CRJJ This has on membe sPECifically individual j Parties to t itSelf’ and The MiChigan S ta (23 0f the g each Were ev ABSTRACT THE IMPACT OF DELIBERATION ON JUROR PERCEPTIONS OF TRIAL PARTICIPANTS CREDIBILITY AND OF VERDICTS AND AWARDS IN SIMULATED TRIAL SITUATIONS By Sunday Adefemi Sonaike This study examied the impact that deliberation has on members of trial juries. Attention was given specifically to changes, as a result of deliberation, in individual juror's perceptions of the credibility of the parties to the trial, in their evaluation of the trial itself, and in their verdicts and awards. The subjects were 101 undergraduate students in Michigan State University who simulated six-person juries (23 of the subjects whose groups had less than six members each were eventually dropped in the analysis leaving 13 six-person juries). The juries were shown a two and a half hour video- tape recording of a civil trial re—enacted from the tran- scripts of an actual case. The trial was presented in three segments and contained the evidence for both parties, the opening and closing statements of the two attorneys, and the judge's charge to the jury. Two first after after delibe measures on The second c_ first, plus The more stringe those involv the defendam dEIiberation the jurors wj Prominently C Jurors Change prominently Cf Sunday Adefemi Sonakie Two sets of questionnaires were administered, the first after the trial but before deliberation, the second after deliberation. The first questionnaire obtained measures on several variables related to juror perceptions. The second questionnaire contained some of the items in the first, plus others on reactions to the deliberation. The study found a tendency for the jurors to be more stringent in their assessment of the credibility of those involved with the plaintiff than those involved with the defendant. There was also a strong indication that ideliberation reinforced the predeliberation opinions of the jurors with respect to those issues that features most prominently during jury discussion. The.opinions of the jurors changed most significantly on those issues less prominently discussed, suggesting that the jurors were un- certain about these issues after deliberation. A comparison of the jurors' evaluation of the trial before and after deliberation showed that their findings about the trial improved with deliberation on the socio- emotional dimensions but worsened on the more task-relevant factors. 0n jurors' awards to the plaintiff, the study found a tendency for both these awards and the jury (group) awards to vary widely making the median of individual awards within groups before predictors of the jury awards than the mean individual awards. Sunday Adefemi Sonakie The jurors were more satisfied with their post- deliberation awards the less these awards differed from the group awards. Also, the less the group awards differed from the predeliberation awards of the individual jurors, the more satisfied were the jurors with the interaction as a whole. There was no statistically significant evidence that sex made a difference in any tendency of jurors to change their verdicts and awards. Accepted by the Faculty of the Department of Communication, College of Communication Arts and Sciences, Michigan State University, in partial fulfillment of the requirements for the Doctor of Philosophy Degree. [15/944 26? f Zflfllw Director of Thesis Guidance Committee: THE IMPACT OF DELIBERATION ON JUROR PERCEPTIONS OF TRIAL PARTICIPANTS CREDIBILITY AND OF VERDICTS AND AWARDS IN SIMULATED TRIAL SITUATIONS By Sunday Adefemi Sonaike A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication 1976 Sunday Adef. 1976 (:> Copyright by Sunday Adefemi Sonaike 1976 TO MY PARENTS ii members: and acac Joseph L possibi] StOpped Wessel V Erwin Be in diffe tiOn to FaTaCe 1 selflesE Stand w}- ACKNOWLEDGEMENTS I have to express my deep gratitude to my committee members: to Dr. Gerald Miller for showing that the legal and academic communities have a lot in common, to Dr. Joseph Woelfel for Opening my eyes to the exciting possibilities of research, to Dr. Donald Cushman who never stopped assuring me that I could do it, and to Dr. David Wessel who still found time for me in his busy schedule. I also wish to thank Dr. Bradley Greenberg, Dr. Erwin Bettinghaus and Dr. Lawrence Sarbaugh all of whom, in different ways, were sources of intellectual inspira- tion to me during my graduate career, and Dr. Richard Farace through whom I learned a lot about myself. I have to thank my family too: my wife for her selfless devotion, and Tayo and Timi who could not under- stand why they had to go away for a year. iii TABLE OF CONTENTS LIST OF TABLE S Chapter I II III IV INTRODUCTION THE JURY AS A SOCIAL GROUP The Nature of Small Groups The Jury as a Social Group ,The Importance of Interaction Interaction and Communication Group Basis of Perception FACTORS RELEVANT TO JURY INTERACTION The Juror Standard of Equity Review of Non-Empirical Literature Review of Empirical Literature Methodological Problem in Studying Jury Interaction RATIONALE AND HYPOTHESES Major Variables Perceived Credibility of Trial Parties Relative Prominence of Trial Factors Dogmatism Scale Evaluation of Trial and Interest in Future Jury Duty Information Retained and Range of Responses Award to the Plaintiff An Index of Conflict Satisfaction with Interaction and Group Award Decision on Negligence of Defendant Access to Videotape Trial Record . Foreman Role and Participant's Sex Prior Car Accident Experience Operationalization of the Variables Perceived Credibility of Attorneys and Witnesses iv Page vi VI VII APPENDICES Perception of the Trial Information Retained and Range of Responses Degree of Satisfaction and Trial Factors METHOD Composition of the Juries Procedure Some Methodological Considerations RESULTS Impact of Jury Size A Word on Awards Test of the Hypotheses Change in Credibility Comparison with adult sample findings Prominence of trial-related factors Influence of dogmatism Evaluation of the trial Trial information retained Group awards Satisfaction with interaction Decision on negligence of the defendant Access to trial record Foreman role and sex Prior car accident experiences CONCLUSIONS AND RECOMMENDATIONS Summary of the Findings Discussion of the Findings Limitations of this Study Practical and Research Implications BIBLIOGRAPHY Page 115 117 117 123 131 131 133 157 Table 3a 3b 3c 4a 4b LIST OF TABLES Consistency and discrepancy between jurors' predeliberation verdicts and group verdicts for juries of different sizes Effect of group verdicts on pOStdelibera- tion verdicts of discrepant jurors for juries of different sizes T-tests of mean differences between pre- deliberation and postdeliberation credibility ratings for the plaintiff, the defendant, and their attorneys and witnesses T-tests of mean differences between pre- deliberation and postdeliberation ratings on credibility dimensions for the plain- tiff, the defendenat, and their attorneys and witnesses Ratings of credibility for the contesting attorneys by adult sample and by jurors in this study Jurors' assessments of the degree of prominence given trial-related factors during jury deliberation Correlation between the degree of prom- inence given trial-related factors and change in credibility ratings for the relevant trial participants Page 86 86 88 91 92 93 94 Frequency of agreement between predelibera- tion verdicts and postdeliberation ver- dicts for jurors at three levels of dogmatism Table of correlations for testing specified hypotheses vi 96 98 Table ‘III 103 ll c.1th 12 Ttvs l3 Table Page 7 T-tests of mean differences between pre— deliberation and postdeliberation evalua- tion of the trial for sub-variables that showed significant differences 100 8a Rotated factor matrix for predelibera- tion trial evaluation 101 8b Rotated factor matrix for postdelibera- tion trial evaluation 101 9 Central tendencies of responses by jurors to four randomly selected information- retention items before and after de- liberation 102 10a Comparison of mean predeliberation awards within juries and mean of group awards 104 10b Mean, median and mode for individual predeliberation awards and group awards 105 10c Comparison of median and mean of in- dividual awards within groups with group awards 106 11 Relationship between variation in in- dividual predeliberation awards and duration of deliberation 107 12 Chi-square test of satisfaction with the interaction for jurors who conformed to the decisions of the juries and those who did not 109 13 T-tests of mean differences in satisfac- tion between juries allowed access to the videotape trial record and juries without such access , 110 14 Frequency of men and women chosen as jury foremen 111 15 T-tests of mean differences in awards in- fluence and satisfaction with interaction between jury foremen and regular jury members 111 vii Table 16a 16b l7 l8 Table 16a 16b 17 18 Change in negligence verdicts for male and female jurors Predeliberation verdicts for male and female jurors T-tests of effect of sex and prior car accident experience on awards of jurors Change of verdicts for jurors with prior car accident experiences and jurors with- out such experiences viii Page 112 113 114 115 TheI tant instit- under the c, U31 States. called to s Winick notes System than from Whose 1 Ninick, 1961 rely more he haps any 0th Um United 5 as five perc tine (813115: America. We ultimately f (De TOQquevi NOt SU on the 3161- 1 t CHAPTER I INTRODUCTION The system of trial by jury has emerged as an impor- tant institution in America, recognized by and protected under the constitution of the United Stateszuulthe individ- ual states. Every year, more than a million jurors are called to serve on jury trials throughout the country. Winick notes that Americans place more reliance on the jury system than any country in the world -- including England from whose legal system the American system derives (C. Winick, 1961a). This is understandable since Americans rely more heavily on the legal system in general than per- haps any other country. Blaustein and Porter estimate that the United States has so many courts that perhaps as many as five percent of its lawyers serve as judges at any one time (Blaustein and Porter, 1954; C. Winick, 1961b). De Tocqueville, commenting on the working of democracy in America, wrote that most major, as well as minor, problems ultimately find their way into the courts of America (De Tocqueville, 1945 translation). Not surprisingly there has emerged a vast literature on the merits and demerits of the jury system, including numerous empirical inquiries dating back to the turn of the century. Despite this tremendous interest in the jury system, however, only a small group of researchers have concerned themselves with the decision-making process operating within the jury during deliberation. Particular attention has been given to the social and psychological characteristics that individual jurors bring with them into the jury experience, and how these characteristics relate to the interaction of jury members and to the outcomes of their deliberation. However, little attention has been given to the other side of this coin, namely, how the judicial context, and more specifically the jury delibera- tion process, affects the perceptions and verdicts of individual jurors. The need for serious exploration of this question was firmly established by Miller (1975). Based on experi- ences gathered during research into the impact of video- tape technology on legal trials, Miller observed that juries may have some difficulty comprehending and utilizing court instructions, legal issues regarding probable cause, legal concepts such as negligence, reasonable doubt, and so on. He further noted some reluctance on the part of jurors to solicit additional information or clarification from the judge or court reporter, and declared: In addition to these issues, answers to questions regarding: (l) the degree to which jurors explic- itly consider inadmissible testimony during deliberation; and (2) the degree to which the deliberation process affects and alters individual perceptions and verdicts, are of great interest to jurists and communication scientists alike. These questions demonstrate the existence of a unique judicial problem. That is, how do legal pro- cedures affect jury deliberation and verdicts? (Miller, 1975, p. 14) The dearth of studies of the decision-process in juries is traceable to a large degree to the problem re- lating to access to juries during deliberation. Existing statutes dictate that the deliberation of juries be kept confidential and forbid the invasion of jury privacy. In effect, many of the studies of the jury process have relied on post-deliberation interviews with jurors which tend to have low reliability (Zimroff, 1974). A majority of studies used written accounts of trials to elicit the responses of individual, simulated jurors (Stone, 1969; Kaplan and Simon, 1972; Friend and Vinson, 1974; Jacobson and Berger, 1974). In Miller and Siebert (1975), and Williams et a1. (l975),role-playing juries watched video- taped presentations of trials, while in two Strodtbeck studies (Strodtbeck et al., 1956, 1957), mock juries listened to audio recordings of a trial and then delib- erated. Faced with the legal limitations on access to juries, many of the studies cited above attempted to simulate real trial situations as closely as possible. Strodtbeck et a1. (1957), for example, used actual potential jurors drawn from jury lists and utilized real court officials during the presentation of the trial to the juries. In a similar vein, Williams et a1. (1975) used potential jurors who were shown the videotaped trial in an actual courtroom. In an even more rigorous attempt to satisfy what Anapol (1974) describes as ”ecological validity", Miller and Siebert and their research team not only used potential jurors but had a real judge, a real court bailiff, and professional actors and actresses who re-created the pro- ceedings of court trials from the transcripts of the actual trials (Miller et al., 1974; Miller and Siebert, 1974; Miller and Siebert, 1975). These studies, however, still showed one serious shortcoming; except for the Strodtbeck studies, they did not allow for the impact of deliberation, an important jury activity. Rather, their findings were based on the reactions of individual jurors acting outside the influences of the group. Miller (1975) provided a rationale for this procedures in his research into the impact of videotape technology on legal trials: we wished to study the potential influence of the mode of pre- sentation per SE without possible confounding because of group process variables associated with deliberation" (p. 5). This is a tenable argument; nevertheless deliberation is an integral part of the jury system. Anapol neatly summa- rizes the "case" for deliberation in his criticism of the use of written transcripts and of individual responses per s2 in jury research: these designs (depart) from the concept of eco- logical validity in important ways: the inter- action involved in a jury decision-making pro- cess is lost when the jury does not function as a group and individual decisions are made; sig- nificant channels of communication are lost when the visual and/or audio aspects of the trial are eliminated: the personality characteristics of lawyers and witnesses are not readily transferred to paper; and the loss of the courtroom atmosphere (may) bring about a different attitude towards the task of jury decision-making. (Anapol, 1974, p. 3) In the Strodtbeck studies, the focus was the effect that factors operating prior to jury duty (e.g._§tatus and sex) have on the patterns of deliberation of juries. In I a similar manner, research into the factors relevant to jury interaction have largely examined characteristics of jury members that may have some impact on deliberation. A few researchers have turned this question around and con- sidered the effects that participation may have on jury members. Even here, however, the focus has been the narrow question of the impact of participation on the jurors' perceptions of the jury system or of the problems in group decision-making (see e.g. Erlanger, 1970, p. 354. Also, Moffat, 1945; Strodtbeck, 1957; Bevan, 1958; Broeder, 1959). While this question is legitimate and its exploration has provided important information on the working of the jury, it is no more compelling in its consequences for the jury system than the corollary question of the impact of deliberation on jurors as this relates to the case in_which they are participating as jurors, and to the outcome of their deliberation as a jury. Consequently, the purpose of this study is to de- termine the several aspects of the impact that deliberation may have on the members of trial juries. Specifically, we are interested in changes in individual juror's perceptions of the parties to the trial and of the trial itself, that may be attributable to the group phenomenon of delibera- tion. In the light of the widespread use of non-delib- erated verdicts in contemporary jury studies, the impor- tance of such a study cannot be overemphasized. If, as hinted by Kalven and Zeisel (1966:489), jury deliberation serves the purpose of achieving consensus through elim- inating minority opinions, then the simple aggregation of individual verdicts without deliberation can be used with greater confidence. If on the other hand, deliberation serves purposes other than, or in addition to, achieving consensus, an understanding of the nature of these effects is of critical interest to jury research. Furthermore, since individuals in many judicial districts go through the jury experience more than once, their impressions of one trial and of the deliberation process involved may affect their reactions to, and participation in, further jury duties. An investigation of jurors' perceptions of the trial they are involved in may throw some light on the problems that exist in this regard. Organization of the Study In the next two chapters of this thesis, the literature relevant to the study is critiqued. The lit- erature comes from two main areas of research: small-‘ group, and jury interaction, and is reviewed in that order. Chapter 4 contains the major hypotheses of the study, the rationale for testing these hypotheses, and the procedures that were used in operationalizing the variables. In Chapter 5, we present the design of the study and discuss certain methodological questions that arose in the course of planning the study. This is followed by the results and findings (Chapter 6). In the final chapter, we review the major findings, discuss their importance, and make rec- ommendations for further research in this area. _—____ —— ._ ——_~ A S} of the smali "behavior" ( general feet through rest Some aspect: t0 Jury int< CHAPTER II THE JURY AS A SOCIAL GROUP A synthesis of the literature on the functioning of the small group is essential to an understanding of the "behavior" of the jury. We shall begin by looking at the general features of the group that have been determined through research. Then we shall more specifically discuss some aspects of group functioning that more closely relate to jury interaction. The Nature of Small Groups Several definitions of a group have been put for- ward, corresponding to the various conceptual and theo- retical approaches that have been taken in appraising the group phenomenon. Groups have been defined in terms of one or more of the following characteristics: (1) composition, (2) interaction patterns, (3) organizational structure or roles, (4) interdependency, (5) duration of interaction, (6) setting of interaction, and (7) awareness of membership. These features concern the "what” and "who" of groups. Other approaches have focused on the "why" of groups. Notable among these are definitions in terms of (8) motiva- tion and need satisfaction, and (9) of achievement of 8 group goals. These characteristics fall broadly into the psychological and sociological approaches and we will, in turn, take a look at these several conceptions of the group. Multiple-unit composition is perhaps the most basic feature of groups but one usually taken for granted. A group exists only when two or more persons are involved. Most conceptions of a group begin_with reference to this composition aspect, usually with phrases such as "collec- tion of individuals..." (Bass, 1960), "plural number of separate organisms..." (Smith, 1945), "aggregate of in- dividuals..." (Brodbeck, 1958), "two or more individuals...” (Sherif and Sherif, 1956), "any number of persons...