An Examination of Happiness, Racism, and Demographics on Judgments of Guilt

17
An Examination of Happiness, Racism, and Demographics on Judgments of Guilt KIMBERLY DEAN, JULIE HOLLIDAY WAYNE, DAN MACK, AND KECIA THOMAS1 L'nnwr.wJ. of Georgir A policy-capturing study \*as conducted to evaluate the roles of hoth plaintiff and detcn- dant race and gender on judgments of criminal guilt in either an assault or a theti case. It was hypothesized that defendant gender and race would affect likelihood of guilt ratings. Moreover, mock jurors' dispositional happiness, scores of modern racism, and race and gender were explored as potential influences on their decision-making policies and judg- ments. As predicted, strong support was found for the role of defendant gender on judg- ments ofguilt. However. defendant race did not significantlq impact guilt ratings. Post hoc analqses revealed the importance of race and gender of the victim to juror decisions. In addition. participants' o u n race and gender played signiticant roles in the decision- making policies that uere produced. Limitations of the study are discussed. as are topics for future research. Every day, we use discriminators such as race, religion, or gender to identify people with whom we are similar and those from whom we are different, and we then organize incoming information around these perceptions (Allport. 1958; Ehrlich, 1973; Taylor, 1981). In 1922, Lippmann first applied the term sfeweotype to describe this process. The problem with stereotyping is that all individuals who belong to a certain social group are viewed as possessing a certain trait or traits. Individuals are seen solely in terms of their group membership, rather than as unique human beings, and are responded to based on their group membership and its assumed char- acteristics (Ehrlich, 1973; Hamilton & Trolier, 1986; Katz, 1981). In addition, stereotypes are rigid, resistant to change, and may be used as justification for the status quo (Sherif & Sherif, 1953; Tajfel, 1969; Taylor, 1981). Stereotypes are precursors to prejudice and impact decision making in many aspects of social interaction. One important area to understand the impact of stereotypes on decision mak- ing is juror decision making. It is necessary to understand factors that may lead to 'Correspondence concerning this article should be addressed to Kecia M. Thomas, Department of Psychology, University of Georgia, Athens, GA 30602. e-mail: [email protected]. 816 Journal of Applied Social Psychology, 2000, 30, 4. pp. 816-832 Copyright 0 2000 by V. H. Winston 8 Son, Inc All rights reserved.

Transcript of An Examination of Happiness, Racism, and Demographics on Judgments of Guilt

An Examination of Happiness, Racism, and Demographics on Judgments of Guilt

KIMBERLY DEAN, JULIE HOLLIDAY WAYNE, DAN MACK, A N D KECIA THOMAS1

L'nnwr.wJ. of Georgir

A policy-capturing study \*as conducted to evaluate the roles o f hoth plaintiff and detcn- dant race and gender on judgments of criminal guilt in either a n assault or a theti case. It was hypothesized that defendant gender and race would affect likelihood of guilt ratings. Moreover, mock jurors' dispositional happiness, scores of modern racism, and race and gender were explored as potential influences on their decision-making policies and judg- ments. As predicted, strong support was found for the role of defendant gender on judg- ments ofguilt. However. defendant race did not significantlq impact guilt ratings. Post hoc analqses revealed the importance o f race and gender o f the victim to juror decisions. In addition. participants' o u n race and gender played signiticant roles in the decision- making policies that uere produced. Limitations of the study are discussed. as are topics for future research.

Every day, we use discriminators such as race, religion, or gender to identify people with whom we are similar and those from whom we are different, and we then organize incoming information around these perceptions (Allport. 1958; Ehrlich, 1973; Taylor, 1981). In 1922, Lippmann first applied the term sfeweotype to describe this process.

The problem with stereotyping is that all individuals who belong to a certain social group are viewed as possessing a certain trait or traits. Individuals are seen solely in terms of their group membership, rather than as unique human beings, and are responded to based on their group membership and its assumed char- acteristics (Ehrlich, 1973; Hamilton & Trolier, 1986; Katz, 1981). I n addition, stereotypes are rigid, resistant to change, and may be used as justification for the status quo (Sherif & Sherif, 1953; Tajfel, 1969; Taylor, 1981). Stereotypes are precursors to prejudice and impact decision making in many aspects of social interaction.

One important area to understand the impact of stereotypes on decision mak- ing is juror decision making. It is necessary to understand factors that may lead to

'Correspondence concerning this article should be addressed to Kecia M. Thomas, Department o f Psychology, University o f Georgia, Athens, GA 30602. e-mail: [email protected].

