Predicting juror bias and its effect on group deliberation, verdict, and recall of trial information


Material Information

Predicting juror bias and its effect on group deliberation, verdict, and recall of trial information
Physical Description:
ix, 215 leaves : ; 29 cm.
Brand, Arthur Howard, 1958-
Publication Date:


Subjects / Keywords:
Jurisprudence   ( mesh )
Prejudice   ( mesh )
Clinical and Health Psychology thesis Ph.D   ( mesh )
Dissertations, Academic -- Clinical and Health Psychology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1985.
Bibliography: leaves 201-214.
Statement of Responsibility:
by Arthur Howard Brand.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000541759
oclc - 16996737
notis - ACW5304
System ID:

This item is only available as the following downloads:

Full Text







Copyright 1985


Arthur Howard Brand

Dedicated lovingly to

my parents, Robert and Helen,

and my wife, Ana.


Research ain't pretty. Indeed, after countless days and nights

of reading, planning, running subjects, analyzing data, and writing

results as well as personally manifesting the entire range of symptoms

associated with major forms of psychopathology, it can become quite

ugly. Despite the trauma associated with this dissertational

experience (I do not know of any such experiences that could have been

described as otherwise) I feel a great sense of accomplishment

regarding this final (hopefully) graduate student project. I would

like to take this opportunity to express my gratitude to the following

people and institutions, whose cooperation, assistance, and support

greatly facilitated its completion:

Jeff Ackerman; Gerry Bennet, J.D.; Donna Bishop, Ph.D.; Brian

Boruff; Barbara Burkett, J.D.; Randy Carter, Ph.D.; Joni Congdon;

Alexis Durham, Ph.D.; Louis Fogg, Ph.D.; Anne Freund; Kenneth "Z"

Freundlich, Ph.D.; Leslie Fried; Larry Gage; Jacquelin Medenblik;

Michael "Agent X" Unger, Ph.D.; Mark Yankauer; Flora Zaken-

Greenberg; the Spring 1984 Trial Law, Introduction to Criminal

Justice, and Personal Growth students; The University of Florida

Learning Resources Center; The Shands Teaching Hospital Audio-

Visual Department; and The University of Illinois-Chicago Circle

Computer Center.

My fondest thanks and gratitude are extended to my chairperson

and supervisor, Jacquelin Goldman, Ph.D., of the University of Florida

Department of Clinical Psychology, for her "judicious" advice and

consultation, both in and out of the court room. A very special

thank-you also goes to the other members of my doctoral committee at

the University Florida: Hugh Davis, Ph.D., and James Johnson, Ph.D.,

of the Department of Clinical Psychology; Franz Epting, Ph.D., of the

Department of Psychology; and Sandra Damico, Ph.D., of the Department

of Foundations of Education; whom have offered me intellectual

stimulation, constructive feedback, and support for this project as

well as other aspects of my professional growth and development.

I would like to thank my parents for their dedicated love and

support throughout my life for whatever goals I chose to pursue.

Finally, I would like to express my love and thanks to my wife and

colleague, Ana Kelton-Brand, Ph.D., who aside from teaching me a new

definition of panic, has provided me with love and affection and whose

devoted companionship is the most important thing in my life.



ACKNOWLEDGMENTS.....................................................i v

ABSTRACT................. ............................. ... ...... viii


I INTRODUCTION.............................................1

Methods of Studying the Jury.............................3
Personal Bias and Prejudice in Jurors' Decisions.........7
The Existence of Juror Bias.........................8
Sources of Bias and Jury Decisions.................10
Authoritarian Attitudes and Jury Decisions.........18
The Legal Attitudes Questionnaire.................22
The Reduction of Juror Bias.............................28
The Voir Dire ......................................28
The Evidence ...................................... 36
Judge's Instructions to the Jurors .................44
Deliberation........ .......................... .....46
Objectives ............................................ ..63
Hypotheses .............................................. 66

II METHOD ................................................... 69

Experimental Design ......................................69
Subjects .......................... .....................69
Description of the Case and Charges .....................71
Dependent Measures....................................... 73
Independent Measures ....................................78
Analyses................................................ 86

III RESULTS................................................. 90

Confidence Ratings of Juror Verdict Decisions...........91
Aggravated Assault against Officer White...........91
Aggravated Assault against Officer Rhodes..........94
Discharging a Firearm in Public....................98
Improper Exhibition of a Weapon...................102
Weighted Juror Verdicts................................105
Aggravated Assault against Officer White..........105

Aggravated Assault against Officer Rhodes.........109
Discharging a Firearm in Public...................109
Improper Exhibition of a Weapon...................112
Verdict ................................................ 120
Aggravated Assault against Officer White ..........120
Aggravated Assault against Officer Rhodes.........120
Discharging a Firearm in Public...................126
Improper Exhibition of a Weapon ....................126
Recall of Trial Information............................132
Recall of Situational Evidence....................132
Recall of Defendant's Character....................135
Recall of the Defense Attorney's Trial
Arguments....................................... 135
Recall of the Prosecution Attorney's Trial
Correlational Analyses............................ 145
Affective State........................................147
Deliberation Process................................... 148
Intercorrelations ................................. 148
Deliberation Time.................................152
Deliberation Styles...............................152
Comparison of Jurors' Pre- and Postdeliberation
Verdicts with the Jury's Verdicts....................157
Jurors' Perceptions....................................160

IV DISCUSSION............................................. 162

Confidence Ratings of Juror Verdict Decisions..........162
Weighted Juror Verdicts................................165
Recall of Trial Information............................178
Affective State .. ..................................... 181
Deliberation Process ...................................182
Deliberation Effects...................................186
Jurors' Perceptions....................................189

V SUMMARY................................................190

APPENDIX RECALL QUESTIONNAIRE...................................198

REFERENCES ...........................................................201

BIOGRAPHICAL SKETCH....................................... .......... 215

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy




December, 1985

Chairperson: Jacquelin Goldman, Ph.D.
Major Department: Clinical Psychology

The literature has suggested that biasing factors along the

authoritarian continuum distort verdict-relevant judgement, even in

the context of the trial procedure which attempts to limit and

counteract their influences. This investigation examined these biases

and at what points in the trial they were likely to be operative.

Seventy college students, classified as authoritarian,

egalitarian, or neutral, by subtracting the egalitarian score from the

authoritarian score on the Legal Attitudes Questionnaire (developed by

Boehm in 1968), served on twelve 6-person mock juries: authoritarian,

egalitarian, neutral, and a mixed composition of authoritarians and

egalitarians. These latter two types served as control groups. These

juries viewed a 4-hour videotaped reenactment of a trial concerning

aggravated assault against two police officers and two weapons

charges. At different intervals during the trial, jurors were asked


to render verdicts and their confidence in these verdicts. Each jury

was allowed up to an hour to deliberate to unanimous decisions. These

deliberations were recorded and analyzed. Jurors' recall of different

types of trial information was also assessed.

These data revealed that jurors serving on egalitarian juries

were the only jurors to take a stance of innocent until proven guilty

early in the trial. Additionally, for one charge, egalitarian juries

showed lower proportions of guilty verdicts than any other type of

jury. Authoritarians recalled less situational evidence than other

jurors. Furthermore, for one charge, authoritarian juries showed a

relatively high proportion of guilty verdicts, as did mixed and

neutral juries, despite their low levels of confidence in these

verdicts. The issues concerning which of these jury types were biased

were discussed.

The importance of the testimony, judge's instructions, and

deliberation in reducing juror bias was highlighted. Additionally,

the data suggested that juries comprised of a mixture of

authoritarians and egalitarians were likely to have heated debates,

expressing varying opinions, and might serve to reduce bias.

Finally, methodological issues emphasized the importance of

incorporating deliberations in jury research, the unreliability and

misleading results of studies using posttrial assessment weeks after

the trial, and the need to study the effects of biases on different

types of charges.


The Sixth Amendment of the U.S. Constitution guarantees the right

to a trial by an "impartial jury" in both criminal and civil matters

where the penalty for an offense is greater than six month's

imprisonment. Although an estimated two million persons serve as

jurors in some 200,000 civil and criminal cases each year (Abraham,

1980), this comprises only half of the criminal and civil cases in

federal district course and less than 10% of all cases (both civil and

criminal) in state trial courts (Vago, 1981).

In terms of numbers, therefore, trial by jury would appear to be

the least significant of our present modes of administering justice.

Nevertheless, the trial process has captured the interests of social

scientists for almost 20 years. In particular, the emphasis has been

on both extralegal factors and trial processes, especially

deliberation, which might influence the decision-making of the jury.

Emphasis in these areas appears for several reasons. The symbolic,

practical, and theoretical importance of the jury raises questions

that are open to empirical test.

Both Nemeth (1976) and Sealy (1979) have suggested that the trial

by jury is such an attractive object of study because of its symbolic

importance. That is to say, the jury somehow symbolizes the

systematic and free involvement of ordinary people in the

administration of law. In this context, the jury serves as a

safeguard against threats to freedom (Sealy, 1979). The practical

importance of the jury has been illustrated by Heurmann (1978). Here

it was shown that attributions regarding the case's outcome if the

case would go to trial influenced prosecutors and defense attorneys in

making decisions about plea bargaining or opposing counsel in private

disputes. Finally, social scientists in the fields of communication,

small group behavior, social psychology, and personality have employed

the trial process as a setting ". ideally suited to act as 'test-

beds' for psychological theories" (Sealy, 1979, p. 55).

Perhaps of greatest interest is the fact that, by definition, the

trial process is a method for applying the law impartially. Along

with this responsibility, the jury is given the authority to take away

freedoms and autonomy from the accused, bestow, or settle suits and

declare financial liability. However, the jury is one of the few

channels through which ordinary citizens can impose on society their

own standards or biases concerning what is morally or socially right

and wrong behavior (Cornish, 1968). This concern has been summarized

by Judge Jerome Frank: "Mr. Prejudice and Miss Sympathy are the names

of witnesses whose testimony is never recorded, but must nevertheless

be reckoned with in trials by jury" (1949, p. 122).

Unfortunately, it has been difficult to investigate the issue of

which factors have influence on the tasks of the jury. Ironically,

the very process, in which we entrust tremendous symbolic and

theoretical importance and actual power, also has presented itself as

an engima. Answers to questions regarding the jury have been elusive,

partly because empirical investigations of jury behavior are hampered

by legal restrictions relative to both the jurors themselves and for

the jury process. Nevertheless, since the right to impartial justice

by a jury of one's peers is central in our democracy, researchers have

found ways to study the jury system. Before discussing the findings

of this relevant research, the methods and the methodological issues

associated with the study of the jury will be discussed.

Methods of Studying the Jury

The Chicago Jury Project (Broeder, 1958) led to the important

publication by Kalven and Zeisel of The American Jury in 1966. Along

with the work of Simon (1967), they showed the feasibility of

empirical research in areas previously thought inaccessible and

focused researchers on the sorts of issues that could and should be

addressed empirically.

One of the most significant things about these endeavors was that

they employed actual jurors, trials, and judges. Kalven and Zeisel

conducted posttrial interviews of jurors from 12-person juries in 225

state and federal criminal trials and solicited "verdicts" from the

presiding judges. In this manner they reconstructed "first-ballot

votes" to compare to the actual jury verdicts. They also used judge-

jury agreement as the criteria for the veracity of actual verdicts.

Although the use of real stimuli and subjects allowed Kalven and

Zeisel to maximize the generalizability of their findings, they

encountered difficulties inherent in doing applied research. First,

although 90% of the jurors agreed to participate in the study, only

19% of the judges responded to the questionnaire. Additionally, the

use of posttrial interviews (often months after the trial) for

reconstructing the "first ballot" made it conceivable that some of the

jurors may have confounded their predeliberation decisions with the

group verdicts or were unwilling to admit that they were influenced by

the group to change their minds (Sonaike, 1978). Furthermore, because

of the inaccessability to study the jury in process, the authors were

not able to unravel the possible factors influencing the jury's

deliberation and verdict.

In an innovative attempt to clarify some of these issues, Zeisel

and Diamond (1978) utilized a "shadow jury." For 12 criminal cases

before 3 judges in the U.S. District Court in Illinois, the authors

used the actual jurors, who were peremptorily excused, to remain in

the courtroom and reveal at the end of the trial how they would have

voted. A second shadow jury was constructed by randomly selecting

from the remaining venire, four 12-person juries or "English

juries." Vinson (1982, 1983) used a somewhat different shadow jury.

He constructed a shadow jury which was matched in demographic and

psychological traits of actual jurors serving on a multimillion dollar

antitrust suit in California.

The great advantage of the shadow jury is that it allowed for the

questioning of jurors during the trial and offered access to the

deliberation process. This design was not without drawbacks. First,

it has been argued that, since shadow jurors know that their decisions

do not count, their motivation may be different from actual jurors.

This threatens the external validity of the findings. Additionally,

this method is not easily utilized because of its great cost.

Analyses of simulated juries and trials have begun to answer some

of the questions raised about jury decision-making. Because of lower

costs and the ease with which data can be accumulated, simulation

studies comprise the majority of research in this area. Simulation

lends itself to experimental manipulation and control of variables

which otherwise could not be isolated. These techniques have varied

widely in stimuli presented, types of juror subjects, setting or

location, trial components included in the simulation, dependent and

independent variables, and unit of analysis (i.e., juror or jury;

product or process) (Bray & Kerr, 1982).

It is obvious that the most important methodological issue

confronting the study of the jury is the trade off between internal

and external validity. All research in this area falls somewhere

along this continuum. Both general strategies have advantages and


Several legal and psychological writers (Konecni & Ebbesen, 1982;

Vidmar, 1979; Weiten & Diamon, 1979) typically favored the field

methods which closely approximate realistic situations. However,

Monahan and Loftus (1982) highlight that the principal liability of

these methods is that, in addition to high cost, they are often

illegal, unethical, or impossible to control in order to provide the

opportunity for internally valid research. Additionally, in order to

avoid concern over the possibility of jury tampering, information from

the jurors cannot be obtained until after the trial. This technique

may cause problems with recall and the possibility that aspects of the

trial and deliberation might color posttrial reports.

In contrast, jury simulation studies have been the preferred

method of study by most psychologists (Baldwin & McConville, 1979;

Thibaut & Walker, 1975). As mentioned previously, simulation studies

allow for randomization of variables, replication of procedures,

access to process (e.g., deliberation), and savings in time and

money. These advantages maximize internal validity.

Nonetheless, several criticisms have been launched at simulation

research regarding its generalizability (Bray & Kerr, 1982; Davis,

Bray, & Holt, 1977; Gerbasi, Zuckerman, & Reis, 1977; Vidmar, 1979;

Weiten & Diamond, 1979). Inadequate trial presentations have often

been used. Here, the method of presentation (e.g., brief and

simplistic audio or written hypothetical case studies), the exclusion

of important trial elements (e.g., opening arguments, voir dire, and

judge instructions), and the use of laboratory settings have been

highly unrealistic. Additionally, subjects (usually college students)

are unrepresentative of actual jurors on several dimensions. Jurors

are also often not asked to deliberate despite the research findings

that deliberation has had a significant effect upon trial outcome

(Bray & Noble, 1978; M.F. Kaplan & Miller, 1978; McGuire & Bermant,

1977). They often vote on an artificial continuous measure of guilt

rather than on dichotomous ratings (i.e., guilty or not guilty). They

are also asked to pass judgment on severity of sentence which jurors

are usually not asked to do. Finally, there have been concerns that

subjects often perceive the consequences of the decision task

different from real jurors. Equivocal results preclude firm

conclusions regarding this last issue (Kerr, Nerenz, & Herrick, 1979;

Wilson & Donnerstein, 1977; Zeisel & Diamond, 1978).

