you would be posted a summary of the fi ndings of the research. The reason for this summary is my commitment to the idea that as research subjects you deserve to be informed about the research you were involved i n . You have volunteered you r t ime and I believe you are entitled to something in return - in this case it takes the form of i nformation relevant to the experiment in which you participated. This summary wi l l be presented in fou r parts : ( 1 ) Theory, (2) Design , (3) Results, (4) Conclusion . 225 THEORY The experiment that you took part in was designed to examine some aspects of an attributional/arousal based theory of crowding. This theory suggests that the in itial stage in the experience of crowding involves a violation of personal space. One way of th inking about personal space is to regard it as the distance which we l ike to maintain between ourselves and others. The theory says that spatial i nvasions are a prerequisite to feel ing crowded and spatial invasions may cause the i ndividual to fee l aroused. Once aroused, the theory suggests that i ndividuals wi l l attempt to find an explanation for thei r arousal . This is often referred to as the "attributional search", whereby if the arousal is attributed to the presence and closeness of others, then the individual will experience crowding. D ESI G N The purpose of this section is to g ive you an idea of the way the experiment was "set up" , o r organised. Two variables were manipu lated. The fi rst of these was distance, with two leve ls, either close or far. As some of you wi l l recal l , you were asked to sit very close to the others i n your group, whi le others sat at a more conventional distance. The second variable which was manipu lated was the information which subjects received . Some subjects were told nothing about the purpose of the experiment; others were told that we were examin ing the effects of subliminal noise on 226 behaviour; sti l l others were i nformed that we were examining the effect of varying distances on behaviour; the final g roup were told we were interested in the effect of both distance and noise. R ES U LTS The resu lts reported here are based on a prel iminary analysis of the data. The theory predicts that crowded subjects wi l l fee l more aroused than noncrowded subjects. To measure arousal subjects were asked to report how re laxed or i l l -at-ease they felt. The crowded subjects reported feeli ng no more aroused than uncrowded subjects. The theory predicts that crowded subjects wi l l not perform as wel l as thei r uncrowded counterparts. To measure performance, g roups of subjects were asked to extract words from a master word (OBSERVATIONALLY). Contrary to predictions crowded subjects performed no differently when compared to uncrowded subjects. The interpersonal relationships of crowded subjects were expected to suffer. Three measures were examined ; ratings of how friendly the fel low group members were , how aggressive they were , and also how much the subjects l iked the people in their group. Crowded subjects rated their fel low group members to be as friendly as did noncrowded subjects. A simi lar pattern of 227 results emerged for both aggressiveness and l iking for others in the g roup. To measure subjects' perceptions of crowding two questions were asked. Subjects were asked how crowded they fe lt and also how confined they felt. On both these measures subjects in the close condition reported feeli ng more crowded/confined than subjects in the far condition. Contrary to predictions, info rming subjects that they were exposed to subl iminal noise fai led to reduce perceptions of crowding or confinement. An i nteresting result was that i nforming subjects that distance was a key variable, actually reduced the crowded subjects perceptions of confinement. CONCLUSIO N Unexpectedly we found that the experience of crowding was not arousing for subjects. C rowded subjects did not experience impai red performance on the word task, nor did the experience of crowding result in i nterpersonal hosti l ity. I n general , the present study fai led to support