816

Journal of Applied Social Psychology, 2000, 30, 4. pp. 816-832 Copyright 0 2000 by V. H. Winston 8 Son, Inc All rights reserved.

JUDGMENTS OF GUILT 817

bias against certain defendants. The present study explores the effects of gender and racial stereotypes on perceptions of the defendant’s likelihood of guilt of a crime.

Empirical data supporting the notion that stereotypes function as cognitive heuristics are abundant. Researchers have found that individuals judge an infrac- tion as more likely to recur when it is committed by an individual with whom the infraction is stereotypically associated than when it is committed by someone with whom it is not associated (Bodenhausen & Wyer, 1985). Bodenhausen and Wyer also found that participants recalled less information about a scenario when the offense was stereotypical than when i t was not. That is, the participants seemed to pay less attention to information when a stereotypical explanation for the crime existed. A follow-up study suggested that people use stereotypes about others as heuristics. When life-circumstance information was considered with regard to the defendant, nonstereotyped defendants received more lenient sen- tences. However, when a stereotyped explanation for the crime existed, there was no more leniency when life-circumstance information was given. Individuals used the stereotype as an easy explanation for the crime and as a basis for their evaluations. In accordance with these findings, we expected that people would judge gender- and race-stereotyped defendants more harshly than they would nonstereotyped defendants. Namely, male and minority defendants were expected to be most at risk for evaluations of guilt.

Hypothesis 1. There will a significant relationship between defen- dant gender and likelihood of guilt (specifically, male defendants will be rated as more likely to be guilty).

Hypothesis 2. There will be a significant relationship between defendant race and likelihood of guilt (specifically, minority defendants will be rated as more likely to be guilty).

Emotion Research

Research in the area of mood has demonstrated that mood can play a role in social cognition, although the results have, at times, been rather conflicting. Despite the fact that there are many different theories of how emotions function, there seems to be agreement about some of the chief effects of mood on social cognition. Generally, positive mood has been associated with increased confi- dence in one’s own as well as others’ abilities, a less systematic style of informa- tion processing in the areas of persuasion, and the use of broader categorization systems (Bless, Bohner, Schwartz, & Strack, 1990; Bohner, Crow, Erb, & Schwarz, 1992; Brown, 1984; Isen & Daubman, 1984; Mackie & Worth, 1989). In addition, Bodenhausen, Kramer, and Suesser (1994) demonstrated that

818 DEAN ET AL.

participants in a positive mood rated a stereotyped defendant as more likely to be guilty than a nonstereotyped defendant; however, there was no such effect for neutral-mood participants. Thus, participants in a positive mood use less system- atic information processing and rely more heavily on the use of stereotypes than do neutral-mood participants. One purpose of the present study is to determine whether the effects of being in a positive mood (i.e., greater reliance on stereo- types) occur for persons who are dispositionally happy.

Traditionally, negative mood was found to be correlated with a more syste- matic processing style and a more realistic evaluation of performance. Partici- pants who are induced to feel a negative mood, specifically sadness. show increases in systematic processing as evidenced by careful assessment of argu- ment strength, as well as other differences in cognitive processing (Bless et al., 1990; Bohner et al., 1992; Brown, 1984; Isen & Daubman, 1984; Mackie & Worth, 1989).

Research on disposition (i.e., the chronic tendency to experience a particular type of emotion) has generally demonstrated that there are, indeed, individual differences in day-to-day emotional experiences and that these dispositions may impact other facets of psychological functioning. The work by Bodenhausen and colleagues (Bodenhausen, Kramer, & Susser, 1994; Bodenhausen, Sheppard, & Kramer, 1994) demonstrated that mood does relate to stereotyping beha\ ior and prejudicial judgments of guilt, at least in experimental situations. However, research has not yet examined how a person’s disposition (as opposed to a tran- sient mood state) impacts stereotyping behavior. Bodenhausen, Krainer & Suesser ( 1994) suggested that research examining the impact of affective disposi- tion on stereotyping would prove interesting. At present, there is no empirical information that has addressed such a notion. The present study will investigate whether dispositionally happy participants (like participants in positive moods) rely more heavily on stereotypes than do participants who are either disposition- ally sad or neutral. Based on the research on mood and stereotyping, the follow- ing hypotheses are proposed:

Hypothesis 3. Dispositional happiness will moderate the relation- ship between defendant gender and likelihood of guilt.

H-vpolhesis 4. Dispositional happiness will moderate the relation- ship between defendant race and likelihood of guilt.