‘ (Bales, 1950), etc. While there is usually no direct discussion of multiple-unit composition as a prerequisite for a group, it is clear that a group must involve a mini- mum of two individuals. The nature and scope of this involvement have been at the center of the disagreement in delineating the essence of the group. Researchers are unanimous in the belief that the mere coming together of two or more in- dividuals is not enough to make a group; however, there is less agreement on what other conditions have to be satisfied. Brodbeck (1958), for example, argues that indi- viduals must stand "in certain descriptive (i.e., observ- able) relations to each other” before they would qualify 10 for description as a group. Roles and structure, then, are regarded by Brodbeck as the essence of the group. For Shaw (1971), Smith (1945), Bales (1950), Homans (1950), and Merton (1957), among others, the essential feature of a group is the interaction between the relevant individuals. Shaw further demands that this interaction be enduring (rather than transitory as, e.g., during an elevator ride), and must be such that "each person influences and is in- fluenced by each other person” (p. 10). Besides stipulating that interaction in a group should occur with some fre- quency, Merton adds two other provisos for group qualifica- tion: "that the interacting persons define themselves as 'members'" by possessing "patterned expectations of forms of interaction which are morally binding on them and on other 'members,‘ but not on those regarded as 'outside' the group." Furthermore, each person in a group must be "g3- fined by others (members and non-members) as 'belonging to the group'” (pp. 285-286). In a manner somewhat similar to Merton's, both Smith and Bales see the interaction of group members as involving some perceptual aspects. For Smith, this is "the collec- tive perception of (members') unity," for Bales, the "impression or perception of each other member" formed by each participant during every interaction. Bales further demands that the interactions of a group take place in face-to-face setting, a proviso that is shared by Campbell (1958) in 1 common spa In acteristic; phasized b; Sherif (19 groups. Tl System in V. certain imp and norms. the Organiz 11 (1958) in his presentation of "proximity" (sharing of a common space) as a necessary feature of a group. In addition to Brodbeck (1958), organizational char- acteristics and role relationships among members are em- phasized by McDavid and Harari (1968), and by Sherif and Sherif (1956) as the critical consideration in analyzing groups. This approach regards the group as an organized system in which members interact in accordance with a certain implicit or explicit but standard set of values and norms. The norms regulate role relationships between the organized "parts" or components (members) that make up the groups; these role relationships in turn create and sustain interdependency. This statement from Newcomb (1951) succinctly portrays the implications of 2952s and interdependency for group definition: For social psychological purposes, at least, the distinctive thing about a group is that its members share norms about something. The range covered by the shared norms may be great or small, but at the very least they include whatever it is that is distinctive about the common interest of the group members -- whether it be politics or poker. They also include, necessarily, norms concerning the roles of the group members —- roles which are interlocking, being defined in reciprocal terms ... These distinctive features of a group -- shared norms and interlocking roles -- presuppose a more than transitory relationship of interaction and communication. (p. 3) Conceptualization of the group in terms of the "why" of membership has been the concern primarily of social psychologists who have shown as much interest in the "indi multi-men approach of the ba they have have usua tence as (Bass, 19 (Schedlin as fruitf 1967; Fre T psycholog DeutSCh ( and % sociologi or coneC while the the indiv how it fu T2 in fact, I reality 0; Pei-SpeCth A11p0rt (1 12 the "individual" dimensions of groups as in the results of multi-membership interaction. Researchers utilizing this approach have sought answers to the several ramifications of the basic question: Why do individuals (granted that they have a choice) go into groups? And their "answers” have usually taken the following forms: individuals are motivated to form collectivities because "(their) exis- tence as a collection is rewarding to the individuals" (Bass, 1960), because they "identify with one another" (Schedlinger, 1952), or because each member sees membership as fruitful in achieving certain (common) goals (Mills, 1967; Freeman, 1936). The difference between the sociological and the psychological approaches to understanding the group led Deutsch (1968) to distinguish between sociological groups, and psychological groups. The essential feature of the sociological group, Deutsch argued, is in the interaction or collective consciousness of the interdependent members, while the essential feature of the psychological group is the individuals' percsptions of this interdependence and how it furthers the group goals. This distinction is by no means trivial. It has, in fact, been the basis of a long-drawn debate as to the reality of the group. Operating from a pscychological perspective, some scholars of group dynamics, notably Allport (1924), have argued that ”group" is a mere 13 abstraction, a concept that we use to account for "col- lective individual behavior." The only reality in a group, argues this school of thought, are the individuals, and groupness refers merely to the set of values, thoughts, ideas, habits and norms held in common (in their minds) by individuals in collectivities. Groups, in short, exist only in the minds of men! At the other extreme, Durkheim (1898), and Warriner (1956), among others, maintain that there is more to the group than "the simple sum of the parts." The questions deserving our attention, this school of thought argues, should not be psychological ones but only the ones that concern the interaction of members in a group. Middle positions have emerged between these two extremes, such as the "panels" argument of Cattell (1951a) which regards both the psychological and the sociological questions as equally deserving attention, and the entitativ- ity "theory" of Campbell (1958) in which the "realness" of a group is portrayed as lying on a continuum. These mod- erate positions, interestingly, turn out to be the plau- sible ones and are now more widely accepted. Campbell (1958) argued that the differences in "realness" of physical objects (chairs, tables, persons, etc.) compared to non-physical objects (ideas, names, words, etc.) are really differences in the extent to which our sense modalities are applied in the perception of 14 these objects. We perceive the chair as "real" because we can sss it and EQEEE it, feel its smoothness, EEé£.the the sound that results from tapping it, and perhaps smsll the wood it is made of. Information about non-physical -objects such as "group" comes to us through fewer sources than those employed in perceiving physical objects and may be less immediate and less compelling. Nevertheless, the mental process by which we "perceive" both physical and non-physical objects is essentially the same. Thus, our perception of the chair differs from that of the group as a concept only in the degree of "realness" for which Campbell applies the term "entitativity." Entitativity thus describes the extent to which different objects are psrceived as having real existence, that is existence as an entity. In an attempt to more clearly delineate the group as an entity from the group as the simple sum of the parts, several studies, notably Sherif (1935), Sherif and Sherif (1956), and Asch (1951, 1952, 1955, 1956), have established the reality of "group behavior" as distinct from the be? haviors of individuals outSide the context of the group. However, other studies of the influence of psychological variables on conformity (e.g., Crutchfield, 1955; Cervin, l955a,b, 1956; Bass, 1955, 1957; Barron, 1952; Beloff, 1958; Rosner, 1957; Block and Block, 1952) have demonstrated just as strongly that the nature and extent of group 15 influences on the individual is dependent to a large extent on his idiosyncratic characteristics. These studies and others of the same vein have been reviewed extensively elsewhere (Hare, 1962) and it is not our intention to duplicate this effort. For the purpose of this investigation, it is enough to note some general conclusions that have been generated by these studies. Generally, the greater the ambiguity of the object, the greater will be the influence of other group members in determining the judgment of the subject (Hare, 1962, p. 31). The ambiguity could involve situations in which the subject is confused as to the nature of the object or is simply unsure of the appropriateness of his interpretation. The object too could be something tangible (as for example a mechanical contraption) or something less tangible (as for example the prospect of a war between two nations). The danger of unfriendly attack from an external source has always triggered a strong bond between members of a nationality group perhaps because of the uncertainty (or ambiguity) as to the future that accompanies such a situa- tion. ‘ Secondly, the reactions of individuals to the various situations faced in a group (including ambiguity of the object) will vary depending on their personalities. In Cervin (1955a, 1955b, 1956), subjects with all combina- tions of two personality traits, anxiety and neuroticism, 16 were given social approval or disapproval of their opinions in discussion with two confederates. While no differences in behavior were observed between emotionally stable and unstable subjects when their opinions were praised, the high-neurotic subjects were more rigid in holding their opinions when under disapproval. Similar relationships between non-conformity and personality variables (neurot- icism, authoritarianism, submissiveness, etc.) have been found in experiments by Crutchfield and the other re- searchers cited earlier. Finally, the more most individuals desire member- ship in a group, the less likely they are to persist in deviant behavior. This is intuitively obvious since a deviant runs the risk of being ostracised or of losing membership. Goffman (1967), discussing the aspects of human interaction worthy of scientific interest, made the follow- ing comment: ... the proper study of interaction is not the individual and his psychology, but rather the syntactical relations among the acts different persons mutually present to one another. None- theless, since it is individual actors who con- tribute the ultimate materials, it will always be reasonable to ask what general properties they must have if this sort of contribution is to be expected of them. (p. 2) Thus, it seems there is strong argument for examining both psychological and sociological questions in the study of group dynamics. 17 A close examination of the quotation from Goffman above suggests that the sociological and the psychological groups may not be two types of groups at all; rather, their essential features may represent two aspects of a group. The individual comes into the group with certain idiosyn- cratic properties. Nevertheless, his perceptions within the group are inherently tied to the group consciousness. Furthermore, the different approaches to group definition, far from being unique and independent, do in fact overlap to a considerable extent. It requires only a little effort to see that the researchers, coming as they are from dif- ferent social science disciplines, are focusing on the dif- ferent but interrelated aspects of the same phenomenon. The members of a group must of necessity interact (directly or indirectly), and this interaction must be other than transitory. To regulate the interaction, some order of procedure becomes necessary. This is provided by norms and rules of behavior that manifest the underlying roles. The "commonness" created by conformity to norms and by interlocking roles become in fact a measure of the inter— dependency between members. Thus the different approaches to the definition of a group are not incompatible with one another and groups differ only in the extent to which they emphasize one or more of the features. In summary, then, we can say that a group, as dis- tinct from an "aggregate,' is composed of more than one 18 individual who, to differing extents, (a) engage in fre- quent and usually face-to-face interaction, (b) perceive themselves and are perceived by others as members, (c) play roles defined and regulated by norms, (d) jointly promote interdependent and rewarding goals, (e) have a collective perception of their unity, and (f) tend to act in a unitary manner towards their environment. The Jury as a Social Group Based on the conclusions drawn above, we should be on safe grounds to contend that a jury has the necessary attributes of a group. It has collective membership -- usually six or twelve. Its members engage in frequent, face-to-face interaction governed by well-established judicial rules and regulations. Members of a jury jointly promote interdependent goals and the requirement for con- sensus makes this more so than with the typical group. The fact that the individuals are subsumed under the common label of "jury" is proof that they are perceived by others as members of a group, and it is unlikely that any member of a jury will fail to perceive himself as a member of the group. Finally, in its deliberation and return of a I unanimous verdict, the jury acts in a unitary manner towards its environment. Juries have in fact been described as "a type of ad hoc problem-solving group" and been empirically studied 19 (Strodtbeck and Mann, 1956, p. 3). However, certain.dif- ferences between the group process as it operates in a jury and the process within other groups deserve to be mentioned. First, membership in a jury is only partially voluntary. Juries are drawn from the list of registered voters in a community who have been declared eligible for "jury service.‘ While, presumably, an individual can successfully object to serving in a particular jury, he cannot skip jury duty as a whole save through the formal procedure for exemption. Second, communication within juries follows an un- usual pattern. In the initial part of a trial, while the case for and against the defendant is presented, members of the jury interact only through their joint presence and common location or proximity. They are expressly forbidden to exchange opinions or discuss the case in issue. Com— munication between the members happens only during the deliberation period and serves the definite goal of arriving at some unanimous agreement on the guilt or innocence of the defendant, based strictly "upon the law and the evidence" presented in court._ In short, the jury system does not encourage fraternizing among members. Third, the level of formality around the actions of the jury throughout its standing is uncommonly high, no doubt as a consequence of the highly formal structure of 20 the legal system as a whole. Perhaps the greatest mani- festation of the formality of jury interaction is the fore- man who has to be chosen within the group before delibera- tion begins. Unlike the less formal groups, no room is given for the "natural" emergence of the jury leader. Fourth, there is a requirement on the jury to re- turn a unanimous verdict following its deliberation. This is a condition that is usually absent in other groups. While all groups probably desire consensus in the pursuit of their goals, most groups are satisfied with acceptable compromises. With the jury, on the contrary, split ver- dicts are unacceptable and a "hung" jury, especially in criminal cases, will usually be compelled to retire again and redeliberate until unanimity is achieved, or a retrial will be held with fresh jury. Finally, members of a jury are protected from facing grssp consequences of their behaviors and decisions for once a verdict is rendered, the group dissolves. This is a most important distinction when juries are compared to most on-going groups in which group members may be held accountable for their past and present actions. The Importance of Interaction We have already presented "interaction" as one of the primary factors for differentiating a group from an aggregate of individuals. The concept is not however re- stricted to the group-communication literature. In fact, 21 it is not restricted to the social sciences, for in physics and mathematics, there are ideas of interaction quite similar to our use of the word in the social sciences. McCall and Simmons (1966) illustrate this point with the example of Newton's Law of Universal Attraction: which states that the force of attraction between two bodies equals some constant times the ratio of the product of the masses of these bodies to the square of the distance be- tween them. This law, more generally known as the law of gravity, is a relational law defining attraction between bodies as a reciprocal or mutual kind of influence. Attrac- tion is conceived as a two-sided dependence, an interdepen- dence, rather than a one-sided dependence of effect upon cause. In statistics, interaction is said to have occurred when two or more independent variables produce results that cannot be ascribed simply to the effects of either of them or to the effects of their simple additive combination (see e.g. Blalock, 1972, pp. 230, 337-340; and Blalock, 1965, pp. 374-380). McCall and Simmons note that whenever a relationship of...inf1uence between two events cannot be resolved into a simple function of one but must instead be treated as a joint function, as a mutual or reciprocal influence, we have a case of interaction (McCall and Simmons, p. 47). Shaw in fact defines the group as two 93 more pgrsons who 22 are interacting with one another is such s manner that each person influences and is influenced by each other person (Shaw, 1971, p. 10). Shaw emphasizes that interaction re- . quires mutual influence: If one person, A, sees another person, B, with whom he wishes to speak and so approaches him, A is in- fluenced by B but not vice versa; hence interaction does not occur, and A and B do not constitute a group. However, if B notices that A is attempting to get his attention, B may be influenced to also approach A. In this case, A and B are interacting and so compose a group. Or consider the case of a person, A, who is looking up at the sky and is approached, independently, by two other persons, B and C, who also begin looking in the same direction. Again, no group exists, despite the fact that B and C have been influenced by A, because A has not been influenced by B or C. These three persons become a grou if they enter into a discussion (inter- action concerning the object of their attention... interaction requires mutual influence, and an aggregate of individuals is a group only if inter- action occurs (p. 10). The use of the concept of interaction in the phys- ical and the social sciences may appear different at first glance; they both, however, refer to the same issues. Interaction, in the physical science interpretation, may more adequately be labelled interaction effect, the result of the mutual influence of two or more bodies (or persons) working in conjunction. In the social sciences, the emphasis is on the process, rather than the effect; never- theless, it is also recognized that certain effects do re- sult from interaction. It is now obvious why "interaction" occupies such a principal position in the conceptualization of groups. 23 In contrast to the behavior of individuals in individual contexts, the behavior of individuals in group contexts cannot be fully explained by the examination of the simple individual characteristics. "Interaction" in the group con- text precludes a simple one-to-one relationship between in- dividual "attributes” within and outside groups, and this is the basis of the arguments of Sherif, Asch, and the other exponents of "group behavior", earlier discussed. Nevertheless, groups are still made up of individuals and few individuals totally lose their unique features as a result of group membership. Interaction and Communication It is typical to regard interaction as synonymous with communication and Shaw's reference to discussion as the condition for "groupness", in the above quotation, is a recognition of this fact. However, there are occasions when a distinction between communication and interaction may be required. Blau and Scott (1962) noted the following as one such distinction: The concept of social interaction focuses prin- cipally upon the formal characteristics of social relations: such terms as frequency, initiative, superordination, and reciprocity indicate its dimensions. The concept of communication, on the other hand, directs attention to the meaningful content conveyed in the encounter, and its char— acteristics are described by such terms as flow of messages, obstacles, positive and negative reactions, and exchanges. (p. 116) In a similar vein, Homans (1950) included under interaction 24 both verbal and nonverbal communication and declared: ...the word communication... may mean the content of the message, signal, or "communication" being transmitted, or the process of transmission it- self,...or the sheer fact, aside from content or process of transmission, that one person has com- municated with another. Only to the last of these do we give the name of interaction... (p. 37) Interpreting Blau and Scott, and Homans, Boster (1976) concluded that communication, in a group, is "the content of interaction, the meaning transmitted from person to person." (p. 8). In short, interaction and communication are "two sides of the same coin" (p. 7). This distinction between interaction and communica- tions seems quite relevant to the jury situation, perhaps more so than to other kinds of groups. As we have already mentioned, exchange of opinion between members of a jury is forbidden until all the evidence has been presented. Nevertheless, it is apparent that by virtue of their prox- imity during the presentation of evidence, and the realiza- tion of a common purpose, members of a jury must be inter- acting even before deliberation. The content of this inter- action, for which we have reserved the label communication, must of course be different from the content of interaction when the jury formally gets together to deliberate. Group Basis of Perception Manis (1955) found that the self-concepts of men living in a dormitory in four-man living units were in- fluenced by others' perceptions of them over a period of 25 months of living together. In Miyamoto and Dornbush (1956), individuals' ratings of themselves on four personality traits for groups of 8 to 48 persons were analyzed. The subjects' self-perceptions were compared with the self- perceptions of others in the group, with their perceptions of others' attitudes, and with their perceptions of the attitudes of "members of most groups" or a ”generalized other". The researchers found a relationship between the subjects self-perceptions and the feelings of others in their groups. This relationship was stronger when the self-perceptions were compared to the subjects' percep- tions of others' attitudes, and stronger still when self- perceptions were compared to the subjects' assessments of the generalized attitude. These and other studies (Cartwright, 1952; Allport, 1954; Zander, 1958 to name a few) have provided some under- standing of the vital link between the perceptions of in- dividuals and the groups or cultures to which they belong. It is now generally accepted that the individual's perception, at any given time, is a function of the attitudes of the society transmitted in culture, the more transient ps5- ceptions of the small group involved in the action at the monent, and an idiosyncratic com- ponent which results from the personality of the perceiver and the perceived and other unique situational factors. (Hare, 1962, p. 81) The degree to which the influence of the group pervades an individual's self-perception depends on the degree to which that individual is attracted to the group 26 and how much value the group members place on his participa- tion (Festinger, Torrey and Willerman, 1954; Stotland et al., 1957). The reasons are easy to see since an individual who is highly attracted to a group is likely to pay greater heed to the general opinions of the group. Furthermore, if his participation is valued, the groupmembers are likely to be more explicit in their assessment of him. One of the areas of others' influence that has re- ceived wide interest is that of the relationship between intimacy and perceptual accuracy. We can summarize the findings in this regard as follows. Generally, subjects tend to assume that they are liked by persons whom they like (Tagiuri, 1957); also, if a subject likes two other persons, he tends to perceive them as liking each other (Kogan and Tagiuri, 1958a). Pairs of friends tend to be more accurate in the perception of each other's personalities than pairs of non- friends, partly because of increased knowledge of each other as a result of continued interaction (Bieri, 1953; Taylor, 1957), and partly as a result of a desire to pro- ject one's own values on one's friend (Fiedler et al., 1952; Davitz, 1955). ~This last point conveys the tendency for subjects to describe persons whom they like best as more similar to themselves than those they like least (Lundy et al., 1955), a phenomenon most apparent perhaps in marital relationships (Corsini, 1956). 