the predictions of the theory which it set out to examine. I hope this brief note is informative and serves to answer any questions which you , as subjects, have about the experiment. If you have any further questions 228 then please contact me. During worki ng hours I may be found in my office (P.322) on the third floor of the Psycho logy Department. Te lephone : work 69 099 exten'sion 7924, or home 76 234. Thank you for your participation. Keith Tuffi n 229 APPENDIX D: ANOVA TABLES FOR KEY DEPENDENT MEASURES FOR STUDY ONE (GROUP ANALYS IS) SUMMARY ANOVA TABLE FOR ARO USAL SOURCE Main Effects Distance Information 2-Way Interactions Distance Inform. Explained Residual Total 6 .725 2.890 3.845 3.465 1 0.200 OF 4 1 3 3 7 84.920 56 95. 1 20 63 1 .684 2.890 1 .282 1 . 1 55 1 .457 1 .51 6 1 .51 0 F 1 . 1 1 0 1 .906 0 .845 0 .762 0 .961 P 0.361 0. 1 73 0 .475 0.520 0 .468 SUMMARY ANOVA TABLE FOR CROWDING SOURCE Main Effects Distance Information 2-Way Interactions Distance Inform. Explained Residual Total 1 41 .227 1 33.402 7.825 1 1 .662 1 52.890 OF 4 1 3 3 7 1 1 4.670 56 267.560 63 MS F 35.307 1 7.242 1 33.402 65. 1 48 2 .608 1 .274 3.887 1 .898 21 . 841 1 0.666 2.048 4.247 230 p 0.000 0.000 0.292 0 . 1 40 0 .000 SUMMARY ANOVA TABLE FOR CONFI N EME NT SOURCE Main Effects Distance Information 2-Way Interactions Distance Inform. Explained Residual Total 44.495 36.602 7.892 6 .362 50.857 DF 4 1 3 3 7 1 30 .240 56 1 81 .097 63 1 1 . 1 24 4.783 36.602 1 5.738 2.631 1 . 1 31 2. 1 21 0 .9 1 2 7.265 3. 1 24 2.326 2.875 231 p 0.002 0.000 0.344 0 .441 0.007 SUMMARY ANOVA TABLE FOR PER FO R MANCE SOURCE Main Effects Distance Info rmation 2-Way Interactions 1 41 4.437 1 9. 1 41 1 397.297 OF 4 1 3 Distance Inform. 1 1 07.922 3 Explained 2522.359 7 Residual Total 1 2873.332 56 1 5395.691 63 353.609 1 9. 1 41 465.099 369.307 360.337 229.881 244.376 F 1 .538 0.083 2.023 1 .607 1 .567 232 p 0.204 0.774 0. 1 21 0. 1 98 0. 1 64 SUMMARY ANOVA TABLE FOR PU NITIVEN ESS SOURCE Main Effects Distance I nformation 2-Way Interactions Distance Inform . Explained Residual Total 2.805 0 .640 2. 1 65 0.585 3.390 OF 4 1 3 3 7 1 5.050 56 1 8.440 63 0 .70 1 0 .640 0 .722 0 . 1 95 0.484 0 .269 0.293 F 2.609 2.381 2 .685 0 .726 1 .802 233 p 0.045 0. 1 28 0.055 0.541 0 . 1 05 SUMMARY ANOVA TABLE FOR FRIEN DLI N ESS SOURCE Main Effects Distance Information 2-Way Interactions Distance Inform. Explained Residual Total 2.020 0.062 1 .958 2.478 4.499 OF 4 1 3 3 7 42. 1 31 56 46.629 63 0.505 0.062 0 .653 0 .826 0.643 0 .752 0.740 F 0.671 0.083 0.867 1 .098 0 .854 234 p 0.61 5 0.774 0.463 0.358 0.548 SUMMARY ANOVA TABLE FOR AGG RESS I O N SOURCE Main Effects Distance Information 2-Way Interactions Distance Inform. Explained Residual Total 5 . 1 95 0 . 1 1 4 5.081 0.926 6. 1 21 OF 4 1 3 3 7 56.984 56 63.069 63 MS 1 .299 0 . 1 1 4 1 . 694 0.309 0.874 1 .0 1 7 1 .001 F 1 .277 0 . 1 1 2 1 .665 0.304 0.860 235 p 0.290 0.739 0 . 1 85 0.823 0.544 SUMMARY ANOVA TABLE FOR LIKI N G SOURCE Main Effects Distance Information 2-Way Interactions Distance Inform. Explained Residual Total 2. 1 53 0.526 1 .627 2.947 5.099 OF 4 1 3 3 7 33.025 56 38. 1 24 63 0 .538 0 .526 0.524 0 .982 0 .728 0 .590 0 .605 F 0.91 2 0.891 0.920 1 .666 1 .235 236 p 0.463 0.349 0.437 0 . 1 85 0.299 SUMMARY ANOVA TABLE FOR HOSTI LITY SOURCE Main Effects Distance Information 2-Way Interactions Distance Inform. Explained Residual Total 8.662 4.306 4.357 2.227 1 0 .889 OF 4 1 3 3 7 1 50 . 1 85 56 1 6 1 .074 63 2 . 1 66 4.306 1 .452 0 .742 1 .556 2.682 2 .557 F 0.808 1 .605 0.542 0 .277 0.580 237 p 0.526 0.21 0 0.656 0.841 0 .769 APPENDIX E : ANCOVA TABLE FOR CROWDING (STUDY ONE) CONTROLLING FOR AROUSAL. SUMMARY ANCOVA TABLE FOR CROWDING SOURCE Covariates Arousal Main Effects Distance I nformation 2-way I nteractions Distance Inform. 42.793 1 1 8 .61 9 1 1 1 .646 6.861 1 6.482 1 4 1 3 3 F 42.793 26.249 29.655 1 8. 