Specifically, it is expected that participants who score highly on happiness will rely more heavily on stereotypes when assigning judgments of guilt than will participants who score low on the happiness measure. Therefore, i t is expected that happy participants will assign more guilty ratings to defendants who are people of color than they will to White defendants, when compared to the guilty

JUDGMENTS OF GUILT 81 9

assignments made by sad participants. Likewise, i t is expected that happy participants will assign more guilty ratings to male defendants than to female defendants, as compared to participants who score lower on happiness.

Modem Racism

Racial attitudes have become increasingly more complex and ambivalent. Though few people today would say that they think that discrimination is accept- able, stereotypes flourish and people are willing to justify beliefs that are dis- criminatory on nonracial grounds (McConahay & Hough, 1976). Although blatant forms of racial discrimination are less frequently expressed now than they have been in the past, more subtle forms of racism persist (McConahay, Hardee, & Batts, 1981). Individuals who are modern racists may seem to comply with new egalitarian racial nonns and express them verbally, without actually internal- izing them (Pettigrew & Martin, 1987).

McConahay (1 986) developed the Modern Racism scale and construct in order to assess the subtler forms of racism toward Blacks that are prevalent today. His construct examines reactions to the following four principal tenets: (a) Dis- crimination no longer exists because Blacks can compete freely in the market- place and enjoy the things that they can afford; (b) Blacks are pushing too hard, too fast and into places where they are not wanted; (c) These tactics and demands are unfair; and therefore, (d) Recent gains are undeserved, and the prestige- granting institutions of society are giving Blacks more attention and status than they deserve (McConahay, 1986). In addition, modern racists simultaneously believe that racism is bad, yet they believe and will espouse the four tenets stated previously because they believe them to be societal facts. For example an indi- vidual may answer that he or she is opposed to busing designed to integrate schools by saying that he or she favors neighborhood schools and quality educa- tion, By espousing such high abstract ideals, individuals can respond in a racist fashion and hold racists beliefs, yet justify those beliefs to themselves and others on the basis of something with which no one could disagree. Who doesn’t favor “quality education” (McConahay, 1986)?

Therefore, another hypothesis of this study is that modern racism would influence the judgments of defendant guilt that are made by participants. That is, participants who score higher on this modern measure of racism and prejudice will be more likely to use racial stereotypes in judgments of the likelihood of guilt than will participants who score lower on this measure.

Hypothesis 5. Participant modem racism (MR) will moderate the relationship between defendant race and likelihood of guilt (spe- cifically, participants with a high MR score will make harsher judgments of stereotyped defendants than will participants with

820 DEAN ETAL.

low MR scores). I t is expected that participants high in MR will be more likely to assign guilty ratings to defendants who are Black or Hispanic than to defendants who are White.

Demographic Characteristics

In addition, the present study seeks to explore the effects of demographic fac- tors on stereotyped or prejudiced judgments made about in-group and out-group members. Specifically, participant race and gender will be examined from an exploratory perspective in order to investigate whether either variable serves to moderate the relationship between defendant characteristics and ratings of guilt. No direction of these potential moderators is stated, given that these questions will be examined from an exploratory perspective.

Hypothesis 6. Participant gender will moderate the relationship between defendant gender and likelihood of guilt.

Hypothesis 7. Participant minority status will moderate the rela- tionship between defendant race and likelihood of guilt.

A summary of all of the hypotheses is presented in Figure I

Method

Participants

Forty-one students2 at the University of Georgia participated in the study as one of several options for partial course credit. Eleven men and 30 women partic- ipated individually or in groups of 2 to 15. The ages of the participants ranged from 18 to 22 years. with a mean age of 19 years. There were 35 Whites. 4 Blacks, 1 Native American, and 2 participants who did not list their race. The major areas of study for the participants in the sample, included natural science (4), social science ( I I ) , business ( 6 ) , journalism (5), communication and educa- tion (7), family ( 2 ) , and undecided (4). while 3 participants did not list their maj ors.

ZForty-tuo students completed the survey but one participant’s data was deleted because of an inordinate amount of missing data. However. because our data were analyzed at a within-subject level the effective sample size for the present study is actually the number of crime reports (i.e., 23) evalu- ated by each ofthe 41 participants ( in this case 943). The number o f participants i n the prescrit study is actually rather large given that other published policy capturing studies have much smaller N s (e.g., Dougherty, Ebert, & Callender, 1986; Sherer, Schwab, & Heneman, 1987; Velenzi & Andrews, 1973).

JUDGMENTS OF GUILT 821

Modern Racism T

Likelihood of Guilt ]

RD Race of Defendant SD Sex of Defendant RP Race of Participant GP Gender of Partripant

Figure 1. Summary of individual and cross-level hypotheses.