Judgments of the merits of jury research must be made with the

above issues in mind. Several authors have discussed the utility of

both research strategies (Bray & Kerr, 1982; Monohan & Loftus,

1982). In essence, both investigated the extent to which our theories

of human behavior can predict and explain findings. Hypotheses are

continuously being formulated and tested. Although it would be ideal

to find a correct balance between experimental control and

authenticity, this is not always possible. However, it has been

suggested that the more numerous and independent ways in which effects

or results are demonstrated to support a hypothesis, the less numerous

and less plausible any single rival invalidating hypothesis becomes

(Campbell & Stanley, 1963). Therefore, both applied and experimental

studies are worthwhile contributors to the study of the jury.

With a discussion of the methods and methodological issues

involved in jury research completed, the literature in this area can

now be reviewed with a greater appreciation of its utility and

limitations. Most of the available research has been some type of

simulation study; however, relevant field studies will also be

included in the review.

Personal Bias and Prejudice in Jurors' Decisions

To what extent and by what means a juror's personal decision, as

well as the group process of reaching a consensual decision, is

influenced by biasing factors has long concerned the legal and

scientific community. Erlanger (1970) suggested that personal

characteristics of jurors affect both the individual's decision and

the process through which a jury comes to a collective verdict.

The Existence of Juror Bias

In one of the earliest jury simulation studies, Weld and Danzig

(1940) showed that 25% of their jurors reached a definite decision

early in the trial and testimony served only to strengthen their

certainty. About the same time, the United States Supreme Court had

to rule on legal challenges to jury panels which were partially based

on the biases of the jury (Fay vs. New York, 1947; Thiel vs. Southern

Pacific Co., 1946). Indeed, the ruling in the former case was

actually reversed because of a biased jury panel. Such findings and

rulings are what probably lead Judge Frank (1949) to express his

previously quoted concern about extra-evidential biases.

In the context of this atmosphere, the Chicago Jury Project was

undertaken in the 1950s. It was, therefore, expected that the

findings of Kalven and Zeisel (1966) would confirm Judge Frank's

charges that the typical jury lacked the discipline and understanding

to render fair decisions in most cases (Frank, 1949). However, their

findings were taken to contradict Frank. They found that judge-jury

disagreement only occurred in 34% of the trials examined.

Furthermore, only 19% were total disagreements (i.e., one acquitted

and one convicted). The remaining 15% encompassed hung juries and

differences on charge and penalty.

The authors determined that 54% of the reasons for judge-jury

disagreement were due to issues of evidence (e.g., credibility of

first-time defendants, threshold of proof beyond reasonable doubt).

Feelings about equitable issues (e.g., fault of victim, disproportion

of punishment, selective enforcement of the law) accounted for 29% of

the disagreements and only 11% of the disagreements were due to

attitudes toward the defendant (e.g., age, status, attractiveness).

With these two types of extra-evidential influences (i.e., equitable

issues and attitudes toward defendants) accounting for only one-third

of the judge-jury disagreements it was interpreted that extra-

evidential factors did not adversely affect the ability of the jury to

administer impartial justice.

The findings of Kalven and Zeisel (1966) are of great importance;

nevertheless, their conclusions may have been misleading. In addition

to the methodological problems cited earlier, the authors failed to

consider that evidentiary issues, such as inferring credibility of

certain types of witnesses and proper threshold for decision, may

themselves be influenced by biasing factors (Wasserman & Robinson,

1980). Other authentic studies have also suggested that biasing

factors might be operative in influencing jury decision-making. The

results of the shadow juries of Zeisel and Diamond (1978) (to be

discussed in more detail later) were supportive of this notion.

The shadow jury studied by Vinson (1982, 1983) was shown to come

to a decision before testimony and then only sought information to

support their premature decision. This was supportive of the earlier

findings of Weld and Danzig (1940). Perhaps the effects of bias are

most potent prior to the presentation of evidence. Indeed, Freundlich

(1983) recently found that in four out of six criminal charges 75% of

the mock jurors studied had decided on a guilty verdict after the

opening argument and became increasingly confident in their decision

for the rest of the trial. Thus, it appears that verdict-relevant

biases are apparent before evidence is presented and, as Pyszczynski

and Wrightsman (1981) suggested, may serve as a "thematic framework"

that serves to guide juror information processing.

Sources of Bias and Jury Decisions

Several authentic studies have suggested the presence and

potential impact of biasing factors in juror and jury decisions.

Simulation studies have allowed psychologists to attempt to explain

the sources of bias that might render one particular juror more

susceptible to prejudice than another. A number of demographic,

situational, personality, and attitude dimensions have been

investigated (for reviews of this literature, see Davis, Bray, & Holt,

1977; Gerbasi, Zuckerman, & Reis, 1977; Nemeth, 1981; Saks & Hastie,


The most frequently studied juror demographic variable has been

sex. The data are mixed with respect to conviction proneness of each

sex across different types of charges. Nemeth (1981), in an extensive

literature review, cited several studies in which males were found to

be more conviction prone and/or more punitive, other studies in which

females were more conviction prone and/or punitive, and still other

studies in which no sex differences were obtained. Nemeth reported

that only in the case of rape have females been consistently found to

convict more and be more punitive across different studies.

There have been several other studies which have reported sex

differences in jurors with respect to demographic variables,

personality styles, attitudes, and deliberation styles, which might

influence verdict decisions. Strodtbeck and Mann (1956) showed that,

during a simulated deliberation, actual female jurors were more likely

to assume traditional roles dealing with social-emotional issues,

while males tended to be more task-oriented. Strodtbeck, James, and

Hawkins (1957) reported that actual male jurors, in contrast to female

ones, showed higher participation, influence, satisfaction, and

perceived competence during mock deliberations. Using a written case

summary of a rape trial, Rumsey and Rumsey (1977) found that although

there were no sex differences in verdict, females were generally more

certain of guilt and attributed more responsibility to the

defendant. Mills and Bohannon (1980) showed differential verdict

responses in highly socialized and empathetic male and female mock

jurors. Another study by Moran and Comfort (1982), using 319 felony

jurors (22.7% of the sample who responded) impaneled in the Dade

County, Florida, state courts, found differences between male and

female jurors who convicted. Males who convicted were found to have

more children, lower income, and reported themselves to be less

concerned with responding in a socially desirable manner; females who

convicted tended to report a belief in a just world and that their

actions could affect society as well as higher authoritarian

beliefs. These variables, however, only accounted for 10.7% and 11%

of the variance for males and females, respectively. Finally, two

recent simulation studies (Freundlich, 1983; Freundlich & Goldman,

1982) showed that female mock jurors were not different from males on

verdict or deliberation style, but reported higher levels of

depression, anxiety, and hostility throughout the trial.

Clearly, the results with respect to sex are inconsistent and can

only lead to tentative suggestions that this may be an important

variable to consider when conducting jury research.

Age of the juror has been shown to be associated with verdict.

Several investigations found that younger jurors tended to acquit more

than older jurors (Scroggs, 1976; Sealy & Cornish, 1973; Stephan &

Tully, 1977), although one study did not find age differences (Reed,


The race of the juror also appears to be important although these

findings are also equivocal. Evidence has been presented that black

defendants are more likely to be convicted and given harsher sentences

by white jurors (Broeder, 1965; Greenberg, 1959). However, more

recent studies show that both black and white jurors were more likely

to find the defendant guilty if the defendant was similar in race to

them (M. Miller & Hewitt, 1978; Ugwuegbu, 1976).

Other demographic variables, such as education, occupation,

income and political persuasion, have been investigated. In the

Strodtbeck et al. (1957) study those jurors with high status

occupations showed higher participation, influence, satisfaction, and

perceived competence during mock deliberations than those with low

status occupations. James (1959) found that more highly educated

jurors reported that procedures and instructions most influenced their

trial decisions. Jurors with lower education levels, on the other

hand, were reportedly more likely to be influenced by personal life

experiences and opinions from the trial. Reed (1965) also found that

as education increases the likelihood of voting guilty increases.

Simon (1967) found that jurors with higher monthly incomes were less

likely to vote not guilty than those with lower monthly incomes.

Nemeth and Sosis (1973) found that mock jurors from politically

conservative colleges gave harsher sentences to lower status

defendants than higher status defendants. Jurors from a liberal

college did not appear to make this distinction.

Although available data have suggested that demographic variables

may indeed influence jurors' trial behavior and decisions, these

findings are equivocal. It might be reasonable to suggest that these

variables may interact with each other and other variables such as

type of case, defendant characteristics, and personality of the juror.

The personality of the juror has been of interest to those

studying juror decision-making. Here, both transient (state)

characteristics and/or relatively permanent (trait) characteristics of

jurors have been relevant. Fluctuations in affective state has been

the most commonly studied transient characteristic in jurors. Using

an inferential measure of affective state, Simon (1967) showed that

jurors who convicted were more anxious than those who acquitted the

defendant in an incest trial. Goldman, Maitland, and Norton (1975)

found similar results using a more direct measure of affective state

in a simulated criminal insanity case. The same study found that mock

jurors showing greater levels of hostility over the trial duration

tended to give more severe penalties. Finally, Freundlich and

Goldman (1982) showed that lower levels of emotionality (i.e.,

depression, hostility, and anxiety) were reported more often in mock

jurors of innocent-deciding juries. One cannot safely determine the

relationship of the affective state to the decisions of the jurors

from these data; although one possibility is that as a juror

experiences greater emotion, it may be easier to make a decision to


Another variable, level of moral judgment, appears to potentially

clarify the relationship between affective state and juror

decisions. For example, Goldman and Thomas (1979) found that low

moral level mock jurors showed greater increases in hostility. Using

a less affective-arousing simulated trial, Goldman and Casey (1980)

demonstrated that low moral level mock jurors reported more extreme

affective responses, especially depression and hostility. Since level

of moral judgment, a cognitive variable, was not found to influence

verdict in these studies, the potential mediating effects of this

variable between affective state and juror decisions must remain


Researchers have also studied trait or global personality

characteristics. Several personality constructs have been employed

here. One such construct is the belief in a just world (Lerner,

1970)--the belief that one is likely to get what he deserves and

deserves what he gets. This idea was used to explain the fact that

under certain conditions innocent victims are blamed for their


Zuckerman and Gerbasi (1973) tested this idea using a simulated

rape case. They found that mock jurors who scored high on the Just

World Scale (Rubin & Peplau, 1973) assigned more responsibility to the

rape victim. This supported Lerner's hypothesis that derogation of

the victim serves to maintain a belief in a just world.

Demanding harsher treatment of the defendant may also serve to

maintain a belief in a just world. Rubin and Peplau (1975) reported a

study by Izzett which employed mock jurors in a simulated negligent

homicide case. Here, jurors scoring high on the Just World Scale

formed less favorable impressions of the defendant and assigned

stiffer sentences than did jurors scoring low on this measure.

Gerbasi and Zuckerman (1975) found similar results using a simulated

murder case with a population of real jurors. The results of the

above studies suggest that belief in a just world may be a biasing

factor in juror decisions. However, to what extent deliberation

interacts with this personality construct remains unknown since none

of these studies had their jurors deliberate. Also, severity of

punishment was used as a dependent measure; however, in real trials

jurors only decide on the guilt of the defendant.

Internal-external control (Rotter, 1966) has been another

personality construct studied for its relationship to sentencing

behavior and attribution of responsibility to a defendant. Phares and

Wilson (1972) showed that externals tended to favor the defendant in

several auto accident cases. In a similar experiment, Sosis (1974)

found internals to be more likely to give harsher sentences and hold

the defendant responsible for his crime than externals. The authors

explain their findings in terms of projection of self-perceptions of

responsibility. Unfortunately, the criticisms against the studies on

belief in a just world are relevant here too.

The relationship between character structure--assessed in terms

of socialization, empathy, and autonomy--and mock juror behavior was

recently investigated in a study by Mills and Bohannon (1980), who

used felony jurors impaneled in Baltimore, Maryland. Their data

showed that males with high socialization scores were associated with

the belief that the jurors' task was to reach some decision and

expressed satisfaction in making their decision. Females with high

socialization scores, however, expressed more uncertainty about making

the right decision. Additionally, highly socialized males gave more

guilty verdicts, whereas highly socialized females gave fewer. High

empathy scores in males, in contrast, were consistently associated

with not guilty verdicts as were high autonomy scores for both

sexes. Again, unfortunately, there were no jury deliberations.

The construct of ego development by Loevinger and colleagues

(Loevinger, 1976, 1979; Loevinger, Wessler, & Redmore, 1970) and its

association with verdict in mock jurors has also been investigated.

Ego development, a measure of psychological sophistication or

maturity, has been used by Freundlich and Goldman (1982). Here, they

found that high ego level mock jurors reported experiencing lower

levels of emotionality and were more likely to vote not guilty than

those with low ego levels. Freundlich (1983) supported this finding

by showing that low ego level mock jurors were more likely to vote

guilty on a simulated case of aggravated assault than those jurors

with high ego level. Indeed, high ego level jurors were more likely

to render not guilty verdicts even when exposed to guilt-biasing

pretrial publicity. These studies had jurors deliberate as a jury.

That deliberation is an important part of jury decision making was

supported by the finding that ego level was related to deliberation

style (Goldman, Freundlich, & Casey, 1983). It therefore appears, at

least tentatively, that ego level might be an important biasing

variable in jury decision making.

Level of moral development (Kohlberg, 1969, 1976), a construct

not dissimilar to that of ego development, has also been shown to be

associated with mock juror's decisions about verdict and sentencing.

Arbuthnot (1983) found that, in an ambiguous simulated murder case,

the jurors' decisions about the degree of guilt and severity of

sentence decreased significantly with increasing stages of moral

development. Unfortunately, this study was performed with a written

summary of the trial and no jury deliberations took place.

Nevertheless, this study, taken with the studies using ego

development, suggests that further research employing such constructs

would seem warranted.

Not all studies have found that juror biases are related to

decisions regarding verdict. Sealy (1981) studied fifty 12-person

mock juries comprised of a representative adult population of Greater

London. They listened to verbatim audiotape recordings of either a

rape or a theft trial and rendered verdicts before and after

deliberation. Sealy's results showed that few variables correlated

with verdict either before or after deliberation. That is,

characteristics, such as occupational status, sex, educational

experience, dogmatism, rigidity, and authoritarianism, were not a

particularly significant factor in trial decisions. The only

significant findings were a slight tendency for younger (up to age 25)

and older (above age 40) to acquit, and that jurors with more

favorable attitudes toward the jury system tended to convict. Sealy

concluded that consideration of evidence is of greatest concern to

jurors and, although background and attitude variables may influence

how this is done, they have little or no influence on their final


Despite Sealy's findings and conclusions, one might still

consider that there are verdict-relevant attitudes which might be

especially important in studying juror biases. Indeed, recent reviews

of the literature on juror biases have been replete with studies

investigating the association between authoritarian beliefs (and

related constructs) and juror decision making (Gerbasi et al., 1977;

Hans & Vidmar, 1982; M.F. Kaplan, 1982). Because of the consistent

findings using this attitude construct and its importance to the

present investigation, the literature in this area will be reviewed in

a separate section presently.