1 90 1 1 1 .646 68.482 2 .287 1 .403 5 .494 3.370 238 P 0.000 0.000 0.000 0.252 0 .025 239 APPENDIX F: ANOVA TABLES FOR SELECTED DEPENDENT VAR IABLES SHOWING IND IVIDUAL ANALYSIS FOR STUDY ONE. SUMMARY ANOVA TABLE FOR AROUSA L SOURCE Main Effects Distance Information 2-Way I nteractions Distance Inform. Explained Residual Total 33.675 1 4.450 1 9 .225 1 7.325 51 .000 OF 4 1 3 3 7 1 665.800 31 2 1 71 6 .800 31 9 8.41 9 1 4.450 6.408 5.775 7.286 5.339 5.382 F 1 .577 2 .706 1 .200 1 .802 1 .365 P 0. 1 80 0. 1 0 1 0.31 0 0.357 0.220 SUMMARY ANOVA TABLE FOR CROWDING SOURCE Main Effects Distance Information 2-Way Interactions Distance Inform. Explained Residual Total 706. 1 37 667.01 2 39. 1 25 58.31 3 764.450 OF 4 1 3 3 7 2 1 28.550 31 2 2893.000 31 9 1 76.534 25.876 667.0 1 2 97.770 1 3.042 1 .9 1 2 1 9.438 2.849 1 09.207 1 6.007 6.822 9 .069 240 p 0.000 0.000 0. 1 28 0.038 0.000 SUMMARY ANOVA TABLE FOR CON F I N E M E NT SOURCE Mai n Effects Distance Information 2-Way Interactions Distance Inform. Explained Residual Total 222.475 1 83.01 3 39.462 31 .81 3 254.287 OF 4 1 3 3 7 231 7.200 31 2 2571 .487 31 9 55.6 1 9 1 83.0 1 3 1 3. 1 54 1 0.604 36.327 7.427 8.06 F 7.489 24.642 1 .771 1 .428 4.89 1 241 0.000 0.000 0. 1 53 0.235 0.000 242 APPENDIX G : EXPERIMENTAL PRE- INSTRUCTIONS (STUDY TWO) The three sets of experimental pre-instructions differed for each of the three i nformation conditions. These three conditions were ; contro l , spatial confirmation , and spatial disconfi rmation . Con trol con dition. "Hel lo , thank you for volunteering for this experiment. Before the experiment begins you wi l l be provided with the opportunity to practice some of the kinds of tasks which you wi l l be asked to complete later. The experimental assistant, Helen, wi l l now hand you a copy of the practice anagrams booklet and an answer sheet. Please read the i nstructions on the cover of the booklet but do not begin the fi rst item unti l you are instructed to do so." [25 seconds si lence to al low subjects reading time] "You may begi n the first item, and each subsequent item on hearing the sound of the prompt." 243 PROMPT [30 seconds] PROMPT [30 seconds] PROMPT [30 seconds] " If you have not already finished , please stop work now. Helen wi l l col lect you r answer sheets and practice anagram booklets. You wi l l then be shown i nto the experimental room." Spatial co nfirm ation cond ition . These instructions were exactly as for the control condition , with the fo l lowing addition after "experimental room". " I n this experiment we wi l l be looking at g roup performance and g roup i nteraction under varying interpersonal distances. In everyday life we are often exposed to distances between ourse lves and others that we are unconscious of but that may affect our behaviour. The aim of the experiment is to examine the effect of interpersonal distance on behaviour. You should, therefore , expect to sit at an unusually close distance to other subjects." S p at i al di sconfi rmation con dition. 244 These instructions were identical to the spatial confi rmation condition , with the exception of the last sentence which read as fol lows: "You should, therefore , expect to sit at a normal distance to other subjects. " 245 APPENDIX H : PRACTICE AND TEST ANAGRAMS PRACTICE ANAGRAMS The three practice anagrams were presented to subjects in the form of a four page booklet. The cover of the booklet was titled , "PRACTICE ANAGRAMS" and contained the fo l lowing instructions: "This booklet contai ns three practice anagrams simi lar to the type you wi l l be expected to complete later in the experiment. Each item is presented on a separate page, with the fi rst item appeari ng on the next page. Your task is to rearrange each group of letters so they make a mean ingfu l Engl ish word. You have 30 seconds to complete each item. When you hear the prompt on the tape recorded message, please turn the page and beg in the fi rst item . Do not turn the page for the next item unti l you hear the prompt. Do not write in this booklet, but p lace you r answers on the answer sheet provided." The three anag rams appeared, one on each of the next th ree pages of the 246 booklet. The actual anagrams were taken from Feather & Simon (1 971 ) and were as fo l lows: evol t r, middel , tmomen. The actual anagrams for these words were : revolt, m iddle , moment. During the debriefing session subjects were asked if they found these easy to complete and nearly all subjects reported getting all th ree correct. TEST ANAGRAMS Two sets of test anagrams were requi red , an easy set and a difficult set. The easy set was the same as that used by Feather & Simon ( 1 971 ) , whi le the difficu lt set they had used was comprised of both difficult and i nsoluble items. The requirement for the current study was for a set of difficult items; therefore, a new set was developed. The set of easy items was as fol lows : w it hn i, w a dn et, bunmer, pol iec, onersp, damaeg, gen cha, teffec, mn ega a, ariver. The actual anagrams for these words were : withi n , wanted, number, police , person , damage, change, effect, manage, arrive. The set of difficu lt items were as fol lows : ddeoii, t cajeb, aseccr, ppoerl , bilmre, ccorus, ibesr d, iowutt, ptrodi, ratply. The corresponding anagrams were as fo l lows : iodide, abject, scarce, prope l , l imber, crocus , b rides, outwit, torpid, paltry. 247 To ensure that the two sets of anagrams were in fact sufficiently "easy" or "difficu lt", the two sets were pilot tested on ten individuals for each set. As one of these two sets of anagrams had never previously been used, it seemed important to establ ish that the two sets were in fact different , and that one was much more difficult than the other. The resu lts of this proved to be satisfactory with the pi lot g roup on ly averaging 3.5 correct out of 1 0 possible for the difficult set. On the easy set the scores reflected the ease of the items with an average of 8.1 correct out of a possible 1 0. 248 APPENDIX I : POST-EXPERIMENTAL QUESTIONNAIRE (STUDY TWO) INSTRUCTIONS The fo llowing questions are designed to assess your reactions to the experiment. For each question p lace a cross (X) in the space which most accurately reflects you r answer. Consider, for example, the following question : How warm do you expect the weather to be today? Extremely hot _ : _ X _ : _ : _ : _ :_: _ : _ : _ : _ Extremely cold The above response indicates an expectation of a very hot day. 1 . Before entering the experimental room, how close did you expect you wou ld be seated to other subjects? Not at al l close _ : _ : _ : _ : _ : _ : _ : _ : _ : _ Very close 2. Duri ng the experiment to what extent did you experience rapid heartbeat? Very much _ : _ : _ : _ : _ : _:_: _:_: _ Not at al l 3. Was there a leader in your group? YES/NO (Ci rcle one response) If so, identify by letter A B C D E (Circle one response) 4. Was your personal space invaded? YES/NO (Circle one response) If so, to what extent? Severely _ :_:_:_:_:_:_ :_:_:_ Slightly 5. In completing the tasks were you able to concentrate? Unable to _ :_._._:_._:_ :_:_._ Able to concentrate concentrate 6. How bothered were you by the temperatu re in the experimental room? Very bothered _:_:_:_:_:_:_:_:_ :_ Not at al l bothered 7. How wi l l ing would you be to participate in another, simi lar experiment? Very wi l l ing _:_:_:_: _ :_ :_:_:_: _ Not at all wi l l ing 8. To what extent did you experience nervous tension? Not at al l _ :_:_:_:_:_:_:_:_:_ Very much 9. How bothered were you by invasion of your own space? Not at al l _:_:_:_:_:_:_:_:_:_ Very bothered bothered 249 250 1 0. Rate how nervous each group member was (excluding yourse lf) . G roup member A Calm · . . . . . . . . Nervous ---------- B Calm · . . . . . . . . Nervous ---------- C Calm · . . . . . . . . Nervous ---------- D Calm · . . . . . . . . Nervous ---------- E Calm · . . . . . . . . ---------- Nervous 1 1 . During the experiment did you find the experimenter helpful or unhe lpfu l? Very helpfu l _ :_:_:_:_:_:_:_:_:_ Very unhelpfu l 1 2. How bothered were you by the number of people in the room? Not at al l _:_:_ :_:_:_:_:_:_:_ Very bothered bothe red 1 3. H ow relaxed or i l l at ease did you feel during the experiment? Relaxed _:_:_:_:_ :_:_:_ :_ :_ I I I at ease 1 4. How crowded did you feel during the experiment? Not at al l _ :_ :_ :_:_:_:_:_:_ :_ Very crowded crowded 1 5. How aroused or unaroused did you fee l duri ng the experiment? Aroused _:_ :_ :_:_ :_:_:_ :_ Unaroused 251 1 6. To what extent did you experience rapid breathing? Very much _ :_:_ :_ :_:_:_:_:_:_ Not at al l 1 7. How much did you l ike the people in your group? Disliked _ :_ :_ :_ :_:_ :_:_: _ :_ Liked very much 1 8. How enjoyable was you r experience as a subject in this experiment? Very enjoyable _ :_:_:_ :_ :_ :_:_ :_:_ Not at al l enjoyable 1 9. How bothered were you by the closeness of other subjects? Not at all _ :_:_ :_ :_:_ :_ :_:_ :_ Very bothered bothered 20. How confi ned did you feel during the experiment? Not at al l _. _ : _:_ :_:_:_:_ :_:_ :_ Very confined confi ned 2 1 . How sluggish or alert did you fee l during the experiment? Sluggish _ :_:_:_:_: _ :_ :_:_ Alert 22. During the experiment how bothered were you by noise? Very bothered _ :_:_ :_ :_ :_ :_ :_ :_ :_ Not at all bothered 23. General ly do you prefer to work alone or in a group? Alone _: _ : _ :_:_ : _ :_: _ :_: _ I n a group 24. To what extent did you experience tense muscles? Not at al l _ :_:_ :_: _ : _ : _ :_:_: _ Very much 25. How bothered were you by the lighting in the experimental room? Not at al l _ :_:_:_ :_:_: _ :_: _ :_ Very bothered bothered 26.How pleasant did you find the experimental room? 252 Not at all _ : _ : _ : _ : _ : _ : _ : _ : _ :_ Very pleasant pleasant 27. To what extent did you experience sweating palms? Very much _ : _ : _ : _ : _ : _ :_:_:_:_ Not at all 28. How distracted did you feel during the experiment? Not at al l _:_:_: _ : _ :_:_:_:_:_ Very distracted distracted 29. How cramped did you feel during the experiment? Very cramped _ :_:_: _ :_:_:_: _ :_:_ Not at al l cramped 30. Rate how friendly each g roup member was (excluding yourself) Group member A Not at a l l _:_:_:_:_:_:_:_:_:_ Very friendly friendly B Not at al l _:_:_:_:_:_:_:_:_:_ Very friendly friendly C Not at al l _:_:_:_:_:_:_:_:_:_ Very friendly friendly D Not at al l _:_:_:_:_:_:_:_:_:_ Very friendly friendly E Not at al l _:_:_:_:_:_:_:_:_:_ Very friendly friendly 3 1 . Rate how aggressive each group member was (excludi ng yourse lf) G roup member 253 A Very _:_:_:_:_:_:_:_:_:_ Not at al l aggressive aggressive B Very _:_:_:_:_:_:_:_:_:_ Not at al l aggressive agg ressive C Very _:_:_:_:_:_:_:_:_:_ Not at al l aggressive agg ressive 254 D Very . . . . . . . . . . . . . . . . . . ---------- Not at al l aggressive aggressive E Very _ : _ : _ : _:_ :_ :_:_:_: _ Not at al l aggressive aggressive 32. How happy were you with the performance of your group? Very happy _ : _ : _ : _:_:_:_ : _ : _: _ Not at al l happy 33. How comfortable did you feel during the experiment? Very comfortable _:_ :_:_:_:_:_ :_:_:_ Not at al l comfortable 34. How dul l or jittery did you feel duri ng the experiment? Dul l _ : _:_:_:_:_ :_:_:_ Jittery 35. How stimulated or relaxed did you feel during the experiment? Stimu lated _ : _ :_:_:_ :_ :_: _ :_ Relaxed 36. How excited or calm did you feel during the experiment? Excited _ :_ :_ :_:_:_ :_ :_:_ Calm 37. How sleepy o r wide awake did you feel during the experiment? Sleepy : : : : : : : : Wide awake --------- 255 APPENDIX J : CROWDING RESEARCH NOTES TO SUBJECTS (STUDY TWO) KEITH TUFFIN DEPARTMENT OF PSYCHOLOGY MASSEY UN IVERS ITY Fi rstly I wou ld l ike to thank you again for vo lunteering as research subjects for the experiment which was conducted earl ier this year. You may recal l that fo l lowing the experiment you were debriefed. The major purpose of the debriefi ng was to reveal the deceptions invo lved and also to inform you of the questions which the research hoped to answer. You were also to ld that at the conclusion of the research you would be posted a summary of the fi ndings of the research. The reason for this summary is my commitment to the idea that as research subjects you deserve to be informed about the research you were i nvo lved in . You have volunteered your time and I bel ieve that you are entitled to something i n return - in this case it takes the form of information relevant to the experiment in which you participated. This summary wi l l be presented in four parts : ( 1 ) Theory, (2) Design , (3) Results, (4) Conclusion . 