27 Friendship ties within larger groups appear to en- hance accuracy in assessing the perceptions of others in the group. Dymond et a1. (1952) and Gronlund (1955a) have shown that group members who have many friends in the group are usually more accurate in their assessment of the in- formal group structure and of the characteristics of others in the group than are relatively isolated individuals. The intimacy provided by marital and friendship ties is apparently more superior for accurate perception than close contact on a work basis. Block and Bennet (1955) asked a female psychologist to describe her typical interactive behavior with her professional associates and with some friends. The subject's descriptions regarding her friends known off the job were found to be in closer agreement with their own descriptions than did her des- criptions of interactions with professional associates. Married couples have also been found to make more accurate predictions of each other's responses than couples who were dating (Kirkpatrick and Hobart, 1954). In contrast to on-going groups, juries are usually ad-hoc, existing only for the duration of the trial. Furthermore, the interaction between members of a jury tends to operate more on a "work basis" than at the in- formal level of friendship. In this sense, we may expect the influence of the group on individual jurors' percep- tions to be less pronounced than in on-going groups, and 28 their judgement of others' perceptions to be less accurate. However, juries have other features that may compensate for their ad-hoc nature: they provide highly intensive interaction while they last! More than most on-going groups, juries emphasize the exchange of opinions through deliberation and the need for arriving at some common grounds. Thus each member of a jury may, perhaps without realizing it, be expecting to be influenced by, and to in- fluence, the other members during the deliberation exercise. An indication of this influence may be a modification of a juror's position as a result of deliberation. CHAPTER III FACTORS RELEVANT TO JURY INTERACTION The Juror Standard of Equity Although in theory juries are presumed to be factual, returning verdicts based strictly "upon the law and the evidence" presented in court, it is now generally accepted that juries give recognition to "values which fall outside the official rules" (Erlanger, 1970, p. 349). For example, ...(juries) tend to show leniency when the de- fendant has been punished enough (e.g. was hurt in the commision of the crime, has had great family misfortune since then, and the like), when the punishment threatened is "too severe", when another party involved in the crime and equally responsible received preferential treat- ment or was not charged, when the crime occurred in a "subculture", or, in some cases, when the police have used improper methods (Erlanger, 1970, pp. 349-350) It is not surprising therefore, that significant attention has been given by jurists and social scientists to the innumerable factors that are relevant to the inter- action of juries. We shall review the literature in this regard under two categories: writings of non-empirical (but widely accepted) foundation, and those based on hard- core research. 29 30 Review of Non-Empirical Literature The jury system developed as a means for counter- ing the injustices in the feudal society when the landlord was judge, jury, and sometimes accuser and prosecutor as well. The earliest jurors were neighbors of the accused person or contesting parties, and served primarily as char- acter witnesses. They were chosen on the belief that they knew most about the issue under contention, a condition that is no longer tenable under the present system. The change in this basic assumption of the old jury system has been traced to the expansion of population and the rapid flux of living in modern time (see Dudycha, 1955, p. 204). Under the present system, ironically, special measures are taken to ensure that the jurors have no prior relationships with either party in the contention nor pre- judicial knowledge of the issues to be debated. However, a more fundamental assumption still remains: we still assume that the fellow citizens of the contending parties are competent to evaluate and coordinate the evidence, and return a verdict conveying their impressions "beyond all reasonable doubt." Again, this assumption is less realistic today than it was in the society in which the jury system developed and this is responsible for perhaps the strongest attack on the jury system as it is currently organized. As Dudycha notes: We...assume that laymen -- housewives, farmers, businessmen, clerks -- are competent to evaluate 31 and coordinate evidence which runs the gamut from highly technical information to highly emotional appeals. (However), the variety of evidence now presented in litigations did not exist in the society which gave birth to the jury system (and) the basic difficulty in the jury system (now is therefore) the relative incapacity of the average juror to understand all that he must. (pp. 204-205) The problem is aggravated by the composition of the typical jury team. Although juries, in theory, are representative of the population, the jury pool, in practice, has several sources of bias. Strodtbeck, James and Hawkins (1957) noted that in the metropolitan areas of Chicago, St. Louis, and Minneapolis, lawyers, doctors, teachers, policemen and other local and federal employees, including elected officials, are excused from jury service. Aliens, foreign visitors, recent migrants and persons under 21, who are not eligible to vote, do not appear on the jury lists. Also persons who operate "one man" businesses and prospective jurors with pressing personal problems can have their jury service deferred or cancelled. "The net effect is that the professions and the very low education and occupation groups are slightly under-represented" (Strodtbeck et al., 1957, p. 713). Osborne suggests that to relieve the difficulty in jurors' handling of technical information, juries should be selected by "a competent jury committee of three to five members" (Osborne, 1937, p. 6). An even better idea, expressed by Dudycha, is that specialized panels of experts 32 be used as jurors and advisors to the court. ”Cases, for example, that are likely to involve testimony regarding medical problems would then employ a physician on the panel (and) those involving tax evasions or embezzlement might include an accountant or two." Dudycha argues that "such panels could not only decide the guilt or innocence of the defendant but could advise the court in a more flexible fashion than juries are now permitted" (p. 205). Various factors have been determined to influence jury opinions. One of these is prejudice, defined by Dudycha as "some strongly held belief for which little if any evidence exists” (p. 205). No one is entirely free of some prejudice; however, rigid adherence to prejudicial views destroys the open-mindedness desired in jurors and distorts jurors' perceptions in trials. Geographical or regional loyalties are another source of undesirable influence on jury opinions. Dudycha reports the incident of a murder committed in one of the Relocation Centers established in Arizona for Asiatics shortly after Pearl Harbor. The Center was under Federal supervision; however, the area was included in the legal jurisdiction of the county in which the Center was located. "Serious difficulties arose because many believed that a fair jury trial in the county would be impossible. No matter who sat in the jury, the accused would be convicted almost automatically because of the intense hostility against all Asiatics at that time" (p. 207). 33 There is one type of influence on the jury that we may label a product of our age. It concerns the biasing influence of pretrial publicity especially with regard to cases arising from sensational events. Today, the arm of the mass media is longer and stronger than it has ever been. With the sophisticated technology now available, events of significant import are reported across the length and breadth of the nation seconds after they occur, and in some cases even while they are taking place. If these events later become subjects of legal proceedings the con- stitution of an objective jury may present a difficult task. It appears that the prevailing attitude now is that the threat to justice and fair play is posed by the degree of emotional feelings generated by pre-trial pub- licity, rather than by mere widespread knowledge about an event. On the belief that such strong feelings are more likely to exist in and close to the community in which an event took place, the legal system now permits a change of venue for trial if it is shown that bias and prejudice at the original community are so extreme as to preclude the possibility of a fair trial at this community. Further- more, the voir dire allows attorneys and judges to deter- mine the extent to which pre-trial publicity has prejudiced potential jurors, and to eliminate potential jurors on this basis. 34 Other factors are less amenable to the juror's con- trol. The juror appears to be saddled with the most dif- ficult assignment in the courtroom. Without any special skills, he is expected "to discover not only the errors, or the perjury, of the witnesses but also the errors and fallacies of the lawyers, which often are even more puzzling" (Osborne, p. 8). The confusion of the jury, Osborne further notes, is not helped by the "vagueness and uncertainty" of much of the testimony rendered in court, arising from differences in peoples' "power of observation" and in their ability to render complete and accurate re- ports of what they observe (see, e.g., Wall, 1965; Buckout, 1974). Very often, too, witnesses mix "inference and imagination with the fragmentary facts recalled by an in- accurate memory (and) this mixing of inference with memory is one of the common sources of error in testimony" (p. 9). Another source of problems for the jury is the "belligerent attitude" that often exists between the parties to a controversy and between their attorneys and witnesses: This spirit of controversy leads witnesses to fill out and perfect the stories they tell with- out deliberately intending to commit perjury. They especially hesitate to make qualified state- ments and apparently feel that to do so would not be fair to the party for whom they testify. (Osborne, p. 10) The overall effect of the belligerent attitude is the crea- tion of "an atmosphere out of which it becomes difficult 35 if not impossible for the jury to determine just what all the facts and circumstances are" (p. 9).- The "power of sympathy" is one other influence that complicates the task of the jury. The typical jury is composed of men and women who "for all their lives (have) been taught to forgive and forget," and an experienced and clever lawyer can manipulate this sense of sympathy to his advantage. Osborne cites the case of a defandant who was acquitted by a jury against the facts: ...it was a sunshiny day and outside of the country courthouse windows the birds were singing in the trees and they helped the astute defense attorney to lead the jury to give the defendant another chance. (p. 10) Review of Empirical Literature Empirical investigation has been made of several factors that influence the jury's verdict. In a series of studies at Cornell University, the results of which were reported in Weld and Roff (1938), the researchers found that the order in which evidence was presented had an in- fluence on the jury's verdict. In one study, student3" were asked to consider themselves jurors and listened to thirteen sections of a detailed report of a criminal case. After each portion of evidence, the "jurors" indicated on a nine-point scale the degree of their belief in the guilt or innocence of the defendant. When the evidence was pre- sented in the order in which it was taken, evidence for the prosecution which came first, influenced the jurors 36 towards the "guilty" direction and evidence for the de- fense which followed shifted opinion towards innocence. More interestingly, when the order of presentation was changed so that the defense has both the first and last periods with the prosecution in between, the defendant "won" a better chance of being declared innocent. In other experiments using the same technique, Weld and Danzig (1940) found no significant difference between women and men "jurors" in their tendency to hold on to, or change their opinions. Dudycha, in appraising this last study, argued that the findings may not mean that women react in the same man- ner to all evidence as men. "The different social roles that men and women in general play in our society may in- fluence their evaluation of evidence particularly in sex cases, divorce proceedings, abandonment cases and the like" (p. 208). Other studies have also examined the possibility of sex differences in jury activities with particular attention to jury deliberation. Perhaps the earliest ex- periment in this regard was conducted in 1914 by Munsterberg (Munsterberg, 1914). Simulated juries of both sexes were given a non-legal task and asked to deliberate on it. The task was one on which opinion could be divided although there was in fact a correct solution. Members of the "juries" voted on the issue before and after a ten minute 37 deliberation, and the results of the votes were compared. Fifty-one percent of the men voted correctly on their first attempt, but after discussion, 78 percent of them gave correct answers. With the women, on the other hand, 45 percent of them voted correctly the first time, and after discussion, still 45 percent gave correct votes. Munsterberg concluded that there was a sex difference operating and that the women were ”loyal to their opinions." The conclusiveness of Munsterberg's findings has been questioned by Burtt on methodological grounds. The subjects differed "in caliber" on the sex variable: the men were mostly graduate students in an advanced seminar at Harvard, while the women were a group of undergraduates, much less rigorously selected, at Radcliffe College. Burtt argued, not without justification, that "differences of an intellectual character may have accounted for the result" and decided to replicate the experiment in a dif— ferent setting (Burtt, 1931, p. 159). With 240 undergrad- uates as "jurors," Burtt re-created the task situation (but not the task) used by Munsterberg. He found no sig- nificant difference in the effect of discussion for women compared to men. Furthermore, although the overall changes in judgment were not significant, almost twice as many women changed their judgments in the right direction as changed in the wrong direction, while the change for men were more balanced in both directions. With its larger 38 and more homogenous sample, Burtt's study deserves greater confidence than we can repose on the findings of Munsterberg. The present study is an extension in an important way of both Munsterberg's and Burtt's studies. We believe that the impact of deliberation may go beyond differences attributable to sex, and may involve factors other than decisions on the task (e.g. changes in perceptions of, and impressions about, the participants in a trial, and the trial itself). A more recent study by Strodtbeck and Mann examined the influence of sex role differentiation on the delibera- tions of lZ-person juries. Earlier research (e.g., Strodtbeck, 1951, 1956) had demonstrated a task and social- emotional specialization in family interaction and, further, had indicated that the husband or father preponderantly plays the task role and the mother or wife chiefly plays the social—emotional role. Strodtbeck and Mann (1956), therefore, speculated that there may be a "carry-over of H interaction role specializations from primary groups such as the family, to other types of groups such as the jury. To test this hypothesis, they randomly selected 12-man juries from the regular jury pools of the Chicago and St. Louis courts. The jurors listened to a recorded trial, deliberated, and returned their verdict in a simulated court atmosphere with actual bailiffs and court personnel. The deliberations were tape-recorded, transcribed and 39 scored using Bales interaction process categories. The number and forms of acts originated by each juror was com- puted and analyzed for differences between male and female jurors. The researchers found, in line with their expecta- tions, that women contributed more acts aimed at maintain- ing positive social-emotional atmosphere while the men jurors exceeded women in acts that fell in the task com- ponent. In other words, during the jury deliberations, women more than men agreed to suggestions, made jokes, and generally displayed acts directed at keeping up the spirit of the group. Men, on the other hand, gave more sugges- tions, opinions, and orientation, than women. Using the same procedure as above, Strodtbeck, James and Hawkins (1957) analyzed the deliberations of 49 mock juries for differences between the participation levels of male and female jurors in four occupational categories: proprietors, clerical, skilled, and labor. They found that men, in contrast with women, and persons of higher, in contrast with lower, status occupations participated more in the jury deliberations. In addi- tion, they were more satisfied, had greater influence, and were perceived by their peers as more competent for the jury task. Similar results were found in a study of communication groups' discussion of a mental health film by Vaughan and McGinnies (1957). Here, the members who participated more often were of higher social class, had 40 high status within the group, and were more familiar with the topic. Strodtbeck and his co-researchers concluded that "generalized role expectations" operating prior to participation may also exercise important influences on the deliberation of juries. It is important to note a methodological issue about the Strodtbeck studies. Although the primary inter- est of these studies was the interaction of jury groups, the unit of analysis was individual behavior. It is sig- nificant that, notwithstanding differences in interaction patterns between groups, Strodtbeck and his co-researchers were able to observe important regularities based on the contributions of individual members. This again suggests that attention can properly be directed at both the col- lective and individual "attributes" of groups. The collec- tivity undoubtedly exercises strong influences on the in- dividual members; nevertheless, few individuals are ever totally subsumed in a group and factors not reducible to the influence of the collectivity may also be operative during group interaction. Several researchers have shown interest in the re- search direction pioneered by Strodtbeck and his team, namely, the determination of social antecedents of jury interaction patterns. Others have extended this question to cover the influence of psychological characteristics of the accused person and of the jurors, on the outcome of 41 the jury experience. Broeder (1959, 1965a) found a ten- dency for persons with German or British backgrounds to favor the government (prosecution) while Slavs, Negroes and Italians were more likely to acquit. In a similar study based on the observations of 23 jury trials and per- sonal interviews of jurors, Broeder (1965b) concluded that "plaintiff's family and/or marital status affected the thinking of one or more jurors" in five personal injury cases which were won by the plaintiff. Broeder also established that jurors' expertise in particular occupa— tions became the basis of bias in the behaviors of these jurors during deliberation (Broeder, 1965c). I Zeisel (1968), in a study of the social character- istics of jurors in capital punishment cases, showed that race was a good predictor of opposition to the death pen- alty and that among whites, there was also a sharp dif- ferentiation by sex. Regarding the insanity defense, Simon (1967) concluded that college education, high income, or high-status occupation predisposes a juror to convic- tion. One of the most interesting studies in this regard (Bullock, 1961) examined the length of prison sentences of 3,644 inmates of a Texas state prison for possible dif- ferences based on race. Classifying length of prison sentence by type of offense by race, Bullock found that differences in the assessment of prison sentences tended 42 to persist "even though the prisoners are alike in all effective characteristics except race”. Furthermore, Negro offenders who plead guilty of murder...get shorter sentences than do whites who plead guilty of this offense, (but Negro offenders) who plead guilty of burglary get longer sentences (p. 417). Bullock concluded that "in addition to responding to the law, jurors appear also to respond to the race of the offender" (p. 417). The power of ”expert testimony" (e.g. a doctor's evidence) was investigated by Klein (1968). Twelve-person jury groups listened to one of two tapes portraying a mentally healthy man or a psychotic man. With some of the groups, the tape was supported by the opinion of a clinical psychologist (prestige suggestion) that the man was sane or insane: ...when the jury heard a tape of a "sane" man they tended to judge him as either sane or insane de- pending on the prestige suggestion they were given. Conversely, when the jury heard (an) insane man, they again tended to judge him as either sane or insane depending on the prestige suggestion given (p. 28). On the whole, Klein found what she felt may be a tendency for the jurors to vote ”sane" in contrast to "insane", analogous perhaps to a guilty/not guilty (or negligent/not negligent) verdict in trial cases. The jurors did not dif- fer significantly on sex, nor on the extent of their pre- vious jury experience, in the frequency with which they returned sane/insane verdicts. 