Authoritarian Attitudes and Jury Decisions

Authoritarianism is the term used to describe an attitude system,

which ascribes to the belief that one should unquestionably accept

authority from recognized powerful people and institutions. Persons

with this attitude are expected to conform strictly to conventional

social norms and demonstrate rigid prejudice toward those who are

different or do not embrace their point of view, conservatism, and

show hostility toward those who deviate from established social norms

(Adorno, Frenkel-Brunswick, Levinson, & Sanford, 1950).

Ellison and Buckhout (1981) have stated that authoritarian

attitudes are common in people in varying degrees. Thus, it is

possible to speak of an individual as high or low on authoritarian

attitudes. Individuals who appear at the latter end of the continuum

are often called egalitarian. They are believed to not place great

value on conventional norms and have more tolerance for those who


Ellison and Buckhout (1981) asserted that this attitude dimension

very likely plays a significant role in decisions made by jurors. It

has been suggested that as jurors, authoritarians may convict more

readily and render harsher punishments than egalitarians (Adorno et

al., 1950). Indeed, according to Ellison and Buckhout (1981)

authoritarian attitudes have been one of the best predictors of

conviction that they have encountered. This literature will now be


Centers, Shomer, and Rodrigues (1970) conducted a field study

employing an authoritarian measure by Sanford and Older (1950), with

over 1,000 subjects. They found that high authoritarians were more

punitive towards a juvenile defendant than either low or medium

authoritarians. However, high authoritarians were most likely to

change their recommendations for punishment when presented with an

"expert's" views that contradicted their own.

Mitchell and Byrne (1973) studied mock jurors, assessed for

degree of authoritarian attitudes with an acquiescence-free

authoritarianism scale. They were presented with a case history which

led them to believe that the defendant was similar or dissimilar to

themselves. Authoritarians were more likely than egalitarians to

recommend a more severe punishment. When the defendant was perceived

as similar, authoritarians were unsure of the defendant's guilt more

often than egalitarians.

Bray and Noble (1978) also classified mock jurors as high or low

authoritarians utilizing the F-Scale (Byrne, 1974). Subjects listened

to a simulated murder trial. The results showed that high

authoritarians voted the defendant guilty more often and issued longer

sentences than low authoritarians both before and after

deliberation. Additionally, juries composed of high authoritarians

reached decisions of guilty more than low authoritarian juries.

Unlike Mitchell and Byrne (1973), Bray and Noble showed that perceived

similarity to defendant did not affect jurors' judgment.

Werner, Kagehiro, and Strube (1982) recently investigated the

ability of jurors to disregard inadmissible evidence. Mock jurors

were characterized as high or low authoritarians using the Mitchell-

Byrne Authoritarianism Scale (Mitchell & Byrne, 1973) and presenting

them with a written summary of the crime and trial. Their results

showed that high authoritarians incorporated incriminating evidence

almost regardless of admissibility, but not exonerating evidence in

rendering a decision. Furthermore, they were more likely to convict

even when they were presented with judges' instructions explicitly

stating them to disregard inadmissible evidence. Interestingly, low

authoritarians became more lenient in the presence of inadmissible


Lamberth, Krieger, and Shay (1982) investigated the degree to

which changes in juror decision making are mediated by

authoritarianism. These authors performed a series of three

experiments, two using college students and one using individuals

drawn from the venire at the Philadelphia Court of Common Pleas. In

each study, jurors were classified as authoritarian or egalitarian

based on their responses on a counterbalanced F-Scale (Kirby &

Lamberth, 1974), viewed a videotape of a trial based on an actual case

of attempted rape and burglary, and deliberated in 12-person juries.

Across experiments, authoritarian jurors were more likely to render

guilty verdicts on predeliberation ballots than egalitarian jurors and

authoritarian jurors were more likely to change their opinions during

deliberations (as assessed by postdeliberation questionnaires) than

egalitarian jurors. These authors reported that sex, age, and race

were not predictive of change in verdict. Supporting the

interpretation, that the major predictor regarding a change in one's

vote during deliberation was authoritarianism, was the fact that 78%

of those jurors who changed their votes were classified as

authoritarian. Perhaps the most provocative finding was that most of

the juries, which contained a mixture of egalitarian and authoritarian

jurors, were either hung or voted not guilty; while the only juries to

convict the defendant were comprised entirely of authoritarians.

These data suggest that authoritarians may favor guilty verdicts when

left to their own judgments; however, when faced with alternative

points of view (i.e., from egalitarians during deliberation), the

authoritarian juror may be more malleable. Lamberth et al. (1982)

conceptualized their findings by suggesting that the authoritarian

juror is more likely to respond in a manner that increases positive

affect than is the egalitarian juror, and that accepting the arguments

of the perceived power structure generates positive affect. They went

on to postulate that at the predeliberation ballot the perceived power

structure is the prosecution, who represents the full authority of the

law, but that during the deliberation the perceived power structure is

the majority opinions of the group, which may be more egalitarian.

Clearly, these issues are worthy of further study.

Research with variables associated with the concept of

authoritarianism has provided supportive results. For example,

conservative mock jurors (Nemeth & Sosis, 1973) and real jurors

classified as dogmatic (Hatton, Snortum, & Oskamp, 1971; Rokeach &

McLellan, 1969-1970) appear to render harsher sentences than their

opposites. Additionally, mock jurors with stern attitudes toward

punishment of criminals report both greater certainty of guilt and

harsher sentences in simulated cases (M.F. Kaplan & Miller, 1978; M.F.

Kaplan & Schersching, 1980).

The Legal Attitudes Questionnaire

It appears that an attitudinal system such as authoritarianism

and its associated attributes might be associated with juror and jury

decision making. However, a major criticism of this type of research

has been suggested (Boehm, 1968). The independent measures employed

such as the F-Scales (Adorno et al., 1950; Byrne, 1974; Mitchell &

Byrne, 1973), Dogmatism Scale (Rokeach, 1960), and the scale used by

Centers et al. (1970) deal with attitudes towards a wide variety of

things. According to Boehm, "few of the items measure attitudes

specifically relevant to the jury decision-making process" (1968,

p. 739). Indeed, recent writers (Ajzen & Fishbein, 1977) have

reviewed the literature on attitudes and behavior and suggested that

attitudes toward specific behaviors (e.g., the treatment of criminals

in the judiciary system) will provide the best predictor of single-act

criteria (e.g., jury decision making).

Boehm (1968), therefore, developed a measure, with moderate to

high correlations, with the F-Scale (Adorno et al., 1950) and

Dogmatism Scale (Rokeach, 1960), which would be directly relevant to

the jury decision process. This Legal Attitudes Questionnaire

(LAQ)-Form II (Boehm, 1968) used a forced-choice format to yield three

attitude dimensions: authoritarianism, egalitarianism, and

antiauthoritarianism. Authoritarian items expressed right wing

sentiments and unconditional endorsement of acts of constituted

authority, with a trend toward punitiveness and rigidity. Egalitarian

items endorsed liberal, nonextreme, and tolerant positions. An

antiauthoritarian dimension, based on the findings of Weitman (1962),

consisted of extreme left wing sentiments which blame society for

antisocial acts and endorse positions which are intended to thwart


In addition to providing construct validity by showing moderate

to high correlations with psychological measures of authoritarianism,

Boehm (1968) tested its predictive validity. Mock jurors were given

the LAQ as a theoretical voir dire and were provided with one of two

versions of a manslaughter case. One version biased the evidence in

the direction of guilt and the other towards innocence. It was found

that authoritarians rendered more guilty verdicts ("tough" errors)

than were warranted in the "not guilty" case and antiauthoritarians

rendered more not guilty verdicts ("lenient" errors) than were

warranted in the "guilty" case. The type and frequency of verdicts

made by egalitarian jurors were not reported. Boehm did report that,

of the jurors making hypothetical errors in verdict, 72% of the

authoritarians, 52% of the antiauthoritarians, and 42% of egalitarians

were highly confident of their "incorrect" decisions.

Boehm's data support the findings of others (Freundlich, 1983;

Vinson, 1982, 1983; Weld & Danzig, 1940), which have contended that

authoritarians and antiauthoritarians reached their verdicts early in

the trial (i.e., after opening arguments) and resisted changing their

verdict in the face of new information. A field study using the LAQ

("Juror Bias," 1980) found similar results.

Jurow (1971) used the LAQ with adult mock jurors who listened to

recordings of murder trials. It was reported that authoritarians were

more in favor of the death penalty and more likely to convict than

egal itarians.

Alexander and Licker (1975) reported similar results. They found

the LAQ to be a good predictor of individual mock jurors' verdicts in

a simulated murder trial. Here, authoritarians were more likely to

render guilty verdicts than egalitarians.

Most recently, Goldman, Freundlich, and Thomas (1983) found that

authoritarian mock jurors, classified by the LAQ, showed a tendency to

convict more often than other mock jurors in a simulated trial

involving an attempted murder and aggravated assault of law officers.

Of further interest, they found that authoritarian mock jurors tended

to favor conviction more than other mock jurors at all stages of the

trial (i.e., after opening arguments, after testimony, and after


In another study using the LAQ, Buckhout et al. (1979) found that

mock juries composed of egalitarians were more likely to acquit than

authoritarian or mixed juries in a simulated murder case.

Berg and Vidmar (1975) conducted a study to attempt to explain

the differences in authoritarian and egalitarian verdicts. They

divided mock jurors into high and low (egalitarian) authoritarians by

subtracting the egalitarian score from the authoritarian score on the

LAQ. Subjects were presented with a case similar to that used by

Mitchell and Byrne (1973), which included both facts about the case

(situational data) and information about the defendant's character

(which included information making the defendant of high or low

status). The results showed that all subjects were more certain that

the lower status defendant was guilty; however, high authoritarians

were more punitive, especially toward the lower status defendant. Of

particular interest, on a 7-10 day recall task, high authoritarian

subjects recalled more about the defendant's character than about the

situational evidence. Low authoritarian subjects, in contrast,

recalled more situational evidence and less about the defendant's

character. The authors concluded that differences in recall might

partially explain differences in the attribute of guilt by high and

low authoritarians.

Berg and Vidmar (1975) conducted a second study with different

mock jurors, a different authoritarian measure (Byrne & Lamberth,

1971), and a different case (automobile manslaughter). Eliminating

the part in which the subjects decided on verdict, in order to control

for the influence this may have had on recall, similar results to

their first study were found.

The relevance to jury behavior of memory being subject to

systematic biases has been discussed in a recent review by Sherrod

(1985). The research findings were summarized to indicate that people

recall, organize, and reconstruct "facts" to fit this theme.

Similarly, Pyszczynski and Wrightsman (1981) had previously shown that

the opening argument can act as a "thematic framework" into which

"facts" are integrated. Perhaps the findings of Berg and Vidmar

(1975) may be taken to suggest that authoritarians are more likely to

employ such a strategy leading to selective recall of this


The findings of Berg and Vidmar (1975) also supported Boehm's

(1968) observations that authoritarians base their decisions on

impressions of the defendant's character. These results are also

similar to previous research with authoritarians. For example,

Johnson and Steiner (1967) found authoritarians to be "source-

oriented" and nonauthoritarians to be "message-oriented."

To summarize, the literature on authoritarianism, especially as

assessed by the LAQ, has shown fairly consistent findings.

Authoritarian jurors appear more conviction prone than egalitarians or

antiauthoritarians. They are also more likely to give harsher

punishments. Research supports the hypotheses that authoritarians

make their conviction prone decisions early in the trial and may

selectively attend to information, such as the defendant's character

and inadmissible incriminating evidence, to support their bias.

Some tentative suggestions can also be made. There is some

support that defendant characteristics, such as perceived similarity

or status, may interact with the dimension of authoritarianism;

however, these results are less consistent (for a review of the

influences of defendant characteristics on jurors, see Dane &

Wrightsman, 1982).

Another set of interesting findings is that, in addition to

antiauthoritarians, egalitarians may be biased towards leniency. This

is important to clarify with further research since it has normally

been assumed that egalitarians are impartial and unbiased (Boehm,

1968). The issue of reducing juror bias will be discussed shortly.

First, it is noteworthy of mention that many of the studies above

did not employ jury deliberations in their methodology. This issue

will be discussed in a later section where it will be shown that the

dimension of authoritarianism may interact with the deliberation

process (Alexander & Licker, 1975; Bray & Noble, 1978; Buckhout et

al., 1979).

The Reduction of Juror Bias

As was stated in the beginning of this review, the Sixth

Amendment guarantees to each criminal defendant the right to trial by

an impartial jury. In the spirit of protecting this constitutional

right, several aspects of the trial process can be considered means

through which jury impartiality might be assured. These included the

voir dire, the presentation of evidence or testimony, judge

instructions to the jury, and deliberations. Relevant issues and

findings generated from the study of these different trial components

will now be discussed.

The Voir Dire

The term voir dire has been translated to mean "to see, to

tell." It refers to the first phase of the trial in which prospective

jurors are questioned and selected to sit on the impaneled jury.

There are three judicially accepted purposes of the voir dire: (a)

to determine whether prospective jurors are legally qualified to sit

as jurors (i.e., meet age and residence requirements); (b) to

determine whether prospective jurors can make an impartial decision

based only on the law and evidence presented in the trial; and (c) to

allow the attorneys to obtain information that will aid them in

intelligently employing their limited number of peremptory challenges

to dismiss biased prospective jurors (Suggs & Sales, 1978). The first

two purposes are associated with establishing reasons to dismiss

prospective jurors for cause. A prospective juror can be dismissed

for cause when it is very apparent that he is unable to render an

impartial decision because of his experience, occupation, or personal

or financial interests. Although challenges for cause are of

unlimited number, they are subject to the final discretion of the

judge who may overrule the challenge (Fried, Kaplan, & Klein, 1975).

The number of peremptory challenges allowed, in contrast, is fixed by

law or court rule, but they are not subject to judicial discretion.

That is, they can be for unspecified reasons of the attorney. When

any type of challenge is accepted, the prospective juror is dismissed

and a new prospective juror is randomly chosen from the venire and the

above process is repeated (Fried et al., 1975).

Lawyers must, therefore, be shrewd in their use of peremptory

challenges because there is always risk that subsequent prospective

jurors may be more unsuitable (e.g., biased in an unfavorable

direction) than the one who was dismissed (Broeder, 1965). For this

reason, the legal literature has provided suggestions for jury

selection which, for the most part, are based on folklore, common

stereotypes, and untested hypotheses (Mahoney, 1982).