256 THEORY The experiment which you took part in was designed to examine some of the propositions of an information expectancy mode l of crowding. Essential ly the model predicts that people develop spatial expectations (ideas about how close they will be to others) based on information which they have about the particu lar environment they expect to be in. The information expectancy model predicts that when spatial expectations are confirmed the experience of crowding wi l l be reduced, i . e. perceptions of crowding wil l be lessened ; subjects wi l l feel less discomfort and performance deficits will not be evident. Also the model predicts that d isconfirmation of spatial expectancies wi l l serve to exacerbate the experience of crowding, i .e. increase perceptions of crowding, i ncrease discomfort and possible performance deficits. D ES I G N The purpose of this section is to give you an idea of the way the experiment was "set up", o r organised. Two variables were manipu lated. The fi rst of these was spatial expectation , with three levels of information , control (no information) , spatial confi rmation (accurate information) , and spatial disconfi rmation ( inaccurate information) . A l l subjects sat very close with on ly subjects from the spatial confirmation condition expecting to sit at this distance. The second variable which was manipu lated was the level of difficulty of the 257 anagrams which all subjects completed. Subjects were in it ial ly led to expect easy anagrams, th is expectation being either confirmed in the case of easy anag rams or disconfi rmed in the case of difficu lt anagrams. RESU LTS These resu lts are based on a pre l iminary analysis of the data. The i nformation expectancy model is based on the premise that subjects wi l l use available information i n formulating expectations about spatial concerns. This prediction was supported by the results which showed that subjects who were led to expect crowding reported anticipating less distance between themselves than subjects who expected "normal" spacing , o r subjects who were not told what to expect. A key prediction was that confi rm ing spatial expectations (accurate i nformation) wou ld reduce perceptions of crowding. Crowding was measured by aski ng subjects how crowded, confined and cramped they felt. On none of these three measures was the prediction supported. A further prediction which was partial ly supported was that disconfi rmed nonspatial expectations would increase perceptions of crowding. This was so for the measure of crowding but not the other two measures. The prediction that confi rming subjects spatial expectations wou ld reduce thei r perceptions of arousal and discomfort was unsupported. However subjects confronted with difficult anagrams (disconfi rming their nonspatial expectation) 258 reported being sl ight ly more i l l-at-ease. The same subjects also reported feel ing sign ificantly less comfortable. Contrary to predictions either confi rming or disconfirmi ng spatial expectations fai led to i nfluence subjects' i nterpersonal re lations. Concerning the treatment of Joh n ny Rocco , a significant i nteraction was noted with spatial confi rmation subjects becoming more punitive when nonspatial expectations were disconfi rmed, whi le the reverse occurred for subjects whose spatial expectations were disconfi rmed. Unexpectedly, subjects whose nonspatial expectations were disconfi rmed rated their fe l low group members as less aggressive. No sign ificant d ifferences were noted on subjects' reports of negative affect (hosti l ity , depression or anxiety) . Group performance was measured by asking subjects to derive a number of smal ler words from a master word. Spatial expectations appeared to have no i nfluence on performance, whi le subjects whose nonspatial expectations were disconfirmed, outperformed their counterparts whose nonspatial expectations (easy anagrams) were confi rmed. C O N C L U S I O N Contrary to predictions this study has shown that spatial expectations bei ng confi rmed or disconfi rmed has litt le impact on the experience of crowding. Disconfirming nonspatial expectations did exacerbate the experience of crowding where subjects reported feel ing more crowded and less comfortable. They also showed improved performance on the group task and rated thei r fe l low group members as less aggressive. 