43 Landy and Aronson (1969) examined the effect on jury verdicts of different presentations of the same victim and defendant. In one experiment, the victim of a crime of negligent automobile homicide (i.e. involving the death of the victim) was presented in a script to half of the subjects as an unattractive person, and to the other half as an attractive person. In the second experiment, the "reputation" of both the victim and the defendant was manipulated. The researchers found that subjects in the "attractive victim” conditions sentenced the defendant to a greater number of years of imprisonment than subjects in the "unattractive victim” conditions. In addition, subjects in the "unattractive defendant" condition sen- tenced the defendant more severely than subjects in either the "attractive defendant" or "neutral defendant" condi- tion. Vidmar (1972) found that the verdicts of juries reflected the range and type of alternative verdicts open to them” The juries in Vidmar's study read a description of an attempted robbery and consequent killing of a store proprietor. They were asked to return a verdict as to the defendant's guilt under one of seven conditions (including a no-decision control group) with varying number and severity of decision alternatives. Vidmar found that jurors with at least a "moderate" penalty option seldom chose a verdict of "not guilty". Conversely, more than 44 half of the jurors who were faced with only a "severe” pen- alty returned the "not guilty" verdict. In an interesting twist of research design, Simon, in two studies, used "non-normal" subjects as jurors. In Simon (1963), the jurors were three categories of mental patients: paranoids, depressives and psychopaths. As homogenous six-member juries, the subjects listened to a tape recording of a re-enacted trial of a hitherto respect- able Fire Officer charged with an incestuous sex offense involving one of his daughters. The verdict options were finding the defendant (1) guilty, (2) not guilty by reason of insanity, and (3) not guilty. Individual and group ver- dicts were obtained and the jury deliberations were also recorded. Simon found no difference in any way between the three categories of mental patients in their individual pre- deliberation verdicts, nor did a comparison of these in- dividual verdicts with the verdicts of "regular" subjects in an earlier experiment yield a significant difference in tendency to return "guilty"/”not guilty due to insanity" verdicts. However, when the group verdicts of mental and regular juries were compared, Simon found that the juries composed of mental patients were less likely to find the defendant guilty than the regular juries. Furthermore, the mental patient juries were also less able to resolve their differences and reach a unanimous decision (they had a relatively greater percentage of "hung" verdicts). 45 In an extension of this study reported a year later, Simon (1964) used prisoners as jurors. The design was similar to the earlier study; the same tape was used, in- dividual and group verdicts were taken, but the juries had two verdict options: (1) guilty, and (2) not guilty by reason of insanity. Also, two categories of prisoners were used: property offenders and sex offenders. Again, Simon found a greater tendency for the pris- oner juries to return a "hung" verdict, in contrast to regular juries. There was, however, no significant dif- ference between both the individual and group verdicts of prisoner and regular juries, nor between the verdicts of the property and sex offenders. Simon carried out an analysis of the juries' de- liberations and concluded that the lack of significant differences in verdicts between the property offenders and the sex.offenders was due to the fact that "the disagree- ment within each category muddled up the differences across categories." She further concluded that the sex offenders, on the whole, were more sympathetic and under- standing of the defendant (who was on a sex offense charge) than the property offenders. "In the main, the property offenders regarded the defendant with contempt and disgust (because) he committed acts that were personally offensive to their sense of decency and most of them.wanted to punish him by finding him guilty." (p. 341) 46 In comparison to the large number of studies of socially—determined influences on the jury, relatively few studies have focused on the psychological counterparts of such influences. Weld and Roff (1938), and Weld and Danzig (1940) reviewed at the beginning of this section, are perhaps the earliest studies that fall in this category. Mitchell and Bryne (1973), also examined the effects that the degree of attraction for a defendant expressed by a juror and the juror's level of authoritarianism have on the verdict returned by that juror. They concluded that attraction for the defendant influences the judicial de- cisions of high authoritarian jurors (in the direction of favoring the defendant), but has relatively little impact on non-authoritarian (egalitarian) jurors. Perhaps we can also classify under this category two studies into the biasing influences of pre-trial pub- licity. Sue and Smith (1974) found that damaging pre- trial publicity, even when it dealt with evidence that was irrelevant to the case, colored jurors' later evaluations of evidence presented at the trial. More notably, the bias was in favor of a guilty verdict, and it increased jurors' ratings of how convincing the prosecution's case was. Sue and Smith further found a tendency for females to vote "guilty" significantly more often than males, in reaction to the irrelevant pre-trial publicity, and to be more convinced of the validity of the prosecution's case. 47 Differences along sex in the impact of pre-trial publicity was also found by Hoiberg and Stires (1973). Pre-trial publicity that was either heinous (describing lurid details of a rape-murder), or prejudging (implying that the defendant perpetrated the rape-murder), increased the tendency of females to conclude, after the trial evidence, that the defendant was guilty of the crime. How- ever, the biasing effect of pre—trial publicity was sig- nificant only among females categorized as being of low IQ. Two points need to be stressed about these pre- trial publicity studies. First, in the Hoiberg and Stires study, the nature of the case per se (a rape-murder) may already have predisposed the females in the sample towards returning "guilty" verdicts. This is clearly because, being women themselves, they could perhaps more easily identify with (and fear) the situation involved in the crime. Second, both studies based their conclusions on the verdicts of individual jurors who never got to meet as jury groups. A more diversified empirical examination of jury deliberation has been carried out by Anapol (Anapol, 1974, 1975). Anapol (1974) used groups of students and non- students to simulate jury groups. The juries saw a video- taped presentation of a full and real civil trial; then they deliberated and their interactions were in turn video- taped and analyzed. The jurors also completed questionnaires. 48 From his analysis of the jury deliberation, Anapol drew several broad conclusions about the decision-making process in jury groups. First, he noted "a remarkable consistency in the way the juries approach the problem of deciding the case": in each instance, the jury first decided liability and then took up the problem of damages. At times, an individual juror wanted to consider a matter out of turn but the majority soon got the jury back to the general plan of liability first and then consideration of damages. (p. 6) Furthermore, each jury reconstructed for itself "the events of the case and based its decision on liability on this reconstruction": in this reconstruction phase, the jury made much use of the exhibits, examining pictures,...and drawing diagrams of the accident on a blackboard or a legal pad. (p. 7) Second, Anapol noted consistency in the manner in which the juries dealt with deviants or "holdouts". Where there was only the opinion of one individual deviating from that of the other members, "the jury tended to work on the holdout but on a one-juror—at-a-time basis until the holdout joined the majority". When two or more dis- sidents are involved, "the jury tended to break into small groups with three or four jurors working on a single hold- out." This procedure seems to be effective, possibly because each holdout feels isolated and alone in his recalcitrance. It never seemed to occur to the holdouts to band together and go to work on one or two from the other side. (p. 7) 49 Third, there was some relationship between the de- gree of conviction that a jury expressed about its decision and the amount of money it awarded as compensation. "Where the jury was strongly positive about its decision on lia- bility, it tended to award more money to the plaintiff." (p. 7) In discussing the impact of perceived credibility, Anapol noted a fourth regularity in the interaction of his juries. The juries frequently discussed the believability of the witnesses and the attorneys. They con- sidered the possible motives of the witnesses in testifying. They tended to believe the fellow ironworkers of the plaintiff who testified as eye-witnesses of the accident but considered that the ironworkers were probably all friends who would tend to help each other out...there was (also) a tendency to try the lawyers as well as the case...there did not seem to be a situation in which the case was decided solely on the try- ing of the rival lawyers, but in several in- stances the jurors did discuss the attorneys and their reactions to them. Among the items of a personal nature about the attorneys which the juries discussed were personal appearance, clothes and neatness; hair, hairstyle, and the lack of hair; facial expression, smile, and voice; language, vocabulary, mannerisms; preparation and lack of preparation of the case; personal manner, style, and politeness toward witnesses. (p. 8) With some of the juries, the level of credibility of the attorneys was manipulated before presentation of the videotaped trial. The jury members were given what was supposed to be written background information about the two attorneys. In the "high credibility" condition, this material contained such information as school attended d! 50 by the attorney whose image is boosted (e.g. Yale or Harvard), reputation of his law firm, his "vast" experience as attorney, his record of winning cases, his several publications on the subject of the trial, etc. Low pres- tige was manipulated by indicating a lack of these im- pressive qualifications by the attorney portrayed as less competent. The effects of credibility did not succeed in re- versing any of the verdicts, however, there was a signif- icant effect on the size of the award made by the jury to the plaintiff. Anapol argued that in cases in which support for the arguments of both parties is equally power- ful, "it seems reasonable to suggest that the effects of attorney credibility could affect the outcome of the case." (p. 12) In comparing the median sum awarded the plaintiff, with the amount actually awarded during the real trial, Anapol also found support for the conclusion of Miller, et a1. (1974), suggesting that juries responding to a videotaped trial presentation react in much the same way as juries deliberating from a live trial situation. In a further analysis of the same data reported a year later (Anapol, 1975), the researcher examined the relative importance given a number of factors relevant to decision-making during deliberation, for subjects exposed to a written version of the case and those who saw the 51 videotape presentation. He found that the influence of the personality of the attorneys tended to be ranked lower by those responding to the written version and concluded that the videotape version communicated more information on the personality of the attorneys and thus enhanced judgment in this regard. An equally interesting finding was the ranking of the decision-making factors by subjects in the videotape condition which revealed a tendency to give greater pro- minence to issues concerning the plaintiff. The exhibits tendered by the plaintiff, the testimony of his witnesses, the arguments and personality of his attorney received the highest ranks, in contrast to the exhibits, testimony, attorney arguments and attorney personality concerning the defendant. Furthermore, the instructions of the judge to the jury were given relatively less prominence than the summing-up statement of the attorneys. Finally, the in- fluence of the jury foreman and of the other jurors re- ceived two of the lowest ranks, a result that was also found with subjects in the written condition. Two observations concerning these findings deserve to be made. First, the nature of the civil case involved in this study (an injured steel-worker suing a manufactur- ing company for workman injury compensation) may already have predisposed the subjects towards giving greater pro- minence to the plaintiff. The tendency to think in terms 52 of protecting "the underdog" in this situation is very high. This then may explain the higher ranks given to issues relevant to the plaintiff, as against those in— volving the defendant. Secondly, there may be a tendency for jurors to ssmip that they are influenced only by "the law and the evidence". In other words, jurors may in- dicate a preference for concentrating on the task aspects of their duty (evidence, exhibits, arguments, etc.) rather than the socio-emotional aspects (influence of foreman, influence of other juror, etc.), merely because this is what is expected of them. Anapol was reluctant to accept the low ranks given the socio-emotional aspects of the jury deliberation be- cause he also found significant differences in the sums awarded by juries that deliberated, in comparison to the juries that did not. He offered two possible explanations of this phenomenon. First, it could be that jurors are not always clearly aware of exactly what influences them and their rankings are based merely on their effort to protect their self-image by admitting to doing the "right" thing. Second, since the system through which a group award is agreed upon involves a bargaining process, it is possible that jurors may not consider this compromise procedure as "influence". From a methodological perspective, whatever the explanation one accepts, these findings suggest that some caution is desirable in accepting jurors' indications 53 of the factors that influence their deliberation and deci- sions. Interest in pursuing the line of research pioneered by Strodtbeck and his team appears to have slackened since the turn of the decade. Anapol (1974, 1975) just reviewed appear to represent the new direction of interest which is in videotaping and analyzing the deliberation of simulated juries for regularities within the different patterns. Nevertheless it seems to us that further extensions of Strodtbeck's studies would be highly desirable. In par- ticular, the relationship of psychological characteristics of jury members and patterns of jury interaction should be further explored. Even more desirable will be the examina- tion of the other side of this coin, namely the influence of interaction on the perceptions of individual jurors, which we believe may hold the key to the unique features that distinguish the jury from other groups. Methodological Problems in Studying Jury Interaction From the studies on jury process reviewed above, we can note a methodological shortcoming: none of them utilized jury deliberations in actual court proceedings. This is because present statutes dictate that the delib- erations of juries be kept secret and preclude the invasion of juror privacy. In effect, many of the studies on jury deliberation have relied on post-deliberation interviews 54 which tend to have low reliability (Zimroff, 1974). A few studies, such as those of Strodtbeck et a1. earlier reviewed, make the best of the inadequate situation by simulating court proceedings as closely as possible. For example, they use actual potential jurors drawn from jury lists, and utilize real court officials during stimulus presentation. In a program of research into the effects of introducing videotape technology into courtroom trial situations, Miller and Siebert not only used potential jurors but had a real judge, a real bailiff, and profes- sional actors re-create trial proceedings from the tran- scripts of actual trials (see, e.g., Miller and Siebert, 1974; Miller, et al., 1975; Miller and Siebert, 1975). It would appear that this is probably as close to reality as one could ever get in empirical investigation of jury interaction. CHAPTER IV RATIONALE AND HYPOTHESES One of the studies that bears most directly on the issues that we are interested in was done by Harry Kalven and Hans Zeisel (reported in Kalven and Zeisel, 1966, pp. 487-491). Through post-trial interviews of jurors, the researchers reconstructed the first ballot votes (i.e. individual predeliberation verdicts) for 12- person jurors in 225 trial cases. They found that in all the cases in which the jurors were initially un- animous for conviction, the final verdict was for con- viction. Similarly, when all 12 jurors initially favored acquittal, the final verdict was for acquittal. This tendency for the final verdict to fall in line with the first ballot vote was less pronounced as initial un- animity of verdict reduced but the final verdicts still correspond closely with the initial stand of the majority. Kalven and Zeisel concluded that jury decisions are often made prior to deliberation, and argued that the function oftflmadeliberation process may not be so much to decide the case as to bring about consensus in the direction of the majority first ballot votes. 