Design and Materials

The study used a 2 x 3 x 3 x 2 x 2 mixed factorial design. Policy capturing, which involves using a factorial design where independent variables are system- atically rotated in the form of vignette descriptions (Rossi & Anderson, I982), was used in the present study to manipulate the remaining independent variables. This within-subjects approach is more powerful than a between-subjects approach. The following variables served as within-subjects variables: type of crime (assault vs. theft), race of defendant (Black vs. White vs. Hispanic), race of victim (Black vs. White vs. Hispanic), gender of defendant, and gender of victim.

Respondents received a set of 23 “police reports” that were chosen randomly from the total possible pool of 72 such scenarios. The police reports were stan- dardized with respect to location of incident (campus), and all of the infractions occurred at a random time between 12 p.m. and 6 p.m. In order to make the sce- narios as believable as possible, the physical description of the individuals involved was different across scenarios, as were names and addresses of the defendant, victim, and two witnesses.

Before reading the cases, participants completed a Demographic Information questionnaire. Included in this instrument was the McGreal and Joseph (1993) Depression-Happiness Scale (DHS).

Depression-Happiness Scale

This instrument asked participants questions about their feelings in the last 7 days (e.g., “I felt cheerful”). They rated each of the 13 items on a 4-point scale

822 DEAN ET AL.

Table 1

Means. Standard Deviations, and Scale Alphas for Modern Racism and Happiness Scales

Scale M SD a

Modern Racism 12.75 4.79 .88 Happiness 29.54 6.24 .89

ranging from 0 (never) to 3 (qfren). McGreal and Joseph (1993) found strong evi- dence of the scale’s internal validity and concurrent validity with the Beck Depression Inventory (Beck, Rial, & Rickels. 1974). Research with under- graduate students has provided psychometric evidence for the McGreal and Joseph scale. For example, Lewis and Joseph (1995) found an internal reliability of .91 for the DHS, as well as evidence of its convergent validity with other measures of happiness and satisfaction with life. Furthermore, Lewis (1996) pro- vided evidence of the convergent validity of the DHS with the Crown-Crisp Experiential Index. Additional evidence of the psychometric soundness of the DHS has also been found in the adult working population (Walsh, Joseph, & Lewis. 1995).

After they completed the demographic questionnaire, participants read the 23 police reports. After reading each case, participants judged the degree to which they thought that the defendant was guilty or innocent using a single item, “Please rate the degree to which you think the defendant is likely to be guilty of the alleged infraction on a scale from I to 7.” The scale values ranged from 1 (extremely imlikel,v) to 7 (extremel.v likely).

After the participants evaluated each of the 23 crime reports, they completed a postdecision survey that included the five-item MR scale (McConahay. 1986; McConahay & Hough. 1976) and several other political and campus-related questions. Table 1 provides the means, standard deviations, and coefficient alphas for the MR scale and the DHS. In order to examine if there were signifi- cant gender or race differences on either the MR or DHS scales, t tests were per- formed. Results revealed no gender differences on either the DHS, r(39j = .45, ns, or the MR scale. t ( 3 8 ) = 1.56, ns. Similarly, there were no racial differences on either the DHS, t ( 3 8 ) = 1.27, ns. or MR scales, t(37) = 1.91, ns.

Results

Hierarchical linear modeling (HLM) was used to analyze the data. A two-step HLM analysis was conducted. The first step of the analysis examined whether

JUDGMENTS OF GUILT 823

linear effects were found between the independent variables (i.e., gender of victim, gender of defendant, race of victim, race of defendant, and type of crime) and the dependent variable, likelihood of guilt. The second step examined any potential moderators of this relationship. Specifically, HLM examined the effects of inter-individual characteristics such as participant gender and race. modern racism, and dispositional happiness on the intra-individual decision-making pro- cesses found during Step 1.

Tests n f lndividiral-Level Lineor Efects

Each of the factors used in creating the scenarios (race of victim and defendant; gender of victim and defendant; type of crime) was investigated to examine the impact of each on judgments concerning the likelihood of defendant guilt.

Gender of defendant. As predicted in Hypothesis I , gender of defendant was found to be significantly negatively related to likelihood of guilt, r = -. 17, t(40) =

-2.91, p < .01. That is, male defendants were rated as more likely to be guilty than were female defendants, regardless of the type of crime or the gender or race of the victim.

Race of dqfendunt. Contrary to Hypothesis 2 , race of defendant was not sig- nificantly related to likelihood of defendant guilt, r = .058, t (40) = 1.44, ns. Thus, participants were not using the race of the defendant in their determination of the defendant’s likelihood of guilt.