There are few existing studies which assessed the effectiveness

of the voir dire in identifying and removing biased jurors. To test

this question, Pawader-Singer, Singer, and Singer (1974) compared

jurors selected at random to jurors selected by lawyers through the

voir dire process. Their findings suggested that jurors selected by

the voir dire were more sympathetic to mitigating circumstances, less

influenced by prejudicial pretrial publicity, and displayed fewer

shifts of opinion during deliberation. The authors concluded that

lawyers may be effective in employing the voir dire to select a

favorable jury panel.

In contrast, other studies have yielded less favorable results as

to the efficacy of the popular use of the voir dire. For example,

Broeder (1965) conducted posttrial interviews of 225 actual jurors who

had served on 1 of 23 consecutive federal district court cases. He

reported that the attorneys conducting the voir dires were

unsuccessful at both identifying and eliminating unfavorable jurors.

In an investigation cited earlier, Zeisel and Diamond (1978)

created shadow juries for 12 trials in a federal court, composed of

jurors who had been peremptorily challenged in each trial. An

"English jury" (i.e., a jury without challenges) was reconstructed for

each case by using the first 12 jurors (accepted or dismissed) who

were randomly called from the venire. Their first and final ballot

votes were estimated. They reported that the average first ballot

votes of English and actual jurors were quite similar. However, the

estimated final ballots showed that in 5 of the 12 cases, the English

juries demonstrated a greater probability of guilty verdict than the

actual juries. Interestingly, judge-jury disagreements occurred on

the same five cases (which were all acquittals). In other words, in

these five cases the actual juries (i.e., the result of peremptory

challenges) voted for acquittal and showed disagreement with the

judges, who like the English juries voted for guilt. These results

might be taken to support the defense attorney's ability to employ

peremptory challenges effectively in the voir dire. However, Zeisel

and Diamond (1978) calculated a "performance index" which indicated

inconsistent and variable performances by all the attorneys and showed

that the average performance scores for both the prosecution and

defense attorneys were near the zero point. They concluded,

therefore, that attorneys demonstrated an inability to distinguish

potential bias in jurors.

Clearly, there is a paucity of concrete information about the

effectiveness of lawyers' selection techniques during the voir dire.

The information that is available has yielded equivocal results. At

best, the present research suggests that general jury selection

strategies may be only minimally effective. Zeisel and Diamond (1978)

suggested that increasing the amount of information on which lawyers

base their decisions might improve the disparate performances noted in

their study.

Indeed, such an approach has been employed by social scientists

since the early 1970s in a broad number of techniques encompassed

under the rubric of "scientific jury selection." Scientific jury

selection involves the use of a number of social science methodologies

and bodies of data. These include pretrial investigative techniques,

such as surveys to assess community opinion, the construction of juror

profiles, the use of information networks in the community to obtain

information about demographic and personal characteristics of

potential juror, expert observation of linguistics and body language

in the courtroom, and the application of sophisticated mathematical

formulas (Covington, 1983; Hans & Vidmar, 1982; Lees-Haley, 1984;

Mahoney, 1982).

There have been numerous anecdotal reports of success with the

use of scientific jury selection. Some of these have been well

publicized trials, such as the Vietnam Veterans against the War trial,

the Wounded Knee trials, the trial of John Mitchell and Maurice Stans,

and the trial of Joan Little (Saks, 1976; Schulman, Shaver, Coleman,

Emrich, & Christie, 1973; Zeisel & Diamond, 1976).

Scientific jury selection has been criticized for a number of

reasons by Hans and Vidmar (1982) in their review of the literature.

It was argued that this selection procedure may have enhanced the

success of the defense attorneys via a "placebo effect." That is, the

defense attorneys may have worked harder. Additionally, the jurors

may have responded to the demand characteristics of the defense

attorneys because they were being chosen by "experts" (McConahay,

Mullin, & Frederick, 1977). It was also suggested that the voir dire

may have served to indoctrinate the jurors to an unusual degree about

being impartial (Mahoney, 1982; Suggs & Sales, 1978). Hans and Vidmar

point out that the weak evidence by the prosecution in such conspiracy

cases could easily account for the noted acquittals. A final, but

obvious criticism was that these reports utilized no control groups.

Horowitz (1980) addressed this issue by actually comparing

conventional jury selection methods used by lawyers to systematic

social science methods. Potential jurors were screened by advanced

law students using either conventional folklore or the social science

technique used by Schulman et al. (1973) in one of four mock criminal

cases. The results showed that in cases where actuarial data showed

strong predictions (e.g., the sale of illegal drugs and a court

martial case) the scientific method was superior, but where they were

weak the conventional method was equal or superior. Horowitz

concluded that scientific methods overall were not superior to

conventional methods unless the interaction with the particular type

of case was considered.

Scientific jury selection has also been subject to ethical

criticisms (Herbsleb, Sales, & Berman, 1979; Mahoney, 1982). One

concern involves a violation of an individual's personal life by

invasion of privacy when investigating the prospective juror. In

addition, often such techniques are extremely costly, which would

limit their availability to many. It would clearly run counter to the

notion of impartial justice if juries were selected to the advantage

of only those with sufficient money. Similarly, it appears that

scientific jury selection has been utilized, for the most part, by

defense attorneys who tend to equate freedom from bias as a tendency

toward acquittal and bias as a vote for conviction.

Based on these criticisms, it would appear that if a scientific

method is to be used it must be an in-court procedure which is less

intrusive than most scientific techniques, inexpensive, easy to use,

and of benefit to both the defense and the prosecution. Mahoney

(1982) suggested that an efficient screening device which can meet

these criteria and identify prospective jurors with extreme verdict-

relevant biases could be useful. Boehm's (1968) Legal Attitudes

Questionnaire, which was discussed earlier, would seem appropriate


Of particular interest to the present study, a field study

("Juror Bias," 1980) was conducted to determine the LAQ's efficacy in

identifying biased jurors in criminal trials. The subjects were

actual jurors from 18 felony trials (a total of 117 out of 216

possible jurors participated). These subjects completed the LAQ and

responded to a posttrial interview (2 weeks posttrial). A control

group of 70 subjects, who were called for jury duty but did not serve,

also completed the LAQ and showed no significant differences as

compared to the sample jurors on authoritarianism.

The results of the field study showed that jurors classified as

authoritarian were found to admit to bias against the defendant for

not taking the stand, reported the belief that defendants in general

were seldom or never innocent, and favored conviction regardless of

the possible sentence significantly more often than either

antiauthoritarians or egalitarians. Authoritarians also showed a

trend toward favoring conviction after only the opening arguments and

tended to overrate the prosecution's performance or underrate the

defense's performance. Authoritarians were found to report the

presence of multiple conviction-prone biases (i.e., those biases just

mentioned more often than antiauthoritarians or egalitarians).

Indeed, authoritarians were twice as likely to reveal more than two of

the biases and four times as likely to reveal more than three biases.

The fact that an expected pattern emerged for each measure of

bias and across 18 different criminal trials clearly suggests that the

jurors whose LAQ responses predicted biases were those most often

displaying actual bias. The results are especially encouraging

considering the many factors and wide variety of biases which could

have masked the pattern of bias predicted by the LAQ. The study's

obvious limitation, however, is its retrospective nature which assumes

that the subjects' recall of their feelings and thoughts during the

trial were similar to their actual thoughts and feelings at that

time. It also assumes that the subjects' recall was not influenced by

the fact that they had already deliberated. These are questions that

can be answered empirically.

To summarize, available research suggests that traditional

methods of conducting jury selection may be ineffective. It also

suggests that scientific jury selection methods have yet to be

sufficiently subjected to empirical test as well as presenting several

ethical problems. The Legal Attitudes Questionnaire (Boehm, 1968)

appears to show promise in addressing these issues. It is important

to mention that several studies which yielded favorable results

regarding the LAQ (Alexander & Licker, 1975; Boehm, 1968; Buckhout et

al., 1979; "Juror Bias," 1980) have recognized that its utility is as

a supplement to voir dire questioning, not as a substitute. To draw

an analogy, the trial lawyer may be able to employ the LAQ effectively

to aid in the decision process involving the use of peremptory

challenges just as the psychologist uses test data to make decisions

regarding classification, treatment, and prognosis. Both require

experience and skill in effective interviewing and interpretation of

test results.

The Evidence

From the literature reviewed thus far, it is clear that the

available voir dire methods are not without fault and are not perfect

at the removal of all prospective jurors with verdict-relevant

biases. Therefore, juries are likely to have some members with

verdict-relevant biases. In any case, juries are supposed to base

their decisions on the evidence presented. A brief review of the

literature will be presented here concerning the effects of the

presentation of evidence on jury decisions, with particular emphasis

on how this interacts with verdict-relevant biases.

Recent publications (M.F. Kaplan & Miller, 1978; M.F. Kaplan &

Schersching, 1980) presented research addressing the question of how

biases interact with evidential information in producing decisions.

This was done within the framework of Information Integration Theory

(Anderson, 1974; M.F. Kaplan, 1977a). Briefly, the theory, as it

applies to jury behavior, states that jurors employ a cognitive system

of weighting the values of different pieces of information related to

the trial (factual or otherwise). As mentioned earlier in this

review, the effects of juror biases are often apparent and least

likely to be confounded before the presentation of evidence (i.e.,

initial impressions and impressions after the opening arguments). The

theory hypothesizes that reducing the weight attached to biasing

information (e.g., initial impressions, verdict-relevant attitudes,

inadmissible evidence, pretrial publicity, coerced confessions) or

increasing the weight of evidential information will serve to reduce

juror bias.

A series of experiments were performed. In the first experiment

mock jurors with either lenient or harsh attitudes toward the

punishment of criminals (a concept similar to the dimension of

authoritarianism) rated defendants' guilt and severity of punishment

for summaries of traffic felony cases, in which the evidence had a

high or low appearance of guilt. The subjects were either led to

believe the source of the evidence was reliable and trustworthy or

unreliable. Harsh subjects gave higher guilt ratings than lenient

subjects in cases across both levels of apparent guilt. However, in

both conditions of guilt, harsh and lenient mock jurors gave

essentially the same guilt ratings when the evidence was characterized

as reliable and trustworthy. When nothing was said about the evidence

or when it was questioned, harsh jurors remained severe relative to

lenient jurors.

The second experiment was similar except that mock jurors were

told that either the evidence was inconsistent and questioned by the

attorneys or that, although the evidence was inconsistent, it was

considered by the attorneys to be true and accurate. When the

evidence was characterized as arguable, harsh jurors showed higher

guilt ratings than lenient ones. Again, convergence of guilt ratings

was noted when the evidence was characterized as not contested.

Additionally, it was shown that lenient jurors were more likely to

discount incriminating evidence as influencing their decision if they

were told that the evidence was contestable. This parallels the

findings using guilt ratings.

These findings support the hypothesis that increasing the weight

of the evidential information serves to reduce the observed effects of

biases. Similarly, Hepburn (1980) found that perceived strength of

evidence is the most important predictor of verdict.

Aside from not having jurors deliberate, a major flaw of these

studies was that the presentation of evidence by written case

summaries did not allow the authors to clarify how various parameters

concerning the presentation of evidence might influence the weight

attached to it. Procedure, medium of presentation, order of

presentation, and eyewitness testimony will be the parameters

discussed here.

Procedure would appear to have an influence on the perception and

evaluation of evidence. Here, there are two types of methods for

presenting evidence: (a) the adversary system has the two opposing

sides present and dispute the evidence, and (b) the inquisitorial

system has the judge ascertain evidence from both sides. The issue of

which of these systems are more likely to help reduce juror biases has

yet to be resolved. While the inquisitorial system has the

theoretical advantage of presenting evidence that is more homogeneous

and appears less contentious (Lind, Thibaut, & Walker, 1973), it has

been shown that mock jurors who are biased toward conviction show more

extreme beliefs in the defendant's guilt in the inquisitorial method

of evidence presentation (Thibaut, Walker, & Lind, 1972).

Additionally, studies have shown that impartial observers perceived

the adversarial system as being more fair and jurors within this

system were more satisfied with the rendered verdict (Walker, La Tour,

Lind, & Thibaut, 1974). Thibaut and Walker (1975) also reported that

trial participants favored the adversarial system because it provided

a means of influencing the presentation of evidence. This finding was

criticized by Sheppard and Vidmar (1980), who found that lawyers may

actually induce biases in witnesses during interviews in the

adversarial system. More research is clearly needed to determine how

the procedure of evidence presentation influences the weight jurors

attach to the evidence and biased impressions.

The medium of presentation of evidence is another parameter which

may influence the differential attachment of weights to evidential and

biasing information. Specifically, the use of videotaped trial

materials in the resolution of civil disputes has been of interest.

G.R. Miller (1976) reported a series of studies with actual

jurors that found no differences between live and videotaped civil

disputes in attribution of negligence, awarding of damages, perception

of attorney credibility, retention of trial-related information, and

interest in the trial proceedings. However, a comparison of live,

monochromatic, and color videotaped trials revealed that over time

jurors viewing the monochromatic videotape showed better retention of

trial-related information than those who viewed the color videotape or

a live presentation. Finally, comparing color versus monochromatic

videotape presentations, it was reported that more trial-related

information was recalled by jurors who viewed monochromatic tapes but

attorneys received higher credibility ratings by jurors who viewed

color tapes.

Fanner et al. (1976) reanalyzed some of the above data and found

that jurors in videotape conditions were more likely to find the

defendant negligent than those in live presentation conditions. They

also found that videotape presentation might influence the jurors'

evaluations of the witnesses.

More research is needed in this area before firm conclusions can

be drawn. Certainly, the data presented by G.R. Miller (1976) and

Farmer et al. (1976) would tentatively suggest that the mode of

presentation of evidence might be an important factor in influencing,

at least, civil dispute-relevant decisions.

The order of presentation of evidence has received attention in

the literature. The idea that primacy or recency effects might be

relevant in the presentation of evidence originates from the

experimental literature where it was shown that such phenomena occur

on free-recall tasks (Murdock, 1962). The courtroom procedure,

however, is more complicated than simple, free-recall tasks. This

might account for the equivocal results found in the literature

(Insko, 1964; Lund, 1925; Pennington, 1982; Thibaut et al., 1972).

A final parameter relevant to the presentation of evidence is

perhaps the most popular current area of psycholegal research.

Evidence in criminal trials is often presented by one or several

eyewitnesses. Recent literature reviews (Loftus, 1979a; Penrod,

Loftus, & Winkler, 1982) have concluded that eyewitnesses are likely

to be inaccurate much of the time. Despite this, their influence on

verdict-relevant decisions of jurors is important. For example,

Loftus (1974) reported that, in a simulated case, the use of an

eyewitness resulted in 72% of the jurors favoring conviction. Without

an eyewitness, only 18% of jurors voted guilty. Discrediting the

witness here did not have a significant effect; however, other studies

have found that discrediting the eyewitness results in jurors ignoring

the witness's testimony (Hatvany & Strack, 1980; Weinberg & Baron,


A few studies have been conducted to clarify what factors in

eyewitness testimony are important in influencing jurors. Witnesses,

who are considered experts (e.g., physicians), have been found to have

the greatest testimonial impact and are seen as being most credible

(Ludwig & Fontaine, 1978; Sonaike, 1978). Indeed, Loftus (1980) has

shown that the use of expert witnesses, who have testified on the

unreliability of eyewitness testimony, results in reduced conviction

rates. Ludwig and Fontaine (1978) reported that opinionated eye-

witness testimony had a significant influence on jurors in cases with

more serious charges, such as first and second degree murder. Wells,

Lindsay, and Ferguson (1979) showed that juror perceptions of

eyewitness confidence greatly enhanced their credibility even though

eyewitness confidence is not associated with eyewitness' accuracy.