55 56 This conclusion calls for several comments. First, the final verdict of a jury may not be as "pre-determined" as Kalven and Zeisel will like to believe. Since juror assignment is randomly determined, there is about equal probability that the majority of the jurors on any one team will initially favor conviction, as that its first ballot votes will be for acquittal. This is even more true in cases that are not so "open-andishut" as to make extensive jury discussion unnecessary. Secondly, while this finding is important at the level of determining the guilt or innocence of a criminal defendant, it is of little utility in the majority of trials, which are civil and usually involve agreeing on adequate monetary compensation for the plaintiff. Finally, the use of post-trial inter- views for reconstructing the individual decisions of jurors, in many cases made months beforelfluastudy, has always raised questions concerning the validity of the measures. Commenting on the post-trial behaviors of the Panther 21 jury, Zimroff noted that some-jurors could not remember precisely what happened in the jury room. Some confused what was discussed afterwards with the press and with each other with what was said during delibera- tions. And recollections conflicted. Several, I felt, explained away evidence that they had not thought much about before in order to justify a decision already made -- perhaps on grounds other than the weakness of the evidence, or perhaps on grounds they thought could not adequately be defended (1974, p. 41). It is therefore conceivable that, at least for some of 57 the jurors, their predeliberation decisions may have been confused with the group verdicts, or clouded by an un- willingness to admit that they were influenced by the group to change their minds. Our study avoids this danger by utilizing a design that allows predeliberation decisions to be noted before deliberation. As we discuss in greater detail the rationale for looking at the impact of deliberation in terms of changes in jurors' preceptions of attorneys' credibility and re- lated factors, we shall be providing an outline of the major variables for the study and the hypotheses relevant to these variables as suggested by the trend of the dis- cussion. Major Variables Measures of the following variables were taken through the two pretest-posttest questionnaire: * individual verdict on negligence (of the de- fendant) * individual award (to plaintiff, if negligence is established) * certainty of adequacy of individual award * perceived credibility of: plaintiff's attorney defendant's attorney plaintiff defendant witness for the defendant witness for the plaintiff expert witness I (police officer) expert witness II (plaintiff's doctor) 58 * evaluation of the trial psychological measure (Dogmatism scale) relative degree of information retained from the trial + evaluation of other jurors' interest + interest in future jury duty + certainty of adequacy of group award (or verdict if no award) + degree of satisfaction with the interaction + access/no-access to videotape trial record during deliberation + degree of prominence given different trial factors demographic information (sex, age, class level) prior car accident experience + foreman/not foreman to the group during delibera- tion The variables with asterisks (*) were repeated in the posttest questionnaire. Those with the plus sign (+) are relevant only to a post-deliberation situation and consequently appeared only in the posttest questionnaire. Perceived Credibility of Trial Parties The aspect of juror perception that is of especial interest to us is the credibility of the different parties in the trial. These include the two attorneys (for the plaintiff and for the defendant), the defendant and plain- tiff, two of their witnesses (one for each side), and two 59 "expert" witnesses, one a police officer, the other the plaintiff's doctor. The importance of credibility to a legal trial can- not be overemphasized. Although there are other inter- pretations of a trial, we believe that it is first and 1 In the words of above all things a test of credibility. Miller and Boster (1975): a trial is by nature historical and retrodictive; the contesting parties scrutinize and investigate the past and eventually select samples of avail- able evidence and information which best support their respective positions...they incorporate into their cases only those items of information and evidence that bolster their client's posi- tion (p. 16). In these circumstances, the duty of the judge and jury is not only to weigh the information and evidence but also to "evaluate the veracity of the opposing evidential and informational sources". (Miller and Boster, 1975, p. 16). There is a rich literature on source credibility, dealing with the trustworthiness of a speaker in a one-to- many communication situation. We do not intend to re- view this vast literature in this study but in operationaliz- ing the perceived credibility of trial participants, we 1Miller and Boster (1975) offer two other interpretations of the trial: the image of the trial as "a rational, rule- governed event," and the image of the trial as "a conflict- resolving ritual". Miller and Boster emphasize that their three images of the trial are not mutually exclusive but each start-off point has a different implication for re- search into psychological and social questions about legal trials. 60 have utilized three dimensions of credibility suggested by Berlo, Lemert and Mertz (1969): competence (qualifica- tion), trustworthiness (safety), and dynamism.2 Our objectives are two-fold. First, we would like to see how (and if) juror perceptions of the credibility of each of the trial parties change from predeliberation to post- deliberation on any or all of the three dimensions. Second, we would like to compare the mean perceived credibility of the two attorneys by our student sample with the findings of an earlier study (Miller et al., 1975) in which non-student jurors were used. One line of reasoning suggests that a possible consequence of deliberation may be to draw attention to the credibility of the major trial parties, and of their attorneys and witnesses. This, we feel, may result in jurors exercising greater caution in assessing these trial participants. More specifically, this phenomenon may re- flect as less willingness by jurors to give high credibility scores after deliberation, in comparison to before delibera- tion. We expect this tendency to be greatest with regard to the major trial parties (plaintiff and defendant) who have the greatest to gain or to lose from the case. 2For some of the other studies that have investigated the dimensions of source credibility, see McCroskey (1966), Whitehead (1968), Baudhuin and Davis (1972), Applebaum and Anatol (1972), Wheeless (1974), Tuppen (1974), and McLaughlin (1975). 61 This tendency to be cautious may, however, work in reverse with reSpect to those witnesses whose testi- monies are respected as a result of their special training in the issues about which they testify (i.e. "expert witnesses" such as doctors and police officers). Perlmutter (1954) found that in discussion groups, members who were seen as influential were assigned desirable personality traits, presumably because the group members would like to believe they were being influenced by "worthy" persons. An off-shoot of this tendency in our case may be jurors increasing their credibility ratings of the doctor and the police witness, as a result of deliberation. We therefore suggest two hypotheses: H1: the postdeliberation credibility ratings given by jurors to the plaintiff and the defendant, and to their attorneys and non- expert witnesses, will be significantly re- duced in comparison to their predelibera- tion credibility ratings. the postdeliberation credibility ratings given by jurors to the expert witnesses will be significantly increased in com- parison to their predeliberation credi- bility ratings. Relative Prominence of Trial Factors A second line of reasoning is that changes in credibility ratings as a result of deliberation may de- pend on the amount of attention given by juries to issues relevant to the specific attorney, witness or trial party. Consequently, we are interested in the degree to which 62 the juries gave attention during deliberation to a number of issues, such as the testimonies of witnesses, sub- missions of the attorney, judge's instruction, etc. We believed that issues that are expressly raised and dis- cussed during deliberation have a greater likelihood of triggering change in jurors' perceptions of the relevant attorney or witness. Furthermore, we will note any con- sistency in the issues that are considered most primary by jurors during deliberation. Our expectations will be tested in the following hypothesis: H there is a significant positive relation- ship between the degree of prominence given by a jury to a particular issue and the degree to which the jurors, as a result of deliberation, change their credibility ratings of the plaintiff or defendant, or of the attorney or witness to which the issue is relevant. 32 Dogmatism Scale Personality differences may also account in part for changes in credibility ratings. One such personality factor is dogmatism which has been extensively explored by Rokeach. The dogmatic personality, according to Rokeach, is "excessively concerned with need for power and status" (1960, p. 69). On one dimension, the high dogmatic in- dividual tends to confuse the communicator of a message with the communication, that is, he is apt to react to a message more in response to his feelings for the 63 information source than for the substance of the informa- tion. On another dimension, the dogmatic person can be described as "rigid-conservative"; he tends to hold strongly to sets of beliefs in a manner that leaves little room for change. Dogmatism is an important concept in group de- cision making since the process of reaching concensus usually demands individuals to modify the stands taken by them prior to deliberation. We are using the 15-item version of the Trodahl and Powell short-form dogmatism scale. We expect that: H4: jurors who score higher on the dogmatism measure will more frequently show no change in their verdict after delibera- tion, in contrast to their counterparts who score lower on the dogmatism measure. H5: there is a significant, negative relation- ship between jurors' scores on the dogma- tism.measure and the degree to which they change their compensation award as a re- sult of deliberation. H6: there is a significant negative relation- ship between jurors' scores on the dog- matism scale and the degree to which, as a result of deliberation, they change their credibility ratings of the two attorneys, the plaintiff and defendants, their witnesses, and the two expert witnesses. Evaluation of Trial and Interest in Future Jury Dupy We are also interested in determining the de- gree to which our subjects consider the trial on the posi- tive or negative side along several continua. In Miller 64 et al., items similar to the ones we are using in this study were utilized as indices of juror interest and motivation. In this study we will carry out the addi- tional task of factor analyzing the items to find out whether the reactions of our subjects to the trial fall into some general dimensions. Since we are also obtaining measures of willingness to participate in future jury duty, we can hypothesize that H there is a positive relationship between the degree to which jurors perceive the trial in favorable terms and the degree to which they are willing to participate in future jury activities. 72 Information Retained and Range of Responses An index of the amount of attention jurors paid to the trial is provided by their responses to a number of questions with ”multiple-choice" answers bearing on the trial. The system by which the questions were drawn up is described in the ”variable Operationalization" section 'which follows. However, in addition to attention, we are also interested in the £3282 of responses in the predelibera- tion situation as compared to the postdeliberation situa- tion. Our main hypothesis in this regard is that H there will be closer agreement between jurors as to the correct response to the questions in the post deliberation situa- tion as compared to their responses in the predeliberation situation. 82 65 H9: the greater the degree to which jurors perceive the trial in positive terms, the greater will be the amount of in- formation they will retain about the trial. Award to the Plaintiff In reacting to the trial evidence, the jurors had the task of deciding, individually and in groups, whether the defendant in the case was negligent, and if so, what should be awarded to the plaintiff as compensation. By the instruction of the judge, the plaintiff qualifies for monetary compensation if the defendant is found negligent and no negligence on the part of the plaintiff is established as having contributed to the accident. The maximum amount that the juries were free to recommend as compensation for the plaintiff was $42,500. Kalven (1958) reported that in civil cases there was a tendency for the group damage award to approximate the average of the original sums suggested by individual jurors. He noted that this phenomenon usually occurred through the process of group dynamics rather than through conscious average effort. In this study, we are interested in finding out how the mean of the predeliberation awards by the individual jurors compare to the group award. If we adopt the same reasoning used with respect to change in credibility ratings, we would expect the group awards to be relatively more conservative than the individual 66 predeliberation awards, a demonstration of the caution that may be introduced by deliberation. Consequently, we hypothesize that H10: the mean of awards by individual jurors prior to deliberation will be signifi- cantly larger than the mean of the group awards. Since our interest is in both predeliberation and postdeliberation reactions to the trial, measures of verdicts and awards were taken in both the pretest and posttest questionnaire. In addition to the group decision, the jurors were asked in the posttest questionnaire to indicate their individual decision "regardless of what your group agreed upon". An "award shift" measure was computed based on the absolute difference between the individual's pretest and posttest recommended awards. Our expectations regarding the impact of group deliberation on shift in award is presented in a number of hypotheses later in this section. An Index of Conflict On the assumption that members of juries whose deliberations were accompanied by relatively little con- flict will be more willing to indicate as their post- deliberation individual awards figures close to those of the recommended group awards, we decided to use, as a crude index of degrees of conflict, the absolute differences between the group awards and the individual postdeliberation 67 awards. However, since conflict is dissatisfying, it seems plausible to us that the farther an individual's post- deliberation award is from the group award, the more un- comfortable he would be, and the less he would be certain that his individual postdeliberation award is an adequate decision. We thus hypothesize that: H11; the greater the difference between the ' award recommended by a jury and the post- deliberation award suggested by individual jurors, the less the certainty that will be expressed by the jurors about the adequacy of their individual decisions. Another index of conflict, we feel, may be the amount of time expended by a jury in reaching its decision. We expect that the more the agreement between members' pre- deliberation decisions, the less time will be needed by the group to reach its decision. We thus hypothesize that: H12: there is a significant, positive rela- tionship between the degree of varia- tion in the awards of individual jurors prior to deliberation and the amount of time juries require to reach an award decision. Satisfaction with Interaction and Group Award One factor that could clearly make a difference in the degree of satisfaction in group interaction ex- pressed by individual members is the extent of shift in I their original stand that they are prevailed upon to make during the interaction. We hypothesize that: 68 H13: the more the difference between group awards and individual predeliberation awards, the less the individual jurors' satisfaction with the interaction. Furthermore, we can expect members of juries whose individual decisions prior to deliberation are similar not only to spend less time reaching group de- cisions, but also to be quite satisfied with the group awards which are also likely to be close to their initial stands. It is thus plausible to assume that: H14: the more the difference between group awards and individual predeliberation awards, the less the individual jurors' satisfaction with the group awards. We also feel that given the opportunity to make an individual decision after deliberation, a juror who felt pressured into yielding ground during the delibera- tion (and is thus relatively dissatisifed) will express this dissatisfaction by recommending a postdeliberation award close to or equal to his predeliberation award. We therefore suggest that: H15: controlling for group award, the less the difference between individual pre- deliberation awards and individual post- -de1iberation award, the less the satisfaction with the interaction. Decision on Negligence of Defendant The above hypotheses dealt mainly with fluctua- tions in the amount of award (an interval variable) rec- ommended as compensation for the plaintiff in the event 69 that the defendant is found negligent. However, since no award will be called for if the negligence of the defendant is not established (or if contributory negligence on the part of the plaintiff is established), we also obtained the decisions of the subjects, individually and in groups, as to whether or not the defendant is negligent. This variable, although dichotomous, can change from pre to post- deliberation measures. Based on the reasoning which we used regarding awards, we can also suggest hypotheses re- garding individual satisfaction with the interaction and with the group decision, based on how the group decision compares to their individual decisions. We will test the hypothesis: H jurors whose predeliberation decisions on negligence differ from the decision taken by the group but who nevertheless stick to their initial stand in their postdeliberation decision will be less satisfied with the interaction than jurors who change their initial de- cision to conform to the group decision. 16' ‘Agcess to Videotspe Trial Record Half of the groups was given the option of re- .ferring to the videotape record of the trial during their deliberation to resolve any disagreements as to the evidence. We are interested in the impact that access tx) trial record of this kind may have on the degree of satisfaction experienced in taking expressed the group deCisions and consequently in the interaction as a whole. 70 We hypothesize that H17: members of the groups allowed access to the videotape trial record will express greater satisfaction with the group interaction than would the members of the groups without such access. Foreman Role and Participant's Sex Under the jury system, a foreman is chosen at the beginning of deliberation. The foreman (who may also be a woman) serves as the "coordinator” during deliberation and as spokesman for the group after deliberation. In Strodtbeck et a1. (1957), a large proportion of the juries chose as their formen the members who opened the discussion and sought either to nominate other members or to focus the group's attention on the responsibility to select foremen. The researchers also provided evidence suggesting (1) that men are more likely to be chosen as foremen than. ‘women, (2) that foremen are relatively very active in terms of participation, and (3) that jurors with high participation shift their pre-deliberation damage award the least in agreeing with the group verdict. This last point, also confirmed by Barton (1956), suggests that high participation in a group, may be indicative of a greater ability to influence the other group members, while foreman-role may provide a weapon for achieving this in- fluence through high participation. We can therefore examine the following hypotheses: 71 H18: significantly more men than women jurors will be chosen as foremen. H19: the predeliberation damage awards of the foremen will deviate from the group awards significantly less than those of the other jury members. 20: the degree of satisfaction with the inter- action indicated by the foremen will be significantly higher than that indicated by the other jury members. The findings regarding the influence of sex on a juror's tendency to change his or her verdict following deliberation are conflicting as was shown in the literature reviewed in the third chapter. On the one hand, Munsterberg (1914) maintained, based on research, that women were more "loyal to their opinions" in the sense that fewer of them changed their minds as a result of deliberation. This con- clusion is countered by Burtt (1931) and Weld and Danzig (1940) who found no statistically significant difference between men and women in their tendency to change their minds or to hold on to their opinions as a result of dis- cussion. However, where change of mind did occur in Burtt's study, women much more than men, showed marked preferences for change in a particular direction. This may suggest that the differences in impact of deliberation on women in con- trast to men may be of a more subtle nature than mere change of verdict. Since we have argued that men more than women will be chosen as foremen and that foremen will exhibit the greatest influence on the group, it seems reasonable to 72 assume that more women then men will be persuaded to lean towards the group verdict and may on the whole show greater shifts from predeliberation to post deliberation verdicts and awards. We thus expect that H21: significantly more women then men will change their negligence verdicts as a result of deliberation. H22: women jurors will change their award to the plaintiff from predeliberation to postdeliberation responses to a signifi- cantly greater degree than would the men jurors. Prior Car Accident Experience It is a truism in the social sciences that our past experiences conditions us to see the world the way we do. Whenever we are faced with situations similar to those we have encountered before, the significance of those past experiences usually influences our reactions to the current situations (see e.g. Boulding, 1956; Dember, 1960; Forgus, 1966). Thus in their behaviors during the trial in this study, jurors who have been involved in serious car mishaps (similar to the one on which the trial is based), or whose relatives to close friends have had such experiences, may be reacting in part to these prior experiences. One form in which this phenomenon may be manifested is in a tendency for jurors with past car accident experiences to be relatively less responsive to their group feelings, and to be in effect, more inflexible 73 about changing their predeliberation verdicts and awards. We therefore hypothesize that H23: jurors who have had prior car accident experiences will change their verdicts significantly less frequently than their counterparts with no such prior experi- ences. H24: jurors who have had no prior car accident experiences will change their awards as a result of deliberation, to a signifi- cantly greater degree than their counter- parts who have had such prior car acci- dent experiences. In addition to the hypotheses stated in this section, any other findings that are relevant to the gen- eral question of this study will be noted and also dis- cussed. Operationalization of the Variables Perceived Credibility of Attorneys and Witnesses Semantic differential scale items were used to assess the perceived credibility of eight participants in the trial: the two attorneys, the plaintiff, the defen- dant, a police witness, the doctor of the plaintiff (who also testified), and two "eye-witnesses" to the accident, one each for the defendant and the plaintiff. The items on each participant rated consisted of 15 bipolar adjective scales selected on the basis of their relevance to the witnesses and attorneys and such that five each of the 15 items related to the competence, 74 trustworthiness and dynamism dimensions of credibility. Each bipolar adjective scale had nine points of the structure shown below: dishonest: : : : : : : : : :honest The order in which the negative and positive end of the scales were presented to the jurors was counterbalanced to insure that the ratings were not systematically in- fluenced by the "primacy" or "recency" of either the positive or negative adjectives. The items were scored such that the extreme end of the negative side of the scale got a value of 1, and the extreme end of the positive side got a value of 9. A "total" credibility measure was computed by sum- ming the values for all 15 scales. In the same manner, measures were obtained for the three dimensions of credibility. Thus a ”total" score of 15 or a "dimension” score of 5 represented maximally unfavorable perceptions of credibility, while a total score of 135 or a dimension score of 45 reflected maximally favorable perceptions of the participant rated. The semantic differential scale ratings were supplemented by three questions requiring jurors to choose as to which of the two attorneys and the two contestants they would select as a friend, and which of the two attorneys they would prefer as an attorney for themselves. 