259 I hope this b rief note is informative and serves to answer any questions which you , as subjects , have about the experiment. If you have any further questions then please contact me. During working hours I may be found in my office (P.322) on the thi rd floor of the Psychology Department. Te lephone : work 69 099 extension 7924 or home 76 234. Thank you for you r participation . Keith Tuffin . 260 APPENDIX K : ANOVA TABLES FOR KEY DEPENDENT MEASURES FOR STUDY TWO (GROUP ANALYSIS) SUMMARY ANOVA TABLE FOR HOW CLOSE S U BJECTS' EXP ECTED TO B E SEATED FROM EACH OTH E R SOURCE Main Effects Information Difficu lty Sex 2-Way Interactions Info. Difficu lty Info . Sex Difficu lty Sex 3-Way Interactions Info. Diff. Sex Explained Residual Total SS DF MS 29.349 27.935 0.480 0.934 1 .51 7 0 . 1 05 1 . 1 94 0 .21 8 4. 1 44 35.01 0 45.600 4 2 1 1 5 2 2 1 2 1 1 36 80.61 0 47 7.337 1 3.968 0 .480 0 .934 0 .303 0.053 0.579 0 .21 8 2.072 3 . 1 83 1 .267 1 .71 5 F 5.793 1 1 .027 0 .379 0 .738 0 .239 0.041 0.471 0 . 1 72 1 .636 2.5 1 3 P 0.001 0.000 0.542 0.396 0.942 0.959 0.628 0.681 0.209 0.0 1 8 SUMMARY ANOVA TABLE FOR PRACTICE ANAGRAMS SOURCE Main Effects Information Difficu lty 2-Way I nteractions Info. Difficu lty Explained Residual Total 0 .21 5 0 .095 0 . 1 20 0 . 1 50 0 .320 OF 3 2 1 2 5 4 .690 42 5.01 0 47 0 .072 0 .048 0 . 1 20 0 .053 0.064 0 . 1 1 2 0 . 1 07 F 0 .642 0 .425 1 .075 0 .470 0 .573 261 p 0.592 0.656 0 .306 0.628 0 .720 262 SUMMARY ANOVA TABLE FOR TEST ANAGRAMS SOURCE Main Effects I nformation Difficu lty 2-Way I nteractions I nfo . Difficu lty Explained Residual Total 396.495 2.042 394.453 0 .052 396.547 OF 3 2 1 2 5 27. 1 20 42 423.667 47 F 1 32. 1 65 204.680 1 .0 1 2 1 .581 p 0.000 0.2 1 8 394.453 61 0 .879 1 0.000 0.026 0 .040 0.961 79.309 1 22.824 0.000 0 .646 9 .0 1 4 SUMMARY ANOVA TABLE FOR C R OWDING SOURCE Main Effects I nformation Difficu lty Sex 2-Way I nteractions I nfo . Difficu lty I nfo. Sex Difficu lty Sex 3-Way I nteractions I nfo. Diff. Sex Explained Residual Total SS OF 66 .901 4 1 .820 2 39.241 1 25.840 1 51 .802 5 1 2.347 2 36.509 2 2.947 1 1 . 1 29 2 1 1 9 .833 1 1 1 83.660 36 303.492 47 1 6.725 0 .9 1 0 39.241 25.840 1 0.360 6 . 1 73 1 8.254 2 .947 0 .564 1 0.894 5. 1 02 6 . 457 F 3.278 0 . 1 78 7.692 5.065 2.031 1 .2 1 0 3.578 0.578 0 . 1 1 1 2 . 1 35 263 p 0.022 0 .837 0.009 0.031 0.098 0 .31 0 0.038 0.452 0.896 0.043 SUMMARY ANOVA TABLE FOR ANNOYED SOURCE Main Effects I nformation Difficu lty Sex 2-Way I nteractions I nfo. Difficu lty I nfo. Sex Difficu lty Sex 3-Way I nteractions I nfo. Diff. Sex Explained Residual Total 92.728 4 2 1 .665 2 1 7.041 1 54.023 1 32. 1 44 5 9 . 1 82 2 22. 1 82 2 0 .780 1 8 .927 2 1 33 .799 1 1 356.233 36 490.032 47 23. 1 82 1 0.833 1 7.041 54.023 6.429 4.591 1 1 .091 0 .780 4.464 1 2. 1 64 9 .895 1 0.426 F 2 .343 1 .095 1 .722 5.459 0 .650 0 .464 1 . 1 21 0.079 0.451 1 .229 264 p 0.073 0.346 0. 1 98 0.025 0.664 0.633 0.337 0.780 0.640 0.304 SUMMARY ANOVA TABLE FOR SOMATIC SOURCE Main Effects In formation Difficu lty Sex 2-Way Interactions I nfo. Difficu lty I nfo. Sex Difficu lty Sex 3-Way I nteractions I nfo. Diff. Sex Explained Residual Total 4.499 4 1 . 1 85 2 0 .368 1 2.947 1 1 0.308 5 3.435 2 5.051 2 1 .823 1 1 2.352 2 27. 1 59 1 1 1 1 2.233 36 1 39.392 47 1 . 1 25 0 .593 0.368 2.947 2 .062 1 .71 8 2 .525 1 .823 6 . 1 76 2 .469 3 . 1 1 8 2.966 F 0.361 0 . 1 90 0 . 1 1 8 0 .945 0.661 0 .551 0 .81 0 0.585 1 .981 0 .792 265 p 0.835 0.828 0.733 0.337 0.655 0.581 0.453 0.450 0 . 1 53 0.647 SUMMARY ANOVA TABLE FOR AROUSAL SOURCE Main Effects Information Difficulty Sex 2-Way Interactions I nfo. Difficu lty I nfo. Sex Difficu lty Sex 3-Way I nteractions I nfo. Diff. Sex Explained Residual Total 1 3.01 3 4 3.81 2 2 6.901 1 2 .300 1 22.073 5 1 8.282 2 1 .284 2 2.507 1 1 9.067 2 54. 1 42 1 1 1 46.487 36 200.639 47 3.253 1 .906 6 .901 2 .300 4.41 5 9 . 