These studies show that eyewitness testimony, regardless of accuracy,

probably has an impact on verdict-relevant decisions.

That eyewitness testimony might lead to biased decisions has been

demonstrated by Hatton, Snortum, and Oskamp (1971). Two groups of

jurors viewed the same film of an accident, but received different

biasing eyewitness testimony. Although no differences in information

recalled from the film were found, jurors were shown to make

inferences congruent with the biased testimony they heard.

The presentation above illustrated that several parameters of the

mode of presentation of evidence--adversarial or inquisitorial

systems, videotape or live presentations, primacy-recency effects, and

eyewitness testimony--might influence the differential weights

attached to evidential and nonevidential information, which, in turn,

might have an impact on verdict-relevant decisions.

The type of evidence presented, especially that which may be

biasing, is also an important factor to consider in this context.

Inadmissible evidence comes to mind first. For example, Sue, Smith,

and Caldwell (1973) reported that jurors were only influenced by

inadmissible evidence when the trial evidence was perceived as weak.

Regarding verdict, Thompson, Fong, and Rosenhan (1981) showed that

inadmissible proacquittal evidence was more likely to have an impact

in the expected direction (i.e., toward acquittal) on juror verdict

than proconviction inadmissible evidence or no inadmissible evidence.

Prejudicial pretrial publicity may be considered a subset of

inadmissible evidence. Mock jury research has shown pretrial

publicity to influence individual juror decisions (Hoiberg & Stires,

1973; Sue, Smith, & Gilbert, 1974). A more authentic study

(Pawader-Singer & Barton, 1975) used subjects from actual jury pools

and found that exposure to negatively biased pretrial publicity

resulted in juries being more likely to convict. Loftus (1979a)

showed that jurors exposed to advertisements sympathetic to insurance

companies were more likely to award less money in damages.

Most recently, Freundlich (1983) showed that, although low ego

level jurors tend toward conviction when presented with negative

pretrial publicity, high ego level jurors were more likely to

overcompensate and render more acquittals. Therefore, although

pretrial publicity shows a biasing impact on verdict-relevant

decision, it appears that ego level (and perhaps other personality and

attitudinal variables) might mediate this effect.

The influence of coerced confessions on jurors was investigated

by Kassin and Wrightsman (1980, 1981). Here, jurors did not regard

evidence which was obtained through coercion but incorporated evidence

obtained through favorable inducement into their decision-making.

This would certainly appear to bear on juror decisions concerning the

weight of evidence.

Finally, temporarily biased conditions in the courtroom have been

studied. M.F. Kaplan and Miller (1978, experiment 3) showed that

obnoxious and annoying behavior in the courtroom resulted in biasing

jurors more toward conviction. Additionally, the closer the source of

these courtroom annoyances to the defendant (i.e., the defense

attorney), the greater the effect of bias.

To summarize, integration information theory (Anderson, 1974;

M.F. Kaplan, 1977a) has been proposed as a framework with which to

conceptualize the decision-making process of the juror. Research

consistent with this theory suggests that an increase in the weight

attached to evidential information will often offset the impact of

verdict-relevant biases. However, several parameters such as the type

of procedure, medium of presentation, order of presentation, type of

evidence (e.g., credibility of eyewitnesses), and transient features

of the courtroom would appear to influence the weight of evidential

and nonevidential information and impressions. It is certainly

difficult to control for these variables in applied studies in order

to see if these findings generalize. Even simulated studies can

usually only control or manipulate one or two of these variables.

What is important, however, is that findings be interpreted within the

context of available research concerning the effects of evidence on

jury decision-making.

Judge's Instructions to the Jurors

After the evidence is completely presented, the judge instructs

the jurors on how the law applies to the factual issues of the case.

The purpose of these instructions is to provide the jurors with a set

of rules designed to guide their decisions within the legal

framework. That jurors view these instructions as important is

illustrated by Kalven and Zeisel (1966), who reported that the more

difficult the case being tried the more likely juries request more

instructions from the judge. Instructions are also believed to

facilitate impartiality (Hastie, Penrod, & Pennington, 1983). There

is very little research which has investigated this.

Two studies (Charrow & Charrow, 1979; Strawn & Buchanan, 1976)

have reported that many jurors have difficulty understanding the

instructions because they are often termed in legal jargon. Monohan

and Loftus (1982) reported a study by Elwork and his associates which

showed that clearly written instructions improved comprehension.

The majority of the simulation research on the efficacy of

instructions to jurors in reducing extra-evidential biases concerns

the instruction to disregard inadmissible evidence. Doob and

Kirshenbaum (1973) found that mock jurors were likely to judge a

defendant guilty when given negatively biasing information about the

defendant's past history regardless of whether they received

instructions from the judge to ignore that information. Kerr et al.

(1976) reported that jurors were more likely to follow instructions if

they concern legal criteria in contrast to the disregarding of

information. Recall that the findings of Thompson et al. (1981)

showed jurors receiving inadmissible proacquittal evidence were more

lenient. They also found that instructions to disregard this evidence

had no effect. Similar findings have been reported by Fontes, Miller,

and Bender (1977). Perhaps jurors consider inadmissible evidence as

containing truthful information (Kadish & Kadish, 1971).

Of greater concern is that there is evidence to show that

instructions to the juror may actually result in effects opposite of

its intention. Wolf and Montgomery (1977) have identified what they

call the "boomerang effect." They found that stronger warnings from

the judge to disregard evidence resulted in an increased weighting of

that information by the jurors. It is especially relevant to this

study that Werner et al. (1982) found this effect to be particularly

prevalent in authoritarian jurors. Borgida (1979) proposed that the

more explicit or strong the judicial instruction, the more it is

perceived by jurors to threaten their decision freedom, resulting in a

reactance response.

In addition to being instructed as to which evidence and

information to consider, jurors are instructed as to the types of

verdicts they may render. Here, both Vidmar (1972) and K.J. Kaplan

and Simon (1972) showed that when charges were severe (e.g., first

degree murder) restricted decision alternatives (i.e., guilty vs. not

guilty) often resulted in increased chances of obtaining a not guilty

verdict. Increasing the choice of three or four alternatives resulted

in a reduction of not guilty verdicts. Most relevant to the present

study, Buckhout et al. (1979) using the LAQ showed that, when jurors

were given a choice of guilty (of first degree murder) or not guilty,

45% of authoritarians voted guilty and 33% of the egalitarians voted

the same. When five alternative verdicts were offered, however, there

was no difference between these groups of juries in verdict decision.

To summarize, jurors are given instructions providing decision

rules with reference to factual issues such as the admissibility of

evidence and the type of verdicts to be rendered. Research suggests

that instructions to disregard evidence have either been ineffective

or have resulted in opposite effects. However, it appears that

increasing the number of decision alternatives, at least in trials

concerning heinious crimes, may reduce verdict-relevant biases.


Through careful selection of jurors, presentation of a strong

case, and provision of a legal framework within which to make

decisions, the legal system attempts to insure that the jury will

reach an impartial decision based on legally acceptable evidence.

Until the point that deliberation begins, however, these safeguards

only operate on the level of the individual juror. The factors

influencing the individual jurors' decisions have been discussed. The

question now becomes: "Does the deliberation process influence

individual juror's decisions?" Phrased in a manner more consistent

with the present study: "Through what processes does deliberation

help to reduce the effects of individual juror biases or preferences

on the trial outcome?" These appear to be important questions given

that deliberation is also supposed to serve as a safeguard for the

justice system and, that once the jury retires to deliberate, they, in

essense, become justice itself.

Process. The process of jury deliberation can refer to any

behaviors that occur during the group's interaction, including rate

and style of participation, verbal behaviors, nonverbal behaviors

(e.g., gestures, facial expressions), and cognitive events occurring

within the individual juror (e.g., opinion changes, interpersonal

attraction). Research has been conducted almost exclusively on the

verbal communication processes during jury deliberation. Here, the

question is "Who says what to whom in what manner?"

One topic, that has been studied using recorded verbal

communications during deliberation, has been rate of participation

among jurors. Research on 12-person juries suggests that in most

juries more than one-half of the total verbal acts are monopolized by

relatively few jurors (Hastie, Penrod, & Pennington, 1983; Simon,

1967; Strodtbeck et al., 1957). These studies also report that the

foremen are responsible for between one-fourth and one-third of all

verbal acts. With respect to nonforeman, jurors with high

occupational status and high education tend to participate more

actively than their lower educated, lower occupational status

counterparts (Hawkins, 1962; James, 1959; Simon, 1967; Strodtbeck et

al., 1957). Indeed, there are data to suggest that these jurors are

more likely to be perceived by their fellow jurors as being more

influential. Additionally, Hastie et al. (1983) showed that the size

of the faction to which the juror belongs is positively associated

with the probability that the individual juror would speak.

Unequal participation in deliberation counters the philosophy of

equal representation in the jury. However, a study by Foss (1976b)

suggested that the number of people who participate in the

deliberation may not be as important as the issue of whether all

points of view are expressed. Using college students as mock jurors,

in juries ranging from five to seven people, and studying the

recordings of their deliberations, he found that juries made their

decisions through an equalitariann" process. That is, each juror's

position was given fair hearing in the deliberation, with pro and con

arguments made for all supported positions by at least one group

member. Foss interpreted this finding as indicating that the

deliberation process prevents a decision based on extra-legal juror

biases unless similar biases are adhered to by all or an overwhelming

majority of the jury members. This suggests that a jury composed of

members with a mixture of biases will render an "impartial" decision.

Another area of investigation in deliberation processes is at

whom comments and arguments are directed. The paucity of research on

the target of jurors' comments is not surprising given the difficulty

in identifying the target based on audiotape recordings. The

literature suggests a pattern in that those who tend to speak more are

likely to be spoken to more often. In most cases the target of juror

communications have been the foreman (James, 1959) and men (Nemeth,

Endicott, & Wachtler, 1976). Other patterns of communication have

been investigated as a means of studying this area. Hawkins (1962)

studied the sequence of channeling of communication in the

deliberating jury. Dividing the deliberation process in four phases,

Hawkins reported that the proportion of communications between members

of opposing factors rose steadily through the first three phases,

while comments directed toward the members of a juror's own faction

and toward the whole group simultaneously dropped. During the final

phase, when the outcome appeared decided, communication between

factors dropped greatly while the amount of communication toward the

whole group increased considerably. Other studies have found similar

patterns (Levin, Farrell, & Perotta, 1981; Simon, 1967). These

patterns confirmed models of the sequence of stages of small group

discussion, which propose increasing conflict between members until

resolution occurs followed by a reconciliation, solidarity, and

commitment to the final decision (Bales & Strodtbeck, 1951).

Foss (1976a) obtained a pattern in deliberations in which the

early phases were characterized by various jurors contributing new

pieces of information which had been missed or forgotten by other

jurors and presenting their own interpretations of the evidence.

During later phases of deliberation, pressures to conform were

observed to be predominant. This pattern would suggest that early

phases of deliberation consist mainly of an exchange of views between

active parties; whereas, later phases may consist of normative social

influence processes so that a group consensus can be reached. The

findings of Hastie et al. (1983) support this explanation. Holstein

(1985) also reported similar findings in that the number of

interpretations offered by jurors in mock deliberations decreased as

the deliberation progressed.

The deliberation process appears to be an anxiety-producing

experience for most jurors (Goldman & Casey, 1980). Therefore, in

addition to investigating who participates and the patterns of

communication in jury deliberations, authors have also studied the

style of jury interactions during deliberation. The Bales Interaction

Process Analysis (Bales, 1951) coding scheme has been most widely used

in this area of research. Here, James (1959) found that it was more

common for a juror to express ideas and opinions more often than agree

with another juror. In contrast, Goldman et al. (1983) showed that

jurors most commonly sought agreement and support from each other.

Perhaps future research will suggest that different styles of

interactions may characterize different phases of the deliberation.

Several findings have indicated that style of interaction may be

influenced by juror characteristics. Here, Strodtbeck and Mann (1956)

employed Bales' system to assess differences in interaction style with

respect to sex. Males were reported as more likely than females to

express the categories "gives information" and "gives opinion";

whereas females were more likely than males to devote their comments

to the categories "shows solidarity," "shows tension release," and

"agrees" (i.e., the metacategory of "positive reactions"). This sex

role differentiation in interactional styles of jury deliberations has

been replicated (Nemeth et al., 1976; Piliavin & Martin, 1978).

Goldman et al. (1983) corroborated these results by finding that male

juries had a higher rate of disagreements than female or mixed-sexed

juries. Additionally, Piliavin and Martin (1978) found that sex role

differentiation was strongest in same-sex groups; mixed-sex groups

tended to exhibit an attenuation of the existence of sex role

differentiation. Mixed-sex juries might then be one means by which to

increase the probability of equal participation and representation.

Nonetheless, it remains to be determined empirically whether equal

participation leads to significantly different jury decisions.

With respect to jurors' education, James (1959) reported that

jurors with no more than a grammar school education tended to offer

comments that were less accurate and more disruptive to the

deliberation process; focused more of their comments on testimony,

personal experiences, and opinions of the trial; and were more likely

to passively accept another's view than were their higher educated


Moral judgment level has also been investigated. Recall the

finding of Goldman and Casey (1980) that the deliberation process

appears to be an anxiety-arousing experience. Here, it was

additionally found that jurors with lower levels of moral judgment

were more likely to demonstrate heightened levels of affect.

Recently, Freundlich (1983) showed that ego level, a measure of

psychological sophistication, is related to deliberation style. He

found that juries composed of mock jurors with high levels of ego

development, who received incriminating pretrial publicity,

demonstrated comments indicating low solidarity (characterized by

discord and conflict), few agreements, and few positive reactions. In

contrast, juries composed of low ego level jurors, receiving the same

pretrial information, exhibited higher rates of solidarity, agreement,

and positive reactions yet were less efficient as noted by the

presence of decision problems. These findings suggest that juries

comprised of jurors who are psychologically sophisticated, although

generating more conflict, also are more efficient at decision making

and more apt to overcome biases. More research is needed here to

further clarify the role of ego level in jury decision making.

Of greater relevance to the present study, Freundlich (1983) also

found a nonsignificant trend in high authoritarian jurors who

responded in a similar manner to those juries which were comprised of

low ego level jurors. Since these juries were not organized according

to authoritarian attitudes and the findings were not significant,

further research is needed to address the issue of whether the

authoritarian dimension has an influence on deliberation style.

The research reviewed above suggested that several

characteristics of jurors may influence the style of interaction

during the deliberation process. However, a more precise role of

these variables must be clarified by further research. Regardless of

these findings, process is only one important aspect of

deliberation. The content of what is said must also be examined.