75 Perception of the Trial A similar procedure was utilized to assess jurors' perceptions of the trial; nine-point bipolar adjective scales were used and their order of positive/negative pre- sentation was counterbalanced. However, the adjectives were changed to more closely express the perceived utility, of the trial exercise. The semantic differential scale items were supplemented by nominal-scale questions on whether the trial was fair and just with regard to each of the plaintiff and the defendant. Information Retained and Range of Responses The measure of retention consisted of 19 multiple- choice and true-false questions, and two open-ended questions, selected from 40 recall items used in Miller, _Bender et al. (1975).3 Six out of 21 recall items in the pretest questionnaire were randomly picked and repeated in the posttest questionnaire. Our intention was to com- pare the range of responses to determine if there was a difference in the degree of agreement among subjects from pre- to post-deliberation situations. 3The recall questions used in Miller et al. (1975) were selected from a large item pool that had been pretested and subjected to item analysis. The 40 items were de— termined to be the most reliable and were distributed approximately equally over the duration of the trial. In selecting our 21 recall items, we have also ensured that they related to events spread out across the trial. 76 Degree of Satisfaction and Trial Factors Nine-point scales were used to determine the de- gree of satisfaction with individual/group decisions and with the interaction, as well as the degree of importance that juries gave to the several trial factors during their ‘deliberation. In all cases, the negative ends of the scale carried smaller ”values" than the positive. Nominal scales and open-ended responses were used for all other 'questions in both the pretest and posttest questionnaire. CHAPTER V METHOD Composition of the Juries College students simulated jury teams in this study. Anapol (1974) provides evidence that student and non-student role-playing juries arrive at identical ver- dicts following deliberation ss groups. This is not counter-intuitive in any way since students do in fact qualify as jurors in their registered districts once they attain the voting age. 4 A Six-person juries were used in this study. number of studies (Kessler, 1972; Anapol, 1974) found no significant differences in the verdicts of six-man and twelve-man juries. In addition, and consistent with the findings of small-group research, the six-man juries were 4Although it was our intention to maintain the size of all groups at six, it turned out that in fact six of the juries had compositions smaller than this number. Jury member- ship was randomly determined and we took extra steps to ensure that s11 members of a jury attend the session set for them. In addition to the jury membership notice and carefully worded appeal for attendance (Appendix A), mem- bers of the juries were reminded by telephone the day pre- ceding their attendance. However, as is inevitable in a study of this kind, a few last minute drop-outs still occurred. In effect one of the juries had five members, three had four members and two had only three members. 77 78 more efficient than their twelve—man counterparts. As Anapol observed: ...the six-man jury seemed to be euqal to and often superior to the traditional twelve-man jury. The smaller jury seems more free from repetitions and wasted motion than the larger jury. It seems to work more efficiently and smoothly than the twelve-man jury...In general, all juries stay on the task problem, but the smaller juries are even better in this respect. (Anapol, 1974, p. 8) The six-man jury is now more generally used by courts in many legal jurisdictions (including Ingham), and the twelve-man jury is usually reserved for serious cases that carry heavy penalties. The subjects for the study were volunteers from several sections of two undergraduate, introductory com- munication courses offered in Michigan State University in Spring and Summer, 1976. The students signed up to participate in the study in response to face-to-face solicitation by the researcher. During this solicitation, the researcher spoke to the class as a whole and tried to "sell" the merits of the study. He pointed out that it was an investigation of "some aspects of jury activities", that it would demand about four hours from each individual in one night, but should prove an interesting experience for volunteers. In addition, the researcher arranged with the Department of Communication to give 0.25 course- credit/point as compensation for participation. 79 Fourteen students expressed interest in participat- ing in the study but could not receive the .25 course credit because of involvement in other research during the term. For these volunteers it was decided that they be paid $5 per person (which would be the approximate equi- valence of .25 university credit). However, to ensure relatively even distribution of motivation across groups, care was taken to avoid concentration of these paid mem— bers in any group. Six-member juries were formed from those who volunteered. Assignment into jury groups was randomly determined and jury members were given appointments to show up for participation as groups. All the sessions were held in the evenings. Be- cause of the problem of space, it was not possible to handle all the groups in fewer than ten sessions. All the sessions were held in the same set of rooms and care was taken to prevent any form of undesirable intrusion while the sessions were on. Procedure The juries were shown an abridged videotape re- cording of a civil trial re-enacted in full from the trans- scripts of an actual case (see Miller and Siebert, 1975, 8O Stimulus I).5 This recording was presented in three segments with five-minute breaks for coffee to minimize boredom (see Gunther, 1972; Miller et al., 1974; Anapol, 1974). Along with the evidence for both parties, the tapes also contained the opening and closing statements of the two attorneys, and the judge's charge to the jury. After the trial had been presented, members of each group retired and deliberated with the objective of reaching concensus ontfluanegligence or otherwise of the defendant. in the case, and the amount of compensation (if negligence is established) due to the plaintiff. 5We decided to use a videotaped recording of a trial in an effort to hold as constant as possible the trial upon which deliberation by the several groups will be based. The case selected was entitled Nu ent vs Clark and was a re-enactment of a trial in which a lady sued for compensa- tion for injuries allegedly suffered in a car accident. The names of the parties were changed. The roles were played by professional actors and actresses but the judge was a real judge in Flint, Michigan. The original action included a "derivative" suit by the husband of the plaintiff (Mr. Nugent) for a refund of hospital bill, automobile repair costs, etc. and the re- enactment ran for about four hours. However, since the juries in this study, in addition to viewing the trial, also had to complete two questionnaires and deliberate, we decided to minimize ”subject fatigue" by abridging the videotaped trial in a manner that would not destroy the relative merits of the evidence for both sides. Con- sequently, Mr. Nugent's derivative action was removed from the tapes leaving the evidence for the substantive action. Also, one witness was dropped from each side. Included in the witnesses who were retained was one "eye- witness" to the accident from each party. These were in- dividuals who were in the two cars along with the plain- tiff and the defendant, when the accident occurred. The abridged version of the trial ran for just under three hours. 81 To ensure that the juries took the experiment serious, the participants were told that the case was a re-enactment of an actual trial and that the study was being conducted in collaboration with the Michigan judi- ciary who were interested in how students viewed certain legal situations (Appendix B). They were of course de- briefed after the study. Two sets of questionnaires were used. The first was administered after the presentation of the trial but before deliberation. It obtained measures on "perceptual" variables already discussed. The second questionnaire was administered after deliberation and contained some of the items in the first questionnaire plus a few others on reactions to the deliberation. Some Methodological Considerations In the course of designing this study, certain methodological questions relating to the pretest-posttest format, and the appropriate unit of analysis had to be resolved. Juries need to remember as much of the evidence as they can while deliberating; in effect, it is almost imperative that deliberation takes place as soon as possi- ble after the presentation of the evidence for both sides. From the perspective of this study, the problem in obtain— ing test-retest reactions to the trial and trial 82 participants is that not much time (possibly less than one hour) may expire between the two points at which the two measures have to be taken. In these circumstances, the confounding effect of testing may be introduced into the measures thus jeopardizing their internal validity. In coping with this problem, we considered two test-retest models discussed by Campbell and Stanley (1963): the one- group pretest-posttest design (p. 7) and the separate- sample pretest-posttest design (p. 53). The one-group model simply involves testing sub- jects before and after treatment (in our case, the group deliberation). The separate-sample design model is some- what more complex and can be illustrated as follows: R O (X) R X 0 The rows represent randomly squivalent subgroups or samples, X is the treatment, and 0 stands for the measures. The first subgroup is measured prior to treatment; the equi- valent second subgroup is measured after treatment. In contrast to the one-sample pretest-posttest, this design is considered by Campbell and Stanley a good solution to the test sensitization problem since no individual is measured more than once. However, we rejected the separate- sample model for the less elaborate one for one important reason: Campbell and Stanley also point out that the separate-sample design is more appropriate with large, 83 survey-quantity populations, such as cities, in which the subgroups can be formed from an almost limitless larger group. Randomization, and the law of large numbers, are the strength of this design. However, since our interest is in examining the impact of deliberation in smsll groups, our subgroups would have to be fashioned out of six-member groups. In these circumstances, we could not rely on the power of large numbers and this design loses its advantages. In using the one-group pretest-posttest design, we took several steps to minimize the influence of testing. First, in a covering letter, we specifically pointed out that some of the items in the posttest questionnaire might look similar to items in the earlier questionnaire but appealed to the subjects to respond to the questionnaire "gs ygp fssl ss this moment." The subjects were further urged to make each response a totally independent act and to avoid being careless in completing the questionnaire. Second, not all items on the first questionnaire were re— peated. Furthermore, the order of presentation of the items that were repeated was re-arranged in the posttest questionnaire. Finally, these items were interwoven with new questions relevant to the group deliberation process, and to demographic data. Another methodological question that had to be resolved concerned the unit of analysis that should be employed in analyzing the data for this study. Since the 84 impact of juries deliberating as groups is the focus of the study, and since the pattern of interaction usually varies across groups, it seemed at first glance that averaging the simple individual responses of group members may not be adequate when several groups are involved. How- ever, on further consideration, it became clear that the group differences should not pose a problem in utilizing individual responses after deliberation since groups do, in fact, differ outside of the experimental conditions of the study. In other words, the differences in interaction patterns are integral parts of the group interaction ex- perience. Since group membership was randomly determined, from a statistical point of view, the differences could be regarded as random variation, or error, rather than bias. Nevertheless, where an issue involved characteristics of a group ss s gpspp, the mean of the individual responses will be applied as an estimate of the group "response". CHAPTER VI RESULTS This chapter is organized into two sections: (1) an examination of the effect of differences in jury group-sizes, and (2) tests of the hypotheses and report of other relevant findings. In all the tests of hypotheses, the criteria for rejecting the null hypotheses of "no- difference" will be significance beyond the 0.05 level. Impact of Jury Size Although the design of this study was for six- person juries, six of the juries (23 subjects) turned out to have less than six members. Since we were concerned with avoiding any interaction of group size with changes in perceptions and verdicts through deliberation, the question arose as to whether the analysis should be restricted to the 13 six—member juries. To resolve this question, we carried out a preliminary analysis to detect the impact of differences in jury sizes, if such an impact existed. The results of the analysis follow. 85 Table 1: Group Size Table 2: Group Size We pendence f fewer obse 86 Consistency and discrepancy between jurors' predeliberation verdicts and group verdicts for juries of different sizes Verdicts Consistent Discrepant 3 5(83%) 1 4 5(63%) 3(377.) 5 4(807.) 1 6 41(577.)- 31(43%) (n = 91) Effect of group verdicts on postdeliberation verdicts of discrepant jurors for juries of different sizes Held on to Changed discrepant discrepant verdict verdict 3 0 l 4 O 3 5 l 0 6 13(42%) 18(587.) (n = 36) could not carry out chi-square tests of inde- or these tables because many of them contain rvations than are needed for such tests. Never- theless, we can draw some important conclusions from 87 inspection of the tables. In the juries with 3, 4, and 5 members, the group verdicts were almost entirely in line with the predeliberation verdicts of the majority of members. By contrast, the degree of consistency and dis- crepancy between predeliberation verdicts and group ver- dicts for the six-member groups were relatively more balanced. Perhaps of even greater importance, the ten- dency to conform to the group verdicts was much greater wittxthe smaller-sized groups. Four out of the five‘ initial ”dissidents" in the smaller-sized groups changed their verdicts after deliberation to conform to the group verdicts. On the other hand, within the six-member groups, only 18 of the 31 dissidents (or 58%) changed their verdicts following deliberation. As a result of this analysis, we decided to restrict further analysis of the data to the six-person juries. A Word on Awards For some of the statistical tests, it was necessary to utilize the differences between the predeliberation and postdeliberation measures on the same variables. In the case of award to the plaintiff, the difference between group award and either predeliberation or postdeliberation awards was also needed. The arithmetic means of these differences, in some cases, yielded negative values (as for example when group awards were subtracted from 88 predominantly lower predeliberation awards). We were in- terested in the absolute values of these means irrespective of the direction of difference that they indicated. This did not pose any statistical problem since the deviations were squared in the statistical tests. Test Of The Hypotheses Change in credibility Table 3a: T-test of mean differences between predelibera- tion and postdeliberation credibility ratings for the plaintiff, the defendant, and their attorneys and witnesses Mean Probabilities Variable Number T- 2- 1- of Pre Post Value tailed tailed Cases * ** Mr. Simmon 78 97.29 93.29 2.68 .009 .0045 Mr. Albright 78 107.10 106.63 0.39 .695 .347 Mrs. Nugent (plaintiff) 77 75.12 72.58 0.88 .381 .190 Mr. Clark (defendant) 77 92.34 94.29 ~1.39 .168 .840 Police Officer 77 82.65 76.30 3.78 .000:* .ooo*:* Ann Nugent 77 76.71 72.53 2.67 .009 .0045 Doctor ' 78 113.41 109.24 2.11 .038* .019* Mrs. Parrish 78 85.67 88.69 -.60 .551 .225 * significant beyond the .05 level ** significant beyond the .005 level Note that one-tailed probabilities are used in the tests because our hypotheses are directional. H1: the postdeliberation credibility ratings given by jurors to the plaintiff and the defendant, and to their attorneys and non-expert witnesses, 89 will be significantly reduced in comparison to their predeliberation credibility ratings. Support for this hypothesis was found with respect to the credibility ratings for the plaintiff's attorney (Mr. Simmon), and for the plaintiff's witness (Ann Nugent) who was also her daughter (See Table 3a). There were no significant differences from pre to postdeliberation credibility ratings for the defendant's attorney (Mr. Albright), nor for his witness (Mrs. Parrish). The dif- ferences between predeliberation and postdeliberation credibility ratings for the plaintiff (Mrs. Nugent) and the defendant (Mr. Clark) themselves were not statistically significant at the .05 level. H2: the postdeliberation credibility ratings given by jurors to the expert witnesses will be significantly increased in comparison to their predeliberation credibility ratings. This hypothesis is not supported as its stands (See Table 3a). The results of the t-test suggest that the direction of change hypothesized is wrong. The dif- ferences are statistically significant but in the direc- tion opposite to the one hypothesized. With both the doctor and the police officer, the credibility ratings were reduced, rather than increased, from predelibera- tion to postdeliberation (Table 3a). Overall, the highest predeliberation credibility rating was that of the medical doctor, followed in de- scending order of magnitude by those of Mr. Albright, 90 Mr. Simmon, Mr. Clark, Mrs. Parrish, the police officer, Ann Nugent, and Mrs. Nugent. This ordinal structure did not change much after deliberation; the defendant, Mr. Clark, his attorney and his witness still maintained their credibility lead over the plaintiff, her attorney and her witness. In an effort to throw more light on these results, we performed separate t-tests on the dimensions of the credibility. These results are in Table 3b. The dimension of credibility that seemed to have been most pertinent in the evaluation of Mr. Simmon was his competence (t-test significant beyond .000 level). The dynamism of his opponent, Mr. Albright, also functioned most importantly in his evaluation at both predeliberation and postdeliberation stages but this was upset by his "poor" showing on the competence and trustworthiness dimensions. A similar situation happened with respect to both the plaintiff and the defendant. With regard to the police officer, there were significant differences on all three dimensions, but only on competence and trust for the doctor. The credibility ratings for plaintiff's witness, Ann Nugent, differed significantly on only the competence dimension. 