1 41 0 .642 2 .507 9 .534 4.923 4.069 4.269 F 0 .799 0 .468 1 .696 0 .565 1 .085 2.246 0 . 1 58 0 .61 6 2 .343 1 .2 1 0 266 p 0.534 0.630 0.201 0.457 0.385 0. 1 20 0.855 0.438 0. 1 1 1 0.31 6 SUMMARY ANOVA TABLE FOR PUNITIVENESS SOURCE Main Effects I nformation Difficu lty Sex 2-Way Interactions I nfo. Difficu lty I nfo. Sex Difficu lty Sex 3-Way I nteractions I nfo . Diff. Sex Explained Residual Total SS OF 1 .986 4 0.292 2 1 .688 1 0 .007 1 3 .979 5 3 .879 2 0 .097 2 0 .007 1 4 .51 4 2 1 0.479 1 1 1 4.500 36 24.979 47 0 .479 0 . 1 46 1 .688 0.007 0 .796 1 .937 0 .049 0.007 2.257 0.953 0.403 0.531 F 1 .233 0 .362 4. 1 90 0 .01 7 1 .976 4.81 0 0. 1 21 0 .0 1 7 5 .603 2.365 267 p 0.31 4 0.699 0.048 0.896 0. 1 06 0 .0 1 4 0 .887 0 .896 0.008 0.026 SUMMARY ANOVA TABLE FOR AGGRESSION SOURCE Main Effects I nformation Difficu lty Sex 2-Way I nteractions I nfo . Difficu lty I nfo. Sex Difficu lty Sex 3-Way I nteractions I nfo. Diff. Sex Explained Residual Total SS OF 21 3 .023 4 33.81 5 2 1 65.763 1 1 3.444 1 52.099 5 33.522 2 1 3.287 2 5.290 1 5 .61 2 2 270.733 1 1 746.947 36 1 01 7 .680 47 53.256 1 6.907 1 65.763 1 3.444 1 0.420 1 6.761 6 .644 5.290 2.806 24.6 1 2 20.749 2 1 .653 F 2.567 0.81 5 7.989 0 .648 0.502 0 .808 0 .320 0 .255 0 . 1 35 1 . 1 86 268 p 0.055 0.451 0.008 0.426 0.773 0.454 0.728 0 .6 1 7 0.874 0.331 SUMMARY ANOVA TABLE FOR LI KI N G SOURCE Main Effects Information Difficu lty Sex 2-Way Interactions I nfo . Difficu lty Info . Sex Difficu lty Sex 3-Way Interactions Info. Diff. Sex Explained Residual Total SS OF 4.929 4 1 .755 2 0 .053 1 3 . 1 21 1 2 .040 5 0 .802 2 0 .701 2 0 .538 1 1 .841 2 8.81 0 1 1 1 7.000 36 25.81 0 47 1 .232 0.878 0.053 3 . 1 21 0 .408 0 .40 1 0 .350 0 .538 0 .920 0 .80 1 0 .472 0 .549 F 2.6 1 0 1 .858 0 . 1 1 3 6 .609 0.864 0 .849 0.742 1 . 1 39 1 .949 1 .696 269 p 0.052 0. 1 71 0.739 0.0 1 4 0 .51 5 0 .436 0.483 0.293 0. 1 57 0. 1 1 4 SUMMARY ANOVA TABLE FOR ANXI ETY SOURCE Main Effects I nformation Difficu lty Sex 2-Way Interactions I nfo. Difficu lty I nfo. Sex Difficu lty Sex 3-Way Interactions I nfo . Diff. Sex Explained Residual Total SS OF 1 .509 4 0.662 2 0.563 1 0.284 1 4.252 5 0 . 1 63 2 3.841 2 0.250 1 1 .972 2 7.733 1 1 1 1 .0 1 3 36 1 8.747 47 0 .377 0.331 0.563 0 .284 0 .850 0 .081 1 .920 0.250 0.986 0 .703 0.306 0.399 F 1 .234 1 .081 1 .841 0.930 2.780 0.264 6.277 0 .81 7 3.222 2.298 270 p 0.31 4 0.350 0. 1 83 0.341 0.032 0.769 0 .005 0 .372 0.052 0.030 SUMMARY ANOVA TABLE FOR HOSTI LITY SOURCE Main Effects Information Difficu lty Sex 2-Way Interactions Info . Difficu lty Info . Sex Difficu lty Sex 3-Way Interactions Info . Diff. Sex Explained Residual Total SS OF 0.900 4 0.860 2 0.030 1 0.01 0 1 1 .770 5 1 .260 2 0.260 2 0.250 1 0.607 2 3.277 1 1 9.81 3 36 1 3.090 47 0 .225 0.430 0 .030 0 .01 0 0.354 0 .630 0 . 1 30 0.250 0.303 0.298 0 .273 0.279 F 0 .825 1 .577 0 . 1 1 0 0.037 1 .299 2 .31 1 0 .477 0.9 1 7 1 . 1 1 3 1 .093 271 p 0.51 8 0.220 0.742 0.849 0.286 0. 1 1 4 0.625 0.345 0.340 0.394 SUMMARY ANOVA TABLE FOR DEPR ESSION SOURCE Main Effects Information Difficu lty Sex 2-Way Interactions Info . Difficu lty Info . Sex Difficu Ity Sex 3-Way I nteractions Info . Diff. Sex Explained Residual Total SS DF 0.833 4 0 .6 1 2 2 0. 1 41 1 0.080 1 0.745 5 0.332 2 0.351 2 0.063 1 0 .462 2 2.039 1 1 1 0.740 36 1 2.779 47 0 .208 0 .306 0. 1 41 0.080 0. 1 49 0 . 1 66 0 . 1 75 0 .063 0 .231 0 . 1 85 0 .298 0 .272 F 0.698 1 .025 0.472 0.269 0.499 0 .556 0.588 0.209 0.774 0.621 272 p 0.598 0.369 0 .496 0.607 0.775 0.578 0.56 1 0.650 0.469 0.798 SUMMARY ANOVA TABLE FOR PER FO R MANCE SOURCE Main Effects Information Difficu lty Sex 2-Way Interactions Info. Difficulty Info. Sex Difficu lty Sex 3-Way Interactions Info. Diff. Sex Explai ned Residual Total 735.986 4 277.875 2 396.750 1 6 1 .36 1 1 1 55.833 5 1 06.625 2 22.51 4 2 26.694 1 1 33.764 2 1 025.583 1 1 2903.667 36 3929.250 47 1 83.997 1 38.938 396.750 6 1 .361 3 1 . 1 67 53.3 1 3 1 1 . 257 26.694 66.882 93.235 80.657 83.60 1 F 2.281 1 .723 4.91 9 0.76 1 0 .386 0 .66 0 . 1 40 0.331 0 .829 1 . 1 56 273 p 0.080 0. 1 93 0.033 0.389 0.855 0.523 0.870 0.569 0.445 0.350