Content. Not surprisingly, relatively little research has been

devoted to the content of deliberations, given the difficulty in

obtaining access to deliberations. James (1959) reported that juries

devoted very little time (15%) to actual testimony. Most of the time

(50%) was found to be spent discussing personal experiences and

opinions or procedural issues (25%). These findings would appear to

contradict the notion of the jury as a task-oriented group; however,

James showed that the majority of the jurors' comments were usually of

high quality. In other words, references to testimony were rated to

be accurate approximately 80% of the time; opinions and personal

experiences were rated as pertinent to the jury's task approximately

70% of the time; and procedural discussions were rated as being

facilitative approximately 70% of the time. Interestingly, she found

that 64% of the jurors' comments were neutral.

Kessler (1973) found results which supported James' data.

Lawyers and law students rated the content of mock jury deliberations

and found that 82% of the discussion to be legally relevant.

Additionally, jurors' comments were rated as not favoring either guilt

or innocence.

Hastie, Penrod, and Pennington (1983) formed sixty-nine 12-person

mock juries comprised of volunteers recruited from Superior Court jury

pools in three counties in Massachusetts. After viewing a 31/2 hour

videotape of a re-enactment of a murder trial, each of these juries

deliberated to a verdict. The contents of these deliberations were

examined closely and it was shown that 53% of the remarks made during

deliberation related to testimony and trial evidence and 25% of the

remarks were references to the judge's instructions about the

definitions of the law. They also rated these deliberations to

indicate that 75% of the remarks made were intended to communicate

information concerning the case from one juror to others. These data

suggested that jurors are indeed task-oriented for a good portion of

the deliberation. That juries monitor their own behavior was evident

from the fact that 40% of the expressed erroneous interpretations of

the judge's instructions were corrected in the deliberation.

Although not formally rating the content of deliberation, Goldman

et al. reported that their mock juries appear mostly "concerned with

details of the evidence, legal definitions, reliability and

credibility of the witnesses and evidence, and consequences of

verdicts" (1975, p. 376).

M. Levine, Farrell, and Perotta (1981) employed a content

analysis system relevant to group development (Farrell, 1976) which

was a modified form of the Bales (1951) interactional system.

Comparing the jury to an unstructured group, they found that juries

exhibited less disclosure of personal information and expression of

feelings while spending a greater proportion of time working and

clarifying the rules guiding deliberation.

In a very recent investigation, Holstein (1985) looked

specifically at the frequency with which juries make interpretations

of evidence and testimony (i.e., when a juror verbalized a formulation

of "what really happened" in the case to the jury). Here, jurors were

recruited who had recently served two months on the jury panels of the

Washtenaw County, Michigan District, or Superior Courts. They formed

forty-eight 5- to 6-person mock juries which viewed a videotape of a

theft case and then deliberated to a unanimous verdict. Holstein

found that at least one interpretation of "what really happened" was

rendered during each of the 48 mock deliberations and that jurors

presented more than one interpretation in 75% of the mock

deliberations. Indeed out of the 48 mock deliberations studied, 15

different interpretations were identified. These findings would

corroborate Foss's (1976b) conclusions that several jurors' positions

are given a fair hearing during deliberation.

The research on process and content has suggested that

deliberation is a complex and dynamic process. Different jurors

appear to participate unequally and may be more influential,

particularly foremen, members with higher educational and occupational

status, and men. Nevertheless, the jury follows the stages and

operates as a task-oriented group adhering to instructions provided by

the court and focusing on decision-related contents. However, process

and content are only important as they relate to the final jury


Product. Doob (1976) illustrated that although juries appear to

operate "within the etiquette" of the law, extra-legal biases can

still be influential toward the final decisions. Therefore, this

discussion will proceed from purely descriptive aspects of

deliberation to what processes and contents might influence the

product or outcome deliberation, with specific attention devoted to

the role of verdict-relevant biases (i.e., the authoritarian


Kalven and Zeisel (1966) were skeptical of the influence of

deliberation on the trial outcome. Recall their study which

reconstructed first ballot votes from posttrial interviews of actual

jurors. They reported that the first ballots were predictive of

actual jury verdicts 90% of the time and suggested, "The deliberation

process might well be likened to what the developer does for an

exposed film: it brings out the picture but the outcome is

predetermined" (Kalven & Zeisel, 1966, p. 489). Their data also

suggested that the majority of the initial opinions were extremely

predictive of trial outcome.

Nemeth (1977) provided support for this majority effect. By

keeping a running total of the number of comments supporting each

decision choice, she found that once the difference between totals for

guilty versus not guilty exceeded seven, the alternative with the

higher talley predicted the juries' final decision in 36 of 37

cases. Furthermore, this critical difference was observed early in

the deliberation. However, the question of whether the majority

effect is the result of the actual number of supporting arguments or

the position which has the most supporters has yet to have been

answered. M.F. Kaplan (1977b) has provided tentative evidence that

the number of unique and nonredundant arguments is the critical

variable here. Hastie et al. (1983) showed that majority factions are

more confident in their decisions and therefore might be more

persuasive. The implication of these findings is that the

deliberation process is ineffective at reducing initial biases,

particularly if those holding such biases are in the majority.

Furthermore, they would suggest that the effects of bias on individual

juror decisions will generalize to the decisions of juries.

More disheartening has been a body of research which suggests

that deliberations usually have the effect of exaggerating the

individual's predeliberation opinions and biases. This has been

referred to as the group polarization phenomenon (Myers & Lamm,

1976). Simply stated, "The average postgroup response will tend to be

more extreme in the same direction as the average of the pregroup

responses" (Myers & Lamm, 1976, p. 603). In other words, if

individual jurors were leaning toward a not guilty verdict, the

deliberation would increase their commitment to a not guilty verdict

even more. Similarly, if individual jurors tended to believe a

defendant to be guilty, deliberation would result in a stronger

commitment to a guilty verdict.

Myers and Kaplan (1976) measured mock jurors' responses before

and after deliberation on eight abstracts of actual traffic felony

cases. As predicted, after discussing cases low in incrimination

value, subjects were more extreme in their judgments of innocence and

more lenient in recommended punishment; after discussing cases high in

incrimination value, subjects shifted toward harsher judgments of

guilt and punishment. Thus, prediscussion preferences (i.e.,

incrimination value of evidence) resulted in polarization of verdict


The group polarization phenomenon has been investigated in jurors

assessed along the authoritarian dimension. These studies are

relevant to the present investigation and will be discussed here.

Vidmar (1972b) divided a group of mock jurors into high and low

dogmatism juries. The dogmatism juries were found to shift toward

harsher sentences and the low dogmatism juries shifted toward more

lenient sentences, despite the fact that both types of juries did not

differ in their predeliberation judgments. It is possible here that,

although biases were hidden, they were still influential.

The study of Buckhout et al. (1979) employed the LAQ to divide

mock jurors into egalitarian, authoritarian, and mixed juries. They

found that "pure" juries shifted more after deliberation in the

predicted directions than did "mixed" juries. Furthermore, "pure"

egalitarian juries shifted more toward leniency than other juries.

These results suggest that one way to reduce the polarization of

verdict-relevant biases is to either have a nonbiased jury or to at

least insure that different biases are equally representative.

Bray and Noble (1978) used the F-Scale (Byrne, 1974) to classify

mock jurors as high or low authoritarians. This study yielded several

results relevant to the effects of deliberation on authoritarian

biases. After listening to an audiotape presentation of a murder

trial, the juries deliberated. Pre- and postassessments of guilt and

sentencing showed that low authoritarians were more lenient after

deliberation than before and high authoritarians were more severe

after deliberation than before. These findings again support the

notion that juries moved in the direction of choice tendencies

initially favored.

Although there are considerable data supporting a predictable

polarization phenomenon after group deliberation, this is not always

the case. Indeed, Kalven and Zeisel (1966) found that, in 9% of their

cases in which there was a proacquittal majority, the juries actually

convicted or were hung. Additionally, 14% of the cases which had

proconviction majorities acquitted or were hung. Nemeth (1977)

explained the differences between these percentages by postulating

that it is easier for a minority to persuade other members of the jury

that they might have a reasonable doubt than to convince them beyond

such a doubt as in a proacquittal majority. She also hypothesized

that minorities holding a position of not guilty are less likely to

give in, which she reported is an important variable in swaying a

majority. She postulated no explanation for the former finding.

Although his findings were not in the same direction as those of

Kalven and Zeisel (1966), Holstein (1985) provided data suggesting

that group polarization or majority effects do not necessarily occur

after deliberation. In the 48 mock deliberations studied, 11 out of

15 interpretations rendered favored not guilty verdicts; however, only

8% of the juries actually acquitted and 45% convicted the defendant.

Holstein also found that 38% of the juries were hung and that this was

positively related to the number of alternative interpretations

postulated during the deliberation.

There are also other data which suggest that deliberation may

serve to reduce the influences of bias. For example, Izzet and

Leginski (1974) showed that deliberation served to reduce a

predeliberation attraction leniency bias. The study of M.F. Kaplan

and Miller (1978, experiment 3), which showed that annoying behavior

in the courtroom resulted in biasing jurors more toward conviction,

found that the biasing effects of the courtroom virtually disappeared

after deliberation. Finally, Carretta and Moreland (1983), using

college students serving on 6-person mock juries who had read a

written summary of a murder trial, found that although inadmissible

evidence affected the jurors' pre- and postdeliberation estimate of

the defendant's guilt, it had no effect on postdeliberation verdict.

Furthermore, it was found that when one group member would mention

inadmissible evidence, there was an 84% chance that another juror

would respond by either reminding the group that such evidence was

inadmissible or by reiterating a trial argument from the opposing

view. This finding suggested that jurors "monitored" one another

during the deliberations to minimize the effects of this biasing


With regard to the authoritarian dimension, Alexander and Licker

(1975) showed that egalitarian jurors who favored a predeliberation

guilty verdict shifted more toward leniency during deliberation when

faced with arguments which vigorously favored acquittal. These

authors, as well as others (Boehm, 1968; Buckhout et al., 1979),

argued that, as expected, authoritarians were more resistant to change

and that egalitarians were more persuadable. However, Bray and Noble

(1978) found that high authoritarian jurors changed their verdicts

from predeliberation to postdeliberation in both directions more often

than low authoritarians. Similarly, Lamberth et al. (1982) found that

authoritarians changed their verdicts after deliberation from guilty

to not guilty. Clarification of this issue of influenceability awaits

further research.

To summarize, how can one account for the various and apparently

contradictory findings with respect to how the deliberation process

influences the outcome of the trial? It has been shown that in many

cases not only does the preference of the majority of the jurors

prevail, but that these preferences are often exaggerated through the

processes of group polarization or conformity influences. In

contrast, some aspects of the deliberation process appear to reduce

initial preferences or biases and may even result in reversals from

predeliberation preferences. These reassuring findings indicated that

deliberations may be task-oriented processes where jurors share

information and alternative interpretations of case material and

definitions of the law and monitor each others' responses to reduce

the influences of extra-legal juror biases.

Information Integration Theory (IIT) (Anderson, 1974; M.F.

Kaplan, 1977a) has been proposed as a theoretical model which can

account for these disparate findings (M.F. Kaplan, 1977b; M.F. Kaplan

& Miller, 1977). In this model, each piece of information available

for a judgment has two properties: (a) the scale value which is the

position on a dimension judgment or valence (i.e., degree of innocence

or guilt), and (b) the weight which is the importance of that piece

of information for the judgment. For each piece of information (e.g.,

evidence, initial impression) of which the individual is made aware, a

scale value is assigned and through an averaging of the given weights

(i.e., "cognitive algebra") the information is integrated into a


Briefly, IIT's explanation focuses on the deliberation as

affecting decisions through information influences (i.e., information

sharing, persuasive argumentation) and normative influences (i.e.,

pressure to conform to the group's norms).1 Data have been provided

which suggest that when jurors are provided with nonredundant

information of scale values similar to their own or information more

extreme in scale value from their neutral initial impressions, the

averaging of weights will lead to a polarized final decision (M.F.

Kaplan, 1977a). This would also seem to explain the majority

effect. If during deliberation jurors are provided with information

of opposite scale value, through differential recall for example, the

averaging of weights will lead to a more moderate decision, a

reduction of initial preferences or biases, or even a reversal of

predeliberation decision (M.F. Kaplan, 1977b; M.F. Kaplan & Miller,

1977). For example, recall that juries comprised of both egalitarians

and authoritarians were less likely to exhibit polarization than
"pure" juries (Buckhout et al., 1979). It is possible that these

juries contained members that recalled different trial information

(Berg & Vidmar, 1975) with different scale values, and in which the

averaging of weights resulted in a more moderate decision.

Of course, this can only be considered tentative speculation

since the studies designed to support IIT only have manipulated the

evidence presented to the juries and have not considered the

possibility that verdict-relevant biases may carry a heavily weighted

1 The information influences have been given greater support by the
available data (M.F. Kaplan, 1977b; M.F. Kaplan & Miller, 1977;
Myers & Lamm, 1976).

informational scale value. The utility of this theory, therefore,

remains to be investigated.


The literature reviewed in the previous discussion has suggested

that biasing factors distort verdict-relevant judgment even in the

context of a procedure which attempts to limit and counteract their

influences. Authoritarian jurors and juries have consistently

rendered harsher decisions. Research has also tentatively suggested

that egalitarian jurors and juries might be more biased toward

leniency (Buckhout et al., 1979). Clarification of this issue is

certainly warranted since authoritarians have been viewed as biased

and egalitarians as nonbiased (Mahoney, 1982).

The questionable efficacy of traditional voir dire methods have

led several authors (Boehm, 1968; "Juror Bias," 1980; Mahoney, 1982)

to suggest methods which might improve the effectiveness of the voir

dire procedure at removing jurors with extreme verdict-relevant

biases. A verdict-relevant measure which assesses the authoritarian

attitude dimension, the Legal Attitudes Questionnaire (Boehm, 1968),

has been suggested for its additional advantages of being a relatively

nonintrusive in-court procedure, which is inexpensive, easy to use,

and may benefit both the defense and prosecution. This is worthy of

investigation since the selection of a jury comprised of citizens that

do not hold extreme biases and whom will render a fair and impartial

decision appears most desirable. This might increase the probability

that the presentation of legally acceptable evidence, judge's

instructions to the jury, and deliberation would be likely to foster

impartial justice.

An emerging theme in research on the effects of juror biases on

trial outcome is the investigation of the processes during the trial

which might account for observed differences. The literature has

suggested that differences in transient affective states (Freundlich,

1983; Freundlich & Goldman, 1982; Goldman et al., 1975), style of

deliberation (Freundlich, 1983; James, 1959; Kessler, 1973), and

recall of trial information (Berg & Vidmar, 1975; M.F. Kaplan &

Miller, 1977; Sherrod, 1985) are worthy of further investigation. The

influence of theoretically bias-reducing trial procedures (i.e.,

testimony, judge instructions, deliberation) must also be identified.