91 Table 3b: T-tests of mean differences between predelibera- tion and postdeliberation ratings on credibility dimensions for the plaintiff, the defendant, and their attorneys and witnesses Credibility No. of Mean T— Probabilities Variable Dimension Cases Pre Post Value 2-tailed l-tailed Mr. Simmon Competence 78 31.55 28.96 3.71 .000** .000** Trust 78 34.06 33.21 1.47 .146 .730 Dynamism 78 31.68 31.24 0.65 .521 .260 Mr. Albright Competence 78 32.06 32.17 -0.16 .873 .436 Trust 78 38.17 38.01 0.31 .760 .380 Dynamism 78 37.77 36.22 3.40 .001** .000** Mrs. Nugent Competence 77 30.84 29.12 1.74 .086 .043* Trust 77 24.60 23.56 0.93 .356 .178 Dynamism 77 17.17 19.13 -1.69 .095 .047* Mr. Clark Competence 77 30.70 31.23 —.66 .511 .255 Trust 77 31.92 32.22 -.47 .639 .319 Dynamism 77 29.74 30.79 -1.67 .099 .049* Police Competence 77 29.95 27.71 3.25 .002** .001** Trust 77 26.56 24.60 1.93 .057 .028* Dynamism 77 26.00 23.84 3.07 .003** .001** Ann Nugent Competence 77 31.82 29.79 2.68 .009* .004** Trust 77 24.94 24.14 1.13 .263 .131 Dynamism 77 19.83 19.32 0.98 .330 .165 Doctor Competence 78 36.99 35.96 1.98 .051 .025* Trust 78 40.86 40.01 1.90 .061 .030* Dynamism 78 33.45 33.22 0.48 .634 .317 Mrs. Parrish Competence 78 26.71 26.12 0.99 .327 .163 Trust 78 27.14 26.13 1.51 .135 .067 Dynamism 78 31.69 31.24 0.83 .411 .205 significant beyond the .05 level ** significant beyond the .005 level Note that one-tailed probabilities are used in the tests because our hypotheses are directional. 92 Compgrison with adult sample findings Table 3c: Ratings of credibility for the contesting attorneys by adult sample and by juorrs in this study * Miller's Study This Study, competence trust dynamism competence. trust dynamism Plaintiff's attorney 27.02 26.18 25.91 24.54 26.49 24.74 Defendant's attorney 28.17 26.67 27.67 24.94 29.69 29.38 * The credibility ratings in Miller et a1. (1975) were measured on seven—point scales. Since our own credibility ratings utilized nine-point scales, it was necessary to adjust our mean ratings on the three dimensions to make them comparable with the Miller findings. Table 3c shows a comparison of our predeliberation credibility ratings with those obtained by Miller et a1. (1975) using a non-student adult sample for non-deliberated verdicts. It was not possible for us to test the differences between these two sets of findings because the standard deviations for the Miller et a1. ratings were not avail- able. However, by inspection, we can conclude that there are no dramatic differences between the two findings. Interestingly, the ratings given the defendant's attorney by the adult sample were also consistently higher than those given to the plaintiff's attorney. 93 Prominence of trial-related factors Table 4a: Jurors' assessments of the degree of prominence given trial-related factors during jury deliberation No. of * Variable Cases Mean Evidence of Mr. Clark (defendant) 77 6.74 Evidence of Mrs. Nugent (plaintiff) 78 6.56 Sketch of accident scene 71 6.16 Submission of Mr. Albright (defendant's attorney) 72 5.74 Submission of Mr. Simmons (plaintiff's attorney) 71 5.49 Judge's Instruction 71 5.01 Integrity of the attorneys 61 4.74 Evidence of Mrs. Parrish (defendant's witness) 75 4.57 Evidence of Ann Nugent (plaintiff's witness) 74 3.77 Evidence of doctor 60 3.62 Evidence of police officer 67 3.54 * Means are in descending order of magnitude. The degree of prominence was measured on a nine-point scale (maximum value = 9) 94 Table 4b: Correlation between the degree of prominence given trial-related factors and change in credibility ratings for the relevant trial participants Trial Mean assessment Difference in Participant of prominence mean credibility Mr. Simmon 5.49 4.00 Mr. Albright 5.74 0.47 Mr. Nugent 6.56 2.54 Mr. Clark 6.74 1.05 Police Officer 3.54 6.35 Ann Nugent 3.77 4.18 Medical Doctor 3.62 4.17 Mrs. Parrish 4.57 3.02 Pearson r: ~0.778. Signifncant beyond .01 level * The degree of prominence and credibility were both measured on nine-point scales with maximum values of 9. H3: there is a significant positive relationship between the degree of prominence given by a jury to a particular issue and the degree to which the jurors, as a result of deliberation, change their credibility ratings of the plaintiff or defendant, or of the attorney or witness to which the issue is relevant. This hypothesis is not supported as it stands. The direction of relationship is opposite to the one hypothesized (Table 4b). However, because the correlation is so large, and in the light of the earlier findings re- garding the pattern of change in the perception of credibility following deliberation, we feel that the 95 indicated relationship deserves to be given further con- sideration. Table 4a contains jurorS' assessments of the degree of prominence which they gave to the trial-related factors during their deliberation. The mean assessments are arranged in order of magnitude. The defendant, Mr. Clark, his attorney and his witness received consistently higher prominence than the plaintiff, Mrs. Nugent, her attorney and her witness. The sketch of the accident scene, which was displayed and used throughout the trial, seemed to have received a lot of attention, as did the instruction of the judge to the jury. The integrity of the attorneys features relatively more prominently during deliberation than the evidence of both the defendant's and plaintiff's witnesses. Perhaps most interesting, the three lowest ranks were for the testimony of Ann Nugent, the medical doctor, and the police officer, three of the four persons about whom credibility ratings changed significantly from predeliberation to postdeliberation. 96 Influence of dogmatism Table 5: Frequency of agreement between predeliberation verdicts and postdeliberation verdicts for jurors at three levels of dogmatism Pre-Post Pre-Post verdicts verdicts agree different Low 18 6 Dogmatism Medium 21 7 High 18 7 Chi-square Significance = = .07903 with 2 d.f. .9613 H4: jurors who score higher on the dogmatism measure will more frequently show no change in their verdicts after deliberation, in contrast to their counterparts who score lower on the dogmatism measure. Table 5 presents the test of this hypothesis which is not supported by the results. scores obtained by the jurors should be noted. The range of dogmatism Although a minimum score of zero and a maximum of 105 were possible, the highest score obtained was 76, and the lowest 12. mean was 43.5 and the median 41.75. The The categories were created using the percentile distribution of scores (less 97 than 31% low, 32%—67% medium, 68%-100% high). Within these constraints, there is no evidence that dogmatism influenced the tendency for jurors to hold on to their verdicts. H5: there is a significant negative relationship between jurors' scores on the dogmatism measure and the degree to which they change their compensation awards as a result of deliberation. The test of this hypothesis is contained in Table 6. The correlation of -.0002, although in the expected direction, is too low to provide support for the hypothesis. H6: there is a significant negative relationship between jurors' scores on the dogmatism scale and the degree to which, as a result of de- liberation, they change their credibility ratings of the two attorneys, the plaintiff and defendant, their witnesses, and of the two expert witnesses. The findings (Table 6) provide only slight support for this hypothesis. Six of the eight correlations are negative as hypothesized. However, only one of the correla- tions (-.2896) is significant beyond the .05 level, although three others are significant beyond the .10 level. Because of the negative direction of these cor- relations, and in the light of the curvilinear relationship indicated by Table 5, we ran scatter diagrams of these correlations in an effort to determine if the low correla- tions were due to non-linearity of relationships. The scatter diagrams did not suggest deviation from linearity as responsible for the low correlations. 98 Table 6: Table of correlations for testing specified hypotheses Correlation No. of Level of Hypothesis Variable (with) Variable Cases r Significance H5 dogmatism change pre to post awards 77 —.0002 .499 H6 dogmatism chance credi- bility, Simmons 77 -.1605 .082 dogmatism change credi- bility, Albright 77 -.1792 .059 dogmatism change credi- bility, Nugent 77 -.l723 .067 dogmatism change credi- bility, Clark 77 .1324 .126 dogmatism change credi- ** bility, Ann Nugent 77 -.2896 .005 dogmatism change credi- bility, Police 77 .1879 .051 dogmatism change credi- bility, Doctor 77 -.0126 .457 dogmatism change credi- bility, Parrish 77 -.0892 .220 H7 trial eval- interest in * uation furture jury 78 .2384 .018 H9 trial eval- recall nation 78 .0332 .387 H11 difference certainty group to post award 60 .1201 .180 award H13 satisfaction difference with inter- pre to group * action award 77 .2327 .021 H14 satisfaction difference with group pre to group ** award award 77 .3939 .001 * significant at the .05 level ** significant at the .005 level 99 Evaluation of the trial H7: there is a positive relationship between the degree to which jurors perceive the trial in favorable terms and the degree to which they are willing to participate in future jury activities. The correlation in Table 6 shows support for this hypothesis. The correlation between "evaluation of trial" and "interest in future jury duty" is .2384 which is significant at the .018 level. Since we obtained trial evaluations both before and after deliberation, we were interested further in how these two sets of evaluations compared with one another. We carried out two additional sets of analysis: a t-test of predeliberation and post- deliberation evaluations, and factor analyses of the pre- deliberation evaluations and the postdeliberation evalua- tions. The mean evaluation of the trial by jurors increased significantly from predeliberation to postdeliberation for four factors, dull-exciting, energetic-tired, fatiguing- refreshing, and stimulatingrtedious (Table 7). The mean evaluation decreased beyond chance from predeliberation to postdeliberation for four other factors: fair-unfair, valuable-worthless, subjective-objective, and confusing- clear. The improvement in evaluation occurred with those factors that could conveniently be labelled as socioé emotional dimensions, crucial to the well-being of the group but not directly relevant to the legal task of the Table 7: 100 T-test of mean differences between predelibera- tion and postdeliberation evaluation of the trial for sub-variables that showed significant differences No. of Mgag T- Probabilities Sub-variables Cases Pre Post Value 2-tailedi l-tailed Dull-Exciting 77 4.01 4.36 -2.15 .035* .017* Energetic-Tired 77 4.06 4.51 -2.84 .006* .003** Fatiguing- ** ** Refreshing 77 3.17 3.81 -3.51 .001 .000 Stimulating- * * Tedious 77 3.55 3.97 -2.18 .033 .016 ** ** Fair-Unfair 78 6.91 6.12 3.25 .002 .001 Valuable- * * Worthless 78 6.29 6.00 2.07 .042 .021 Subjective- Objective 78 5.18 4.42 3.54 .001** .000** Confusing- * * Clear 78 5.49 4.77 3.28 .002 * .000 * significant at the .05 level ** significant at the .005 level jury. with the more task—relevant factors. Conversely, the worsening of evaluation occurred 101 Table 8a: Rotated factor matrix for predeliberation trial evaluation Factor 1 ' Factor 2 Factor 3 interesting-boring .82743* .30546* .24349 dull-exciting .75908* .20107 .48246* energetic-tired .79801* .11136 .49244* active-passive .47710* .18786 .38519 bad-good .50313* .66478* .21402 fair-unfair .14241 .72758* .12375 valuable-worthless .27405 .73012* .20674 wrong-right .03424 .75046* .11760 subjective-objective .06230 .01116 .19267 attention easy-difficult .50048* .26439 .59036* fatiguing-refreshing .46117* .21114 .68346* confusing-clear .15799 .32831* .52035* stimulating-tedious .43748* .23344 .64780* *factor loadings above .300 Factor 1: eigenvalue = 6.319; Z of variance = 78.7 Factor 2: eigenvalue = 1.311; Z of variance = 16.3 Factor 3: eigenvalue = 0.397; Z of variance = 4.9 Table 8b: Rotated factor matrix for postdeliberation trial evaluation Factor 1 Factor 2 Factor 3 interesting-boring .70805* .44515* .09163 dull-exciting .78663* .39839* .09432 energetic—tired .82457* .28448 .08350 active-passive .61588* .26908 .24135 bad-good .40316* .75659* .19910 fair-unfair .20183 .17171 .50882* valuable-worthless .29081 .75701* .28609 wrong-right .24268 .17768 .81816* subjective-objective .03796 .03914 .40625* attention easy-difficult .73766* .21174 .27739 fatiguing-refreshing .70817* .03477 .23074 confusing-clear .39268* .17907 .24395 stimulating-tedious .73726* .24768 .23377 *factor loadings above .300 Factor 1: eigenvalue = 6.113; Z of variance = 79.8 Factor 2: eigenvalue = 0.911; Z of variance = 11.9 Factor 3: eigenvalue = 0.634; Z of variance = 8.3 102 The factor analysis confirms the presence of at least two major dimensions corresponding to the socio- emotional and the task dimensions. In the factor matrix for predeliberation evaluation (Table 8a), the socio- emotional dimensions dominate both the first and third factors while the task dimensions had the highest loadings on the second factor. In the postdeliberation matrix, the first factor, the largest, is still interpretable as socio- emotional, and Factor 3 as task-oriented. The middle factor had high loadings for variables in both dimensions and may be a mixed factor. The socio-emotional dimensions seem to dominate both factor matrices, loading very highly on the largest factors in both cases. This is interesting in the light oftfluatendency to regard jury interaction as a task- oriented activity. Trial information retained Table 9: Central tendencies of responses by jurors to four randomly selected information-retention items before and after deliberation Predeliberation Postdeliberation Median Mode Median Mode Response to Q.l 2.99 3.00 3.01 3.00 Response to Q.2 2.94 3.00 2.94 3.00 Response to Q.3 3.90 4.00 3.93 4.00 Response to 0.4 3.01 3.00 3.01 3.00 H8: there will be closer agreement between jurors as to the correct responses to information- retention questions in the postdeliberation situation as compared to their responses in the predeliberation situation. 103 Four of the twenty recall items randomly selected and repeated in the postdeliberation questionnaire are compared to test this hypothesis. From inspection of the above table, it is clear that jurors' responses to the four items did not change in any significant way from predelibera— tion to postdeliberation. Consequently, we did not carry out any further statistical test on this part of the data. This hypothesis is not supported by the findings. H9: the greater the degree to which jurors perceive the trial in pOSitive terms, the greater W111 be the amount of information they will retain about the trial. The test of this hypothesis is reported in Table 6. The correlation of .0332 between trial evaluation and recall is not significant at the .05 level and does not provide support for the hypothesis. Group Awards H10: the mean of awards by individual jurors prior to deliberation will be significantly larger than the mean of the group awards. Because the test of this hypothesis involves com- parison within groups as well as across groups, it was necessary to compute manually the means of individual awards within the groups. These means, which were then compared with the group awards, are presented in Table 10 below. 104 Table 10a: Comparison of mean predeliberation awards within juries and mean of group awards Mean of individual awards by groups Group awards $15,834 $30,000 9,584 00,001* 20,417 22,500 11,000** Hung** 3,334 00,000*** 10,834 00,000 14,500 00,000 8,767 00,000 11,251 00,001 16,667 00,001 8,834 12,500 18,751 30,000 6,667 00,001 mean: 12,120 mean: 7,917 s.d.: 5,141 s.d.: 12,469 critical t = 1.717 (one-tailed) with 22 degrees of freedom obtained t = 1.036. Not significant at the .05 level. * groups that found the defendant not negligent were coded as awarding $1 (00,0001) to the plaintiff. ** the mean individual awards for this jury was dr0pped in the analysis. *** groups that found both the defendant and the plaintiff negligent were coded as awarding $0 (00,000) to the plaintiff. As indicated by Table 10a, the mean of awards by individual jurors within groups is larger than the mean of all the group awards but not enough to be significant at the .05 level. The wider variance in the group awards is prob- ably responsible for the failure of the differences to reach 105 significance. The juries showed a tendency to be either fully for the plaintiff (and return a high award deciSion), or fully against her. It is important to note from Table 10a that in six of the seven juries (not including the hung jury) whose individual award means were less than the overall mean for individual awards (i.e. less than $12,120), the group awards were also less than the mean of group awards ($7,917). Of the remaining five groups whose individual award means were higher than the overall mean for individual awards, three also had group awards that were higher than the mean of group awards. Table 10b: Mean, median and mode for individual predelibera- tion awards and group awards Individual awards Group awards Median: $00,001 Median: $00,001 Mode: 00,001 Mode: 00,000 Mean: 12,120 °°'°°1 Mean: 7,917 In an attempt to further understand the relationship of individual predeliberation awards to group awards, we computed the median and mode of the individual awards by the jurors in the 12 groups that returned group verdicts. A comparison of these figures with the median and mode of the group awards is presented in Tables 10b and 10c. The median 106 Table 10c: Comparison of median and mean of individual awards within groups with group awards Median of Mean of individual awards individual awards within groups Group awards within groups $15,000 $30,000 $15,834 00,000 00,001 9,584 25,000 22,500 20,417 00,001 00,000 3,334 10,000 00,000 10,834 00,001 00,000 14,500 00,500 00,000 8,767 00,001 00,001 11,251 10,000 00,001 16,667 10,000 12,500 8,834 15,000 30,000 18,751 00,001 00,001 6,667 average: 7,088 mean: 7,917 mean: 12,120 and the mode of the predeliberation awards are clearly better predictors of the group awards than the simple means of the individual awards. H11: the greater the difference between the award recommended by a jury and the postdeliberation award suggested by individual jurors, the less the certainty that will be expressed by the jurors about the adequacy of their decisions. The correlational test of this hypothesis is in Table 6. With a value of .1201, the correlation is in the expected direction but is not significant at the .05 level, and does not provide strong enough evidence in support of the hypothesis. 107 Table 11: Relationship between variation in individual predeliberation awards and duration of de- liberation Spread of awards Duration of (standard deviation) deliberation (in minutes) 17440 25 17250 42 17350 15 8165 50 12006 27 22483 15 17001 25 18285 22 11254 38 8346 8 10327 23 20961 9 Pearson r: -0.389. Not significant at the .05 level. H12: there is a significant positive relationship between the degree of variation inithe awards of individual jurors prior to deliberation and the amount of time juries require to reach an award decision. The correlation of spread of awards with duration of deliberation is -0.389 (Table 11). The relationship is not statistically significant at the .05 level (n = 12; critical r .576). However, the negative direction of relationship indicated is intriguing because it is counter- intuitive and will be discussed further in the next section. “In—- a" .... 108 Satisfaction with Interaction H13: the more the difference between group awards and individual predeliberation awards, the less the individual jurors' satisfaction with the interaction. The data supports this hypothesis. The correlation of .2327 is both in the predicted direction and significant beyond the .05 level (Table 6). H14: the more the difference between group awards and individual predeliberation awards, the less the individual jurors' satisfaction with the group awards. This hypothesis is also supported by the data. The correlation of .3939 is significant at the .001 level. H15: controlling for group award, the less the dif- ference between individual predeliberation awards and individual postdeliberation award, the less the satisfaction with the interaction. The zero-order correlational test of this hypo- thesis shows a correlation of .1803 between satisfaction with the interaction and the extent to which the pre- deliberation and postdeliberation awards are different. When group awards are controlled for, the correlation increases to .2156, which is significant at the .036 level. The data supports the hypothesis. 109 Decision on Negligence of the Defendant Table 12: Chi-square test of satisfaction with the inter- action for jurors who conformed to the decisions of the juries and those who did not Satisfaction Low High Did not conform 7 6 Conformed 5 13 Chi-square = 1.203 with 1 degree of freedom. Significance = .2727 (n = 31) H16: jurors whose predeliberation decisions on negligence differ from the decisions taken by the groups but who nevertheless stick to their initial stands in their postdelibera- tion decisions will be less satisfied with the interaction than jurors who change their initial decisions to conform to the group decisions. Although the chi-square value was not significant at the .05 level it is important to note that of the jurors who changed their verdicts to conform to the group verdicts, almost three times as many were in the high satisfaction category as were in the low satisfaction group (Table 12). 110 Access to Trial Record Table 13: T-tests of mean differences in satisfaction be- tween juries allowed access to the videotape trial record and juries without such access No. of T- Probabilities Variable Cases Mean Value 2-tailed l-tailed Satisfaction Had access 41 6.781 with group -2.16 .034* .017* interaction No access - 36 7.778 Satisfaction Access 41 5.488 with group -2.97 .004** .002** verdict No access 36 7.222 Confidence in Access 29 5.862 postdelib- -0.35 .728 .364 eration de— cision (individual No access 31 6.097 * significant at the .05 level ** significant at the .005 level H17: members of the groups allowed access to the videotape trial record will express greater satisfaction with the group interaction than would the members of the groups without such access. Members of juries allowed access to the trial re- cords and those of juries without such access were signif- icantly different in their satisfaction with group inter- action and with the group verdicts (Table 13). However, the direction of differences is opposite to that hypo- thesized. There is no significant differences between these two groups on the degree of confidence expressed in their postdeliberation decisions. 111 Foreman role and sex Table 14: Frequency of men and women chosen as jury foremen Chosen as Not chosen foremen as formen Female 5 e 35 Sex (134) Male 25 7(222) Chi-square = 0.5513 with l d.f. Significance = .4578 Gamma = -0.3243 H18: significantly more men than women will be chosen as foremen. The chi-square test does not provide support for this hypothesis. The value of gamma is moderately strong indicating that the expected direction of relationship is correct (Table 14). However, the degree of the relation- ship is not strong enough to be significant at the .05 level. Table 15: T-tests of mean differences in awards influence and satisfaction with interaction between jury foremen and regular jury members No. of T- Probabilities Variable Group Cases Mean value 2-tailed l-tailed Difference, Foremen 12 1875.1 pre-award 0.63 .540 .270 and group Members 60 ~43l4.7 award Satisfaction Foremen 11 7.91 with the 1.19 .253 .176 interaction Members 60 7.10 112 the predeliberation damage awards of the fore- 19. men will deviate from.the group awards signif- icantly less than those of the other jury members. H20: the degree of satisfaction with the interaction indicated by the foremen will be significantly higher than that indicated by the other jury members. The tests of these hypotheses are presented in Table 15. Although neither of the t-values was significant at the .05 level, it is important to note that the mean award by the foremen is closer to the mean group award than is the mean award recommended by the regular jury members. Also, the mean satisfaction in the interaction indicated by the jury foremen is slightly higher than that indicated by the regular members. However, the five to one ratio of regular members to foremen should be kept in mind in interpreting this part of the data. Table 16a: Change in negligence verdicts for male and female jurors Pre-Post Pre-Post verdicts verdicts consistent different Female 33 10(24%) Male 24 10(29%) Chi-square = 0.123 with l d.f. Significance = .7263 113 Table 16b: Predeliberation verdicts for male and female jurors Defendant Defendant negligent not negligent Female 29(66Z) 15 Chi-square = 7.353 with l d.f. Significance = .0067 Gamma - .6034 significantly more women than men will change their negligence verdicts as a result of deliberation H21: The analysis yields no support for this hypothesis (Table 16a). Further analysis also did not show significant differences between male and female jurors in the frequency with which their postdeliberation verdicts were influenced by the group verdicts. However, the male and female jurors differed significantly in their predeliberation verdicts (see Table 16b). About 70Z of the females found the de- fendant negligent while the same proportion of male jurors found the defendant not negligent (chi-square test signif- icant at the .0067 level). This is an interesting finding if we consider the-fact that the defendant in this trial is male and the plaintiff female. 114 Table 17: T-tests of effect of sex and prior car accident experience on awards of jurors No. T— Probabilities Variable Group Cases Mean value 2—tailed l-tailed Difference, Female 44 3795.5 pre-award & _ 1.42 .159 .079 post award Male 34 —602.5 Difference, Female 44 3172.8 pre-award & 2.09 .042* .021* group award Male 34 —1l411.3 Difference, Accident 34 5735.4 pre—award & 2.02 .047* .024* post-award No accident 43 —848.5 Difference, Accident 34 6529.5 pre-award & 2.77 .007* .004** group award No accident 43 —10939.1 * significant at the .05 level ** significant at the .005 level 22‘ women jurors will change their awards to the plaintiff from predeliberation to postdelibera- tion responses to a significantly greater degree than would the men jurors. This hypothesis is not supported by the analysis (Table 17). However, there is significant difference on sex in the degree of similarity between predeliberation awards and group awards.‘ The predeliberation awards of the female jurors are relatively closer to the group awards, suggesting, rather interestingly, that the female jurors influenced the group verdicts to a greater degree than did their male counterparts. 115 Prior car accident experiences Table 18: Change of verdicts for jurors with prior car accident experiences and jurors without such experiences Pre-Post Pre-Post verdicts verdicts consistent different Prior accident 25 9 experience No accident experience 32 10 Chi-square = 0 with l d.f. Significance = 1.00 H23: jurors who have had prior car accident exper- ience will change their verdicts significantly less frequently than their counterparts with no such prior experiences. The chi-square test of independence (Table 18) provides no support for this hypothesis. There is absolutely no difference between jurors who have had car accident experiences in the past and those who had not in their tendency to be inflexible in their verdicts. Thirty-two of the 34 jurors who have been involved in automobile mishap also indicated whether they, or the other party, was at fault. Twenty-four said the other driver was at fault and eight indicated that they were at fault. Of those who said the other driver was at fault, 63Z (15 of 24) also found the defendant in this trial negligent. Conversely, jurors who admitted that they were to blame for the prior accidents, almost two to one, found 116 the defendant in this trial not negligent. It seems that these jurors tended to put themselves in the position of the plaintiff in this trial and to return verdicts con- sistent with their own experiences. H24: jurors who have had no prior car accident ex- periences will change their awards as a result of deliberation, to a significantly greater degree than their counterparts who have had such prior car accident experiences. The test of this hypothesis is reported in Table 17. The two groups differ significantly in the degree to which they changed their predeliberation awards as a re- sultcfifdeliberation (t = 2.02, significant at .024 level). However, contrary to our expectation, jurors with past accident experiences changed their verdicts to a greater degree than those who had none. The direction of the change, nevertheless, shows that the jurors with past accident experiences were more stringent in their post- deliberation awards (hence the positive value), while the jurors without these experiences were more liberal in their postdeliberation awards. Jurors with past accident experiences seem to have had more influence on the group verdicts than their "no- accident" counterpart, and the two groups differed signif- icantly in the degree to which their predeliberation awards compared with the group awards (t = 2.77, significant at .004 level). CHAPTER VII CONCLUSIONS AND RECOMMENDATIONS Summary_of the Findings Change in credibility ratings after deliberation was found with respect to the plaintiff's attorney, the plaintiff's witness, the medical doctor, and the police officer. In each case, the change was of a reduction, rather than an increase, in credibility. On the-whole, the medical doctor had the highest credibility ratings before deliberation. He was followed by the attorneys for the defendant and the plaintiff in that order, the defendant and his prime witness, the police officer, and finally, the plaintiff and her prime witness. This ordinal struc- ture did not change much after deliberation, with the defendant, his attorney and his witness still maintaining their credibility lead over the plaintiff, her attorney and major witness. A comparison of our findings with those of an adult sample in Miller et a1. (1975) revealed a remark- able similarity in the ratings of attorneys' credibility in the two studies. Also, in both studies, the defendant's attorney consistently received higher credibility ratings than his plaintiff's counterpart. 117 118 Contrary to our expectation, a strong negative re- lationship was noted between the degree of prominence given to trial-related factors during jury deliberation and the change in credibility ratings for the individuals to whom the issues were relevant (r = -0.778, significant beyond .01 level). Again, issues relevant to the defendant and his team were accorded greater prominence than those relevant to the plaintiff and her team. The medical doctor, the police officer, and the plaintiff's principal witness, three persons about whom credibility ratings changed significantly after deliberation, received the three lowest prominence ranks. There was no significant evidence that dogmatism influenced the tendency for jurors to hold on to their negligence verdicts, nor to their compensation awards, in spite of deliberation. However, there was a slight tendency for high dogmatic jurors to change to a lesser degree than the low dogmatic jurors their credibility ratings of the attorneys, the plaintiff, plaintiff's witness, the medical doctor and the defendant's witness. A comparison of jurors' evaluation of the trial before and after deliberation showed that their feelings about the trial improved with deliberation on socio- emotional, dynamism dimensions such as the degree to which the trial was dull-exciting, energetic-tired, fatiguing- refreshing, or stimulating-tedious. Conversely, jurors' 119 evaluation worsened on task-relevant factors such as the degree to which the trial was fair-unfair, subjective- objective, confusing-clear, or valuable-worthless. Factor analyses of the predeliberation and postdeliberation evaluations showed that the socio-emotional variables had their highest loadings on the largest factors (78.7Z and 79.8Z of the variance respectively). This is an intriguing finding in the light of the general tendency to regard jury interaction as primarily a task-oriented activity. Two hypotheses were tested with regard to the re- tention of trial-related information. No siginficant shift was found in the agreement of jurors as to the correct responses to information-retention questions, after de- liberation in comparison to before the exchange. There was also no evidence that jurors' feelings about the trial (i.e. whether or not they perceived the trial in positive terms) significantly affected the amount of information that they retained about the trial. A comparison of the group awards with the means of individual awards within groups showed that the overall mean of individual awards, as hypothesized, was larger than the overall mean of group awards -- but not enough to be significant at the .05 level. The group awards were almost all of extreme values ($0, $1, or $30,000), and we suggested that this wide variance in the group awards may be re- sponsible for the failure of the difference to reach 120 significance. Nevertheless, the group awards tended to reflect the trend of predeliberation awards dominant within the groups. There seemed to be a tendency for jurors to be more satisfied with their postdeliberation awards the less these awards differed from the group awards. This relation- ship was however not significant at the .05 level (r = .1201, significance = .180). Contrary to expectation, a negative relationship was observed between the spread in predelibera- tion awards within groups and the amount of time jurors required to reach consensus. The correlation of -0.389 seemed moderately high but was not statistically significant because of the small number of observations involved (n = 12). There was evidence that jurors' satisfaction with the interaction was affected by the degree of discrepancy between their predeliberation awards and the awards rec- ommended by their groups. The more the jury awards differed from the predeliberation awards of the individual jurors, the less satisfied were the jurors with the interaction as a whole (r = .2327, significance = .021). Similarly, the jurors were satisfied with the jury awards the more these group awards were similar to their predeliberation awards. Satisfaction with the interaction was also found to be related positively to the difference between individual predeliberation awards and individual postdeliberation awards. This relationship increased when group awards are 121 controlled for (r = .2156, significance = .036). Finally, although the relationship between conformity to group ver- dicts and satisfaction with the interaction was not strong enough to be statistically significant, jurors who con- formed were about three times as many in the high satis- faction category as were in the low satisfaction group. Contrary to our expectation, members of juries allowed access to the trial record were significantly less satisfied with the group interaction and with the group verdicts then their counterparts without such access. The two groups did not differ in the degree of confidence which they expressed in their postdeliberation decisions. There was no significant difference in the propor- tion of men who were chosen as jury foremen, in comparison to women jurors (chi-square = .5513, significance = .4578). However, the value of gamma (-0.3243) was fairly strong indicating that the hypothesis that men tend to be favored as formen is in the right direction. Jury foremen and regular members did not differ significantly in the degree to which their predeliberation awards influenced the group awards, nor in the degree of their satisfaction with the interaction. The mean pre- deliberation awards of the foremen was however closer to the mean of the group awards than was the predeliberation awards suggested by the regular jury members. The 122 foremen and the regular members also did not differ signif- icantly in the degree of their satisfaction with the inter- action, although again the satisfaction of the foremen was slightly higher than that of the regular jury members. There was no support for the hypothesis that female jurors would change their verdicts following deliberation to a greater degree than Would the male jurors (chi-square = .123, significance = .7263). However, the male and female jurors differed significantly in their predeliberation verdicts (chi-square = 7.353, significance = .0067; gamma = .6034). The female jurors, who were of the same sex as the plaintiff, heavily favored a 'negligent' verdict against the defendant. By contrast, the male jurors, who were of the same sex as the defendant, more than two to one found the defendant not negligent. The predeliberation awards of the female jurors were significantly closer to the average group award than were those of the male jurors suggesting, interestingly, that the female members were relatively more influential with the groups (t = 2.09, significance = .021). No support was found for the hypothesis that the female jurors would change their predeliberation awards to a greater degree than their male counterparts. Jurors who have had accident experiences in the past did not differ significantly from jurors without such experiences in their tendency to be inflexible about their 123 verdicts on negligence. However, the two groups differed significantly in the degree to which they changed their predeliberation awards as a result of deliberation (t = 2.02, significance = .024). Contrary to our hypo- thesis, jurors with past accident experiences changed their awards to a greater degree than those who had none. Nevertheless, the jurors with automobile accident history were more stringent in their postdeliberation awards than their 'no-accident' counterparts. The two groups also differed significantly in the degree of their award in- fluence onthe groups, with the 'past-accident jurors' appearing to have had greater success in this regard (t = 2.77, significance = .004). Discussion of the Findings With respect to the changes in credibility ratings, it seems to us that there was a tendency for the jurors to be more stringent in their assessment of the credibility of those involved with the plaintiff than those involved with the defendant. Not only was the competence of the plaintiff's attorney and her witness called more into question, but the mean credibility ratings given to the defendant, his attorney and his witness were consistently higher than those accorded the plaintiff, her attorney and her witness bgth before and after deliberation. A comparison of our findings with those of an adult sample 124 (Miller et a1. 1975) showed a similar trend; the defendant's attorney received higher credibility ratings than his plaintiff's counterpart. The fact that in all cases the changes were of a reduction in credibility ratings after deliberation appears to lend support to ourargument that one impact of delibera- tion may be to make jurors more cautious and in effect less liberal in their credibility evaluations. It seems that .this tendency operates more strongly in jurors' evaluations of the plaintiff and of persons involved with the plaintiff. It is important to note that both the doctor and the police officer, although expert witnesses, gave testimonies that were favorable to the plaintiff -- and both of them had their ratings on competence and trustworthiness reduced after deliberation! The findings, furthermore, support earlier findings (e.g. Klein, 1968) that medical doctors are perceived as experts and that their testimony may be respected partly as a result of their technical training. The medical doctor was given the highest credibility ratings both before and after deliberation. Contrary to our expectation, the police officer did not feature prominently in this regard, but the two attorneys did. Mr. Simmon and Mr. Albright were given the second and third highest predeliberation credibility ratings, and the second and fourth highest postdeliberation credibility ratings. 125 The findings regarding the relative prominence given to trial-related issues during deliberation seemed to shed some light on the earlier findings. Again, the de- fendant's team showed an edge over the plaintiff's; rel- atively more attention was given to factors involving the defendant than to those involving the plaintiff! However, in an absolute sense, and contrary to our earlier argument, the more the attention accorded by jurors to specific factors during deliberation, the less they changed their credibility ratings from predeliberation to postdeliberation. This suggests to us; that the impact of deliberation on jurors' preceptions of credibility may not be as simple and straightforward as we have earlier assumed. It seems that with respect to issues that were more prominently discussed, deliberation served to reinforce the opinions of the jurors held before deliberation. Conversely, with those issues that featured less prominently during deliberation, such reinforcement was absent, the jurors were uncertain about their earlier positions, and coped with this situation by being less liberal in their postdeliberation credibility ratings of the persons to whom the issues were relevant. The results with respect to jurors' evaluation of the trial are among the most intriguing findings of this study. After hours of listenipg to court proceedings, the opportunity to respond provided by deliberation appeared to raise the spirit of the jurors. At the same time, 126 however, the exchange seemed to have brought home to the jurors their fallibility as individuals thereby making them more cautious in their evaluations of the task-relevant dimensions after deliberation. These findings agree with those regarding changes in the credibility assessments of the persons involved in the trial. It is important that where significant changes occurred in credibility assessments (a highly task-relevant factor) following deliberation, the direction of the change was also negative (i.e. a decrease, rather than an increase). The short duration of the trial used in this study might have been responsible for the failure to find significant shifts in the agreement of jurors as to the correct responses to information-retention items after deliberation in comparison to before it. We believe that the results would have been different with a longer trial, or alternatively, if the jurors were presented with more information-retention items both before and after delibera- tion. This possibility should be considered in further research. We were interested in testing Kalven and Zeisel's assertion that the verdicts of juries closely reflect the dominant opinion of members prior to deliberation. For two of our juries, the predeliberation verdicts were equally divided. One jury could not reach a verdict after delibera- tion but of the remaining ten juries, seven returned group verdicts that were in line with the verdict of the majority 127 of members prior to deliberation. This provides some evidence in support of Kalven and Zeisel's conclusions. It may also explain why jurors tend to hold on to their preceptions regarding issues given much prominence during deliberation, since these discussions are also likely to be fashioned around the issues considered most important by the majority of members. Both the individual predeliberation awards and the group awards varied widely making it difficult to use the mean as the best descriptor of the central tendencies of these awards. In effect, the median and the mode of in- dividual awards prior to deliberation proved to be much better predictors