The effects of juror biases on process and outcome of trials may

be of interest to social scientists. However, for this research to be

legally relevant, it must be shown that biased jurors and juries

differ from nonbiased jurors and juries (Gerbasi et al., 1977).

Previous research has employed judge-jury disagreements as a means

with which to evaluate jury performance (Kalven & Zeisel, 1966).

However, others have suggested that it is more important to employ a

theoretically neutral group with which to compare biased groups (Saks

& Hastie, 1978).

The present investigation attempted to address these issues

employing a simulated trial paradigm. This allowed for the

experimental manipulation of jury composition (through an experimental

voir dire) and the use of a theoretically important control group.

Attempts were also made to maximize the genuineness of the trial

presentation, which included an audiovisual presentation of the trial

and the inclusion of all important trial procedures (i.e., opening

arguments, testimony, judge's instructions, and deliberation).

Other methodological and theoretical questions were also

addressed here. It has been popular for field and applied jury

research to exclude jury deliberations and/or to rely on posttrial

interviews (often more than two weeks) to reconstruct the decision

processes of the jury ("Juror Bias," 1980; Kalven & Zeisel, 1966;

Zeisel & Diamond, 1978). The research clearly suggests that

deliberation is important (Caretta & Moreland, 1983; Foss, 1976b;

Hastie et al., 1983; M.F. Kaplan & Miller, 1978). Also, given that

jurors with different biases recall different aspects of the trial

(Berg & Vidmar, 1975) and that memory may also be reconstructive and

unreliable (Loftus, 1979a; Penrod et al., 1982; Sherrod, 1985), it

appeared important to evaluate the accuracy of posttrial recall.

With regard to theoretical issues, the group polarization

phenomenon has been a popular finding in the literature. By

investigating deliberation processes, the factors which may contribute

to this phenomenon might be clarified. Additionally, this study

addressed how different juror biases might interact with the group

polarization phenomenon.

To summarize, there were several unique aspects of this

investigation. This study directly assessed the efficacy of the LAQ

as a voir dire supplement by deriving and comparing biased to

theoretically nonbiased jurors and juries. Its repeated measures

design allowed for it to clarify the effectiveness of different parts

of the trial at reducing biases. It examined the processes through

which biases operate or were reduced during the trial. The study also

determined the difficulties of not studying group deliberation and/or

relying on jurors' posttrial recall to reconstruct the decision

processes of the jurors during the trial.


The effects of verdict-relevant jury biases [i.e., authoritarian,

egalitarian, mixed composition (half authoritarian/half egalitarian),

and neutral] and sex of the jurors were assessed on verdict, a

weighted index of verdict, confidence in verdict decision, and recall

of trial information. The potentially moderating effects of

demographic variables and affective states on these relationships were

also examined. Several directional hypotheses were offered here. It

was predicted that authoritarian jurors and juries would be more prone

to convict, more confident in their verdicts, more likely to recall

information about the defendant's character and the prosecution

attorney's trial arguments, and less likely to recall situational

evidence than other jurors. Although with less confidence, it was

expected that if egalitarians respond in a less biased manner, they

might recall more situational evidence; but, if they tended to be more

prone to leniency, they might recall more of the defense attorney's

trial arguments. The comparisons between these jurors and juries to

the theoretically neutral jurors and juries as well as to the juries

with mixed compositions were of particular interest because of the

expected bias-reducing effects of these latter two types of juries.

Since the literature has yielded many equivocal and unclear findings,

all other hypotheses tested were nondirectional.

The effects of time period in the trial was assessed on verdict,

a weighted index of verdict, confidence in verdict decision, and

recall of trial information. Again, the potentially moderating

effects of demographic variables and effective states on these

relationships were also examined. It was predicted that an

interaction with verdict-relevant jury biases would be found. Here,

the literature predicted that biased jurors would be more likely to

decide on their verdict earlier in the trial and become progressively

more confident in their verdicts. Specific hypotheses were also made

with reference to recall of trial information. Here, it was predicted

that if informational exchange was an integral part of deliberation,

recall of trial information would improve after deliberation. No

predictions could be made as to whether certain types of trial

information were more likely to be recalled (with the exception that

it was likely that interactions with verdict-relevant jury biases

would occur). Examining the effects of time period also allowed for

clarifying the influence of opening arguments, testimony, judge's

instructions, and deliberation on verdict and confidence in verdict.

Additionally, to assess the efficacy of methodologies that study jury

behavior without the inclusion of a deliberation and/or using

posttrial juror polling, the effects of deliberation and the passing

of time (i.e., two weeks after the trial) on verdict, a weighted index

of verdict, confidence in verdict decision, and recall of trial

information was examined. Nondirectional hypotheses were tested with

respect to the relationship between the passing of time and recall of

trial information.

The effects of verdict-relevant jury biases were assessed on

deliberation style. Nondirectional hypotheses were tested for these

relationships. Also, to clarify these above-noted effects, the

effects of deliberation processes were assessed on the changes between

pre- and postdeliberation measures of verdict, confidence in verdict

decision, weighted verdict, recall of trial information, and affective


Of secondary importance, other relationships were examined to

illuminate the above findings; here, the role of recall of trial

information in its relationship with verdict was examined. It was

expected that predeliberation assessment of these recall variables

would indicate that jurors would recall more information that was more

favorable to their verdict decisions. Other hypotheses with respect

to these relationships were nondirectional. Also, the effects of

verdict-relevant jury biases, sex, and time period were assessed on

affective state. Again, nondirectional hypotheses were tested. In

addition, an attempt was made to illustrate the aspects of the trial

process which the jurors reported were most influential in their

decision making. Finally, a comparison between pre- and

postdeliberation juror verdict preferences and the jury verdicts were

ill ustrated.


Experimental Design

The study consisted of a single-factor experiment with repeated

measures on the same elements. Verdict-relevant biases of the juries

and sex were the between-group factors. Time period in the trial was

the repeated within-subject factor.


An initial subject pool consisted of 149 undergraduate psychology

students at the University of Florida. The subjects who agreed to

participate received extra credit towards meeting course

requirements. From the initial subject pool, 72 subjects (36 males

and 36 females) were selected on the basis of protesting to

participate as mock jurors. Seventy subjects (35 males and 35

females) actually participated in the study.

These subjects ranged in age from 18 to 33 years (mean = 20.2, SD

= 2.19). One hundred percent of the subjects were single; 92.9% were

white; 75.7% came from families with a reported annual income of

greater than $25,000; and 75.7% fell within social classes I-III in

Hollingshead's 2-factor index of socioeconomic status. These

demographic data are presented in more detail in Table 1.

Table 1
Demographic Variables

Variables Percent

Marital Status
Single 100

White 92.9
Black 4.3
Hispanic 1.4
Other 1.4

Protestant 31.4
Catholic 30.0
Jewish 20.0
None 12.9
Other 5.7

Social Class
I 28.6
II 25.7
III 21.4
IV 21.4
V 2.9

Annual Family Income
> $30,000 64.3
$21,000-30,000 19.9
$12,000-20,000 8.7
< $11,000 7.1

Description of the Case and Charges

The case used in the present study was entitled The State of

Florida vs. John Doe Smith. Hypothetically, the events of the case

took place just outside of Gainesville, Florida.

John Doe Smith, the defendant, was a 28-year-old, Tennessee-born,

white, married male with three children. He supported his family and

his grandfather. He had a 10th grade education. His wife had

recently contracted tuberculosis and could no longer work. The

defendant worked two blue-collar jobs, for a total of 60 hours per

week. Evidence was presented that he had recently been having several

marital arguments. On two separate occasions these were reportedly

followed by his wife calling her brother and father, who responded by

allegedly assaulting the defendant; once by attacking him with a knife

and once by driving up to the defendant's house and firing one shot

into the house at the defendant. Evidence was also presented as to

the defendant's excellent marksmanship abilities.

During the night (approximately 1:30 a.m.) of the incident in

question, the defendant reportedly had had a fight with his wife about

whether he could go fishing. The defendant left and went to a local

tavern and had two drinks. His wife later found him there with

another woman, with whom the defendant had been drinking. Another

marital dispute erupted. As the Smiths left the tavern together and

got into the car to go home, the defendant was seen striking his wife

with his hand. The owner of the tavern noted the license tag number

and phoned it to the police, reporting the dispute.

Once at home the defendant was packing his car to leave for the

weekend. His wife was inside the house. Two unmarked police cars

with their lights out arrived. This was testified to be standard

procedure when investigating marital disputes. The defendant, now on

the porch of his house, noticed the cars and "tracked" them with his

rifle as they drove by the house. The police officers got out of

their cars, verbally identified themselves, and ordered the defendant

to drop his gun. The defendant asked the police to identify

themselves by turning on their blue lights. (The defendant later

testified that he thought it was his father- and brothers-in-law

trying to retaliate against him for hitting his wife.) The police

refused to turn on their blue lights and repeated their order. They

testified that if they put on the lights, it would have made them an

easy target. As the defendant made his request again, one policeman,

Officer White, ran behind a telephone pole in front of the house. The

defendant fired one shot, which lodged itself 11 inches above the

ground in the same telephone pole. Officer Rhodes, the second

policeman, was running toward the pole and fell to the ground when he

heard the shot. After the shot, the defendant walked calmly back into

his house, reportedly put down his shotgun and had a cigarette and a

cup of coffee.

Later extra police officers came to assist. The defendant looked

out his window and saw the blue lights of the new cars and came out

onto the porch, where he was physically overtaken by two other

policemen. He was arrested and read his Miranda rights. He was

charged with Aggravated Assault against Officer White, Aggravated

Assault against Officer Rhodes, Discharging a Firearm in Public, and

Improper Exhibition of a Weapon (Supreme Court of the State of

Florida, 1981).

The equivocal nature of this case suggested that a jury could

find reasons for both acquittal as well as for conviction. It was

expected that given the presence of such uncertainty, jurors would

rely on their own personal biases to help them arrive at a decision.

Dependent Measures

Individual juror verdicts on all four charges were obtained at

several points during the experiment. This assessment yielded nominal

scale data (i.e., guilty or not guilty).

Individual juror's degree of confidence in their verdict

decisions were also assessed each time the jurors rendered a

verdict. This measure of confidence ratings of juror verdict

decisions was a nondirectional measure. That is, it assessed the

degree of the jurors' confidence in whatever their verdict was, making

no distinction between guilty and not guilty verdicts. This was

measured by instructing the subjects to make a mark on a line to

indicate the strength of their belief in their verdict choice. The

line's endpoints were labeled "No Confidence" and "Total

Confidence." This measure yielded ratio scale data with a range from

0 (no confidence) to 100 (total confidence).

A weighted index of verdict was created for use in this study in

order to have a directional measure of jurors' confidence. That is,

this measure assessed the degree of the jurors' confidence in their

specific (i.e., guilty or not guilty) verdicts. This index was

derived by multiplying the jurors' confidence in each verdict decision

by +1 if the verdict rendered was guilty and by -1 if the verdict

rendered was not guilty. This measure yielded ratio scale data with a

range from -100 (total confidence in not guilty verdict) to +100

(total confidence in guilty verdict).

A jury verdict for each charge was obtained by deliberation at

the conclusion of the trial. Each jury was required to deliberate to

a unanimous decision on each of the four charges. This assessment

yielded nominal scale data (i.e., guilty or not guilty).

The Multiple Affect Adjective Checklist (Zuckerman, 1960) was

employed to assess each subject's affective state at several points

during the experiment. This measure included scales for depression

(40 items), anxiety (21 items), and hostility (28 items). Subjects

were instructed to circle items which described how they felt at that

given point in time. Zuckerman, Lubin, Vogel, and Valerius (1964)

have reported Spearman-Brown split half reliability coefficients of

.92, .90, and .79 for depression, hostility, and anxiety,


Bales' (1951) interactional process analysis was employed to

classify the style of each jury interaction during the deliberation.

This method has often been reliably and effectively employed in

research concerning group problem-solving, especially jury research.

It classifies, into 12 categories, each action or statement one person

makes to another within the deliberation. These categories themselves

may be grouped into "metacategories," which may be classified in terms

of problem areas (e.g., communication, control, decision), responses

(e.g., positive or negative reactions, questions, attempted answers),

or the neutral task area.

Three raters, who were blind to the composition of the jury

conditions, were trained to use Bale's rating system. Two raters were

used in any one given time when rating the tapes. First, the raters

practiced on taped deliberations from a previous simulated jury

study. The raters would agree on the statement to be rated and then

would rate that statement independent of each other. They assigned

each statement made during the deliberation into 1 of the 12 Bale's

categories. The raters practiced until they obtained a statement-for-

statement interrater agreement of over 70% for a randomly chosen 5-

minute segment. This procedure was then used in the actual rating of

the deliberations. It should be noted that although different rater

dyads were used to rate the deliberation tapes, the same dyad would

rate an entire jury's deliberation. Statement-for-statement

interrater agreement was obtained for each jury deliberation. These

reliability indices are presented in Table 2. As can be seen,

interrater agreement was consistently high across juries (and rater

dyads) regardless of length of deliberation. Since the time taken to

reach a verdict varied across juries, all category scores were

transformed to rate (i.e., occurrence per minute) scores to permit

comparisons based on ratio scale data.

Recall of trial information was assessed with a questionnaire

designed for this investigation (see Appendix A). This was necessary

since every trial presents a unique set of information. The recall

Table 2
% Interrater Agreement of
Bale's Interactional Process Analysis Categories

Jury Length of
Number Jury Type % Agreement Deliberation (min.)

1 Egalitarian 89 29.27

2 Authoritarian 89 26.83

3 Mixed 88 35.17

4 Neutral 91 20.50

5 Egalitarian 84 24.52

6 Authoritarian 92 11.33

7 Mixed 80 34.45

8 Neutral 83 57.22

9 Authoritarian 91 3.25

10 Egalitarian 87 27.03

11 Mixed 77 11.42

12 Neutral 81 26.33

questionnaire consisted of a total of 32 questions: 8 about the

defendant's character, 8 about the evidence presented, 8 about

arguments presented by the prosecution (during opening and closing

statements), and 8 about arguments presented by the defense (during

opening and closing statements). These questions comprised indices of

character recall, situational evidence recall, prosecution-biased

recall, and defense-biased recall, respectively.

Questions were worded simply and excluded any biasing

presuppositions. They were open-ended and required that the subjects

respond with only 1- or 2-word answers. This open-ended format of

questioning has been shown to maximize both accuracy and completion of

recall better than either total free recall (i.e., free narrative) or

multiple-choice questions (Lipton, 1977; Penrod et al., 1982). A

similar questionnaire format was utilized in the study of Berg and

Vidmar (1975) concerning juror bias, as assessed by the LAQ, and

recall of trial information. Items were given a score of 1 if the

answer was correct and a score of 0 if the answer was incorrect or

unanswered. This questionnaire yielded ratio scale data.

At the conclusion of deliberation, all subjects were also asked

an open-ended question which inquired as to what aspects of the trial

had the greatest influence on their decisions. This was done in order

to obtain, from the jurors' perspective, important influential

variables. All the subjects' responses were listed and tabulated.

Inspection of these data revealed 10 different variables in the trial,

which were perceived as being most influential to their decision.

They were (a) the judge's instructions, which included the definitions

of the law; (b) the evidence; (c) the attorneys' opening and closing

arguments; (d) the prosecution attorney's arguments; (e) the witnesses

for the prosecution (their testimony and credibility); (f) the defense

attorney's arguments; (g) the witnesses for the defense (their

testimony and credibility); (h) the defendant's testimony and

credibility; (i) the presence of reasonable doubt (i.e., the

prosecution's failure of meeting the burden of proof); and (j)

deliberation. The subjects' responses comprised answers which fell

into one or a combination of these categories.

Independent Measures

Juror bias was assessed by the Legal Attitudes Questionnaire-Form

II (Boehm, 1968). The LAQ was designed to assess verdict-relevant

biases in jurors. It employs forced-choice format with a total of 30

items (divided into 10 triads). Each triad contains one item

representing each of the three verdict-relevant attitude dimensions:

authoritarian, egalitarian, and antiauthoritarian. For each triad,

the subjects were instructed to place a "+" beside the item with which

they most agree, "-" beside the item with which they least agree, and

to leave the remaining item unmarked. The "+," unmarked, and "-"

items were assigned scores of 3, 2, and 1, respectively. The scores

for each attitude dimension were then summed (the range for each

dimension could be from 10 to 30).

Since the authoritarian dimension is a relative one, for each

subject the score of the egalitarian subscale of the LAQ was

substracted from the authoritarian subscale to yield a distribution of

scores (i.e., +20 indicated the highest authoritarianism score; -20

indicated the highest egalitarianism score; 0 indicated theoretical

neutrality). The resulting distribution was divided into thirds

representing authoritarianism, theoretical neutrality, and

egalitarianism, respectively.1

To date, no reliability data have been provided for the LAQ.

Therefore, a 3-week, test-retest, pilot study of the LAQ was conducted

to provide such needed data. Seventy-nine subjects from an

Introduction to Criminal Justice class at the University of Florida

signed informed consent, agreeing to complete the LAQ on two

administrations which were 3 weeks apart and to provide demographic

data. Sixty-eight (86%) of the subjects (35 males, 33 females)

returned completed LAQs on both occasions. The ages of the subjects

were between 17 and 27 years (X = 19.6); 89.5% were white; 95.5% were

single; and 79.4% fell within social classes I-III on Hollingshed's

(1957) 2-factor classification.

Pearsons product-moment correlations yielded the following

significant reliability coefficients for the LAQ scores:

Authoritarian (r = .79, p<.0001), Egalitarian (r = .84, p<.0001), and

Antiauthoritarian (r = .25, p<.05). Of greatest importance, was the

1 The antiauthoritarian dimension was not investigated in the present
study because of its rare occurrence in subject pools of several
studies using the LAQ (Alexander & Licker, 1975; Berg & Vidmar,
1975; Buckhout et al., 1979; Goldman et al., 1983). Therefore,
subjects whose antiauthoritarian score was the highest of their
three LAQ dimension scores (2 out of 114 completed LAQs or 1.7% in
this sample) were eliminated from this study.

significance of the reliability coefficient for the subjects'

authoritarian minus egalitarian scores (r = .85, p<.0001) since this

score was used to create a distribution from which jurors were

categorized. This finding suggests that the relative difference

between the subjects' authoritarian and egalitarian attitudes remained

stable over a 3-week period of time.

Good construct and predictive validity of the LAQ have been shown

by demonstrating consistent differences in verdict and other jury

processes between juries differing with respect to LAQ

classification. These studies were reported in detail earlier

(Alexander & Licker, 1975; Berg & Vidmar, 1975; Boehm, 1968; Buckhout

et al., 1979).

For some of the analyses of this study, the Multiple Affective

Adjective Checklist (Zuckerman, 1969), Bales' (1951) interactional

process analysis, and the recall of trial information also served as

independent measures.


Three weeks prior to the showing of the simulated trial, the pool

of 149 subjects were administered the Legal Attitudes Questionnaire-

Form II (Boehm, 1968), the Multiple Affective Adjective Checklist

(Zuckerman, 1969), and a questionnaire ascertaining demographic

information. At this time, informed consent was also obtained from

all subjects.

Of the 149 questionnaire packets administered, 120 (81%) were

returned. Of these 120, 4 were completed incorrectly and those

subjects were dropped from this study. Two subjects' LAQ scores

showed that their Antiauthoritarian scores were greater than both

their Authoritarian and Egalitarian scores, and they were also dropped

from this study. The remaining 114 subjects comprised the venire or

pool of potential jurors.

Next, a distribution of each subject's Authoritarian minus

Egalitarian score was created. The scores ranged between -13 and +10

with a mean score of 0.25 and standard deviation of 4.89 (.5 standard

deviation = 2.45) and formed a normal distribution. A plot of this

distribution is shown in Figure 1. Based on this distribution, it was

decided that subjects with a score of less than or equal to -2 would

be classified as Egalitarian; subjects with a score between -1 and +1

would be classified as Neutral; and subjects with a score greater than

or equal to +2 would be classified as Authoritarian. Figure 2 shows

the normal distribution of classified jurors' Authoritarian minus

Egalitarian scores who served in experimental juries (mean = 0.19, SD

= 4.88, range = -12 to +10). These data indicate that the jurors who

served on experimental juries were representative of the experimental


Subjects were assigned to 1 of 12 6-person juries on the basis of

their pretested LAQ attitude distributions. In other words, the

administration of the LAQ served as the experimental voir dire.

Specifically, there were a total of three juries in each of four

conditions: authoritarian, theoretically neutral, egalitarian, or

mixed composition (authoritarian and egalitarian). Subjects

classified to be included into one of the four jury conditions were


x x x

x x x



x x x





x x

in 0
q r-4

in 0


,- 4 0 *
1 *)


+ 9z




* 0 o


In 2



O ** X

0 X


+ VI
i 0

o u' o ox Lx
-c- X -.

xxxx x

-^- -- -

I Ij 0



3 0)
0l -0


*i= -


randomly assigned to one of three juries within each condition, with

the exception that all juries contained equal numbers of males and


The subjects viewed a videotape of a simulated trial from the

University of Florida College of Law practice court. Here, advanced

law students acting as the trial attorneys had enacted a criminal

trial involving a case of aggravated assault. The trial enactment

took place in the College of Law classroom, which resembles the

physical appearance of actual courtrooms.

Subjects were seated with their assigned juries in rows. They

were instructed by the experimenter that the study is a collaborative

endeavor by the Departments of Clinical Psychology and School of Law;

that its purpose was to investigate how students sitting on a jury

view a criminal trial; and that they were not to take notes or have

discussions amongst themselves during the trial. They then viewed the

4-hour color videotape on a 5-foot screen in a small classroom.

Eighteen subjects (comprising three juries) viewed the videotape at

one time. This procedure was repeated until all subjects had

completed the procedure. Juries were randomly assigned to viewing


The tape contained the trial in its entirety, beginning with the

opening arguments. The tape was filmed from the perspective of the

onlooking jury. Following the presentation of the testimony and

closing arguments, the jurors were given standard instructions for

deliberation by the judge (Supreme Court of the State of Florida,

1981). These instructions were detailed, lengthy, designed to impress

upon the jurors the seriousness of their task, and to inform them of

the legal guidelines to be used in deciding their verdict.

Juries were then escorted to individual deliberation "chambers"

near the viewing room where the deliberations were tape recorded.

These "chambers" were the same size as a small conference room; each

contained a rectangular table in the center surrounded by chairs.

Before beginning their deliberation, each jury was given the same

instructions by the experimenter which informed them that they were to

first select a foreman and then to deliberate to a final decision

(i.e., guilty or not guilty) on each of the charges. Juries were

allowed up to one hour to deliberate. Jury verdicts, which were

recorded by the foreman, were also obtained following deliberation.

Several measures were given to all subjects repeatedly throughout

the study. Following the opening argument, testimony, judge's

instructions, deliberation, and at a 2-week posttrial follow-up, the

Multiple Affective Adjective Checklist was administered to assess

changes in affective state. Similarly, the jurors' individual

verdicts and degree of confidence was assessed. In addition, the

individual jurors' recall of trial information and their statement of

which aspects of the trial were most influential was assessed with a

questionnaire following the judge's instructions (i.e.,

predeliberation), postdeliberation, and at a 2-week posttrial follow-



Verdict-relevant jury biases (i.e., authoritarian, egalitarian,

mixed composition, and neutral), sex of the jurors, and time period

were the independent variables. The effects of these independent

variables were assessed on confidence in verdict decision, weighted

verdict, verdict, recall of trial information, and affective state.

Here, the BMDP2V program of the BMDP Statistical Software (BMDP, 1983)

was used to perform a series of repeated measures analysis of

variances, with type of jury and sex as the between-subjects factor

and time period as the within-subjects factor. These analyses were

performed to test the hypotheses with respect to the effects of jury

type, sex, and time period to jurors' confidence in verdict

decisions. Specific hypotheses included the predictions that

authoritarians would be more confident in their verdict decisions

throughout the trial than other types of jurors and, that for biased

jurors, the confidence in verdict decisions would increase as the

trial progressed. Additionally, to assess the role of potentially

moderating variables, a series of separate repeated analysis of

covariances were employed in which age; socioeconomic status; and

state of anxiety, hostility, and depression served as covariates.1

The same statistical procedures were used to test the hypotheses

with respect to the effects of jury type, sex, and time period of the

weighted verdict index. Here, it was predicted that authoritarians

1 State of anxiety, hostility, and depression were covariates which
changed over time periods.

might be more confident in guilty verdicts and egalitarians more

confident in not guilty verdicts. It was also hypothesized that type

of jury would show interaction with time period.

To analyze the data on verdict, first, the distributions of the

proportion of guilty verdicts within each jury per charge for each

separate time period were examined. For each charge, the variances of

these proportions were quite variable over the time periods.

Therefore, to stabilize the variances over time, the proportion of

guilty verdicts within each jury were transformed by obtaining the arc

sine of the square root of the proportion of guilty verdicts within

each jury. This is otherwise known as an arc sine transformation

(Winer, 1971). To test the hypotheses of the effects of the

independent variables on the arc sine transformation of the proportion

of guilty verdicts within each jury, the same analyses as described

above, were employed. Here, it was predicted that authoritarians

would show higher proportions of guilty verdicts than other juries.

It was also predicted, although with less certainty, that egalitarian

juries might show lower proportions of guilty verdicts than other

juries. It was also hypothesized that for these two jury types, the

proportion of guilty verdicts after the opening arguments might

predict the proportions following deliberation.

In order to test the hypotheses regarding the effects of the

independent variables on recall of trial information, the same

analyses were, again, employed. Authoritarians were hypothesized to

recall more about the defendant's character and the prosecution

attorney's trial arguments and less about the situational evidence

than other jurors. Egalitarians were expected to recall more

situational evidence and, perhaps, more about the defense attorney's

trial arguments than other jurors. Additionally, it was predicted

that recall of trial information might increase following deliberation

if information exchange was an integral part of the deliberation

process. Furthermore, it was predicted that follow-up assessment

would indicate a decrease in recall of trial information, as well as

changes in confidence in verdict decision, weighted verdict, and

verdict. Of additional interest, the hypothesis that jurors would

recall more information, that was favorable to their specific verdict

decisions, was tested by employing the Spearman's correlational

procedure to analyze the relationship between verdict and recall of

trial information.

In order to test the hypotheses regarding the effects of jury

type, sex, and time period on affective states, the BMDP4V program of

the BMDP Statistical Software (BMDP, 1983) was employed to perform a

repeated measures multivariate analysis of variance. Again, the

between-subjects factors were type of jury and sex and the within-

subjects factor was time period. The MANOVA was used because

protesting revealed that the variables for affective states showed

moderately high intercorrelations. Nondirectional hypotheses were


The effects of jury type on deliberation process was also

examined. To test the nondirectional hypotheses concerning these

relationships, a one-way analysis of variance was performed with type

of jury as the between factor, for each of the scales on the Bale's

(1951) Interaction Process Analysis ratings of the deliberations and

for the length of deliberation time.

Regarding nondirectional hypotheses of secondary interest, to

analyze the effects of deliberation processes on changes between pre-

and postdeliberation measures of verdict, weighted verdict, confidence

in verdict decision, recall of trial information, and affective state

a true score analysis (Cook & Campbell, 1979) was used. That is,

essentially, for each variable, a regression, using the general linear

models procedure of the Statistical Analysis System (SAS Institute,

1982), was performed between pre- and postdeliberation jury

averages. The residuals or true scores obtained from these

regressions were then correlated with the deliberation process



Repeated measures analysis of variances (ANOVAs) were employed to

analyze the effects of type of jury and sex on confidence ratings of

juror verdict decisions, the weighted index of juror verdicts,

verdicts, and the recall of trial information over different time

periods in the trial for each crime. Type of jury and sex served as

between-subjects variables and time period was the within-subjects

variable. To analyze how potentially moderating variables contributed

to these relationships, a series of separate repeated measures

analysis of covariances (ANCOVAs) were used, in which demographic

variables (i.e., age and socioeconomic status) and affective variables

(i.e., state of anxiety, hostility, and depression) served as

covariates.1 The BMDP2V program of the BMDP Statistical Software

(BMDP, 1983) was used to perform these analyses. The conservative

estimate by Geisser and Greenhouse (1958) of the probability value for

F-tests was utilized in cases where the assumption of sphericity

(i.e., that adjacent observations over time on the same variable are

independent or not correlated) was violated in order to reduce the

increased probability of Type I errors when such violations occur.

1 State of anxiety, hostility, and depression were covariates which
changed over time periods.

Confidence Ratings of Juror Verdict Decisions

Aggravated Assault against Officer White

For the crime of aggravated assault against Officer White, the

repeated measures ANOVA on confidence ratingsI of juror verdict

decisions yielded a significant main effect for time period, F(4, 248)

= 29.61, 2<.00001,2 and a significant interaction of time period and

type of jury composition, F(12, 248) = 2.21, _<.02.2 These results

are presented in Table 3. Additionally, this analysis found a

nonsignificant trend for the interaction of sex and type of jury,

F(3, 62) = 2.57, _<.06. The series of separate ANCOVAs revealed that

these relationships were not significantly altered when any of the

demographic or affective variables were used as covariates.

Tukey's studentized range a posteriori test (Winer, 1971) for

making pairwise comparisons between means (df = 248, p<.05) was used

to indicate that jurors' confidence in their verdict decisions were

relatively low after the opening arguments, but that after the

testimony their confidence significantly increased and, following the

judge's instructions (predeliberation) confidence again significantly

increased and remained essentially at the same level after

deliberation and at the 2-week follow-up. The means and standard

deviations for these results are presented in Table 4.

1 This measure of confidence ratings of juror verdict decisions is a
nondirectional measure. That is, it assessed the degree of the
jurors' confidence in whatever their verdict was, making no
distinction between guilty and not guilty verdicts.
2 The conversative estimate for the p-values for F-tests by Geisser
and Greenhouse (1958).

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EIALS8UK0_SOSOK0 INGEST_TIME 2012-02-20T22:46:32Z PACKAGE AA00009100_00001