Ruce of victim. Race of victim did have a significant, negative linear effect on likelihood ratings, r = -.098, t(40) = -2.09, p < .05. These results indicate that when the victim of a crime was White, the defendant was perceived as more likely to be guilty than if the victim of the same crime was a minority. These findings are consistent with the stereotyping literature reviewed earlier (judg- ments are less harsh when a person is a majority group member, rather than a minority ).

Gender of victim. The HLM analysis revealed a significant, positive relation- ship between gender of victim and ratings of likelihood of guilt, = ,146, t(40) =

2.71, p = .01. This finding indicates that when the victim of a crime was female, the defendant was rated as more likely to be guilty than if the victim of the same crime was male.

Gpe qf crime. The relationship between type of crime and likelihood of guilt approached, but did not reach, traditional levels of significance. The relationship between type of crime and likelihood of guilt was negatively related, r = -.093, t(40) = -1.86, p = .07. Although the relationship was not significant, i t suggests that persons who were reported as committing assault crimes were perceived as more likely to be guilty than were those who were reported as committing theft crimes.

824 DEAN ET AL.

Table 2

Cross-Level Effects of Modern Racism and Happiness on Likelihood of Guilt ~~

Predictor ~

r t ratio p value

Intercept Race of victim (intercept)

Happy MR

Race of defendant (intercept)

Happy MR

Gender of victim (intercept)

Happy MR

Gender of defendant (intercept)

Happy MR

Type of crime (intercept)

Happy MR

3.39

-0.10 0.09

-0.02 0.06

-0.05 0.02 0.15

-0.06 0.03

-0.17 0.03

-0.04 -0.09

0.00 -0.03

33.75 -2.5 1

1.44 -0.62 1.65

-1.00 0.45 2.70

-0.69 0.54

-3.60 0.37

-0.79

-1.70 0.02

-0.54

.00

.02

.14 3 2 .10 2 4 3 6 .02 .3 1 .34 .oo .37 2 9 . I0

.40 2 4

Table 3

Esrimation of Variance Accountedfor by Group-Level Variables

Variance Predictor SD component x2 P

Intercept 0.637 .406 1,095.04 .000 Race of victim 0.233 .054 100.39 .000 Race of defendant 0.167 .028 62.98 .013

Gender of victim 0. I60 .026 38.6 1 >.so0 Gender of defendant 0.270 .073 80.73 ,000 Type of crime 0.097 .307 36.89 >SO0

Note. df = 40.

JUDGMENTS OF GUILT 825

Tests of Cross-Level Effects of Participants’ Modern Racism and Happiness Scores

In the first Level 2 analysis, two of the between-group variables (modern rac- ism and happiness) were simultaneously introduced as moderators of the linear effects found in Step 1. Results indicate, however, that modern racism and happi- ness did not significantly moderate any of the linear relationships discussed ear- lier (Table 2 ) . Therefore, contrary to Hypotheses 3 , 4, and 5, modern racism and happiness did not moderate the linear relationships as we hypothesized.

Tests of cross-level effects ofparticipants ’ gender and minority status. In the second Level 2 analysis, two other between-group variables (participant gender and minority status) were simultaneously introduced as moderators of the linear effects found in Step 1. Chi-square analyses indicate that for several of the linear effects found-specifically those involving gender of the defendant, race of the defendant, and race of the victim-there was significant variation that could be accounted for by group-level variables (Table 3 ) . The cross-level analyses indi- cate that gender and minority status did serve as significant moderators for some of the linear relationships (Table 4). Some of these findings were hypothesized a priori; others were not, as indicated in Table 4.

The effects of participant gender and minority status were examined as mod- erators of the hypothesized linear relationships among the independent variables, defendant gender and race, and the dependent variable, participants’ evaluations of the likelihood of guilt. Specifically, participant gender was expected to moder- ate the relationship between defendant gender and guilt. Similarly, participant minority status was expected to moderate the relationship between defendant race and guilt.

The individual-level effects of victim race and gender on ratings of guilt deci- sions were not hypothesized a priori and were therefore treated in a post hoc manner. Additionally, cross-level examination of these potential linear effects was also post hoc.

Gender of defendant. This analysis examined Hypothesis 6 in order to deter- mine whether participant gender moderated the significant negative linear rela- tionship between gender of defendant and likelihood ratings. Recall that this significant linear relationship indicated that males were more likely to be per- ceived as guilty of a crime than were females. Participant gender was found to be a significant moderator of this relationship, I- = -.259, t(38) = 2 . 0 1 , ~ = .05. This result indicates that participant gender negatively affected the strength of the relationship between gender of the defendant and likelihood of guilt. In other words, female participants were more likely to rate male defendants guilty of a crime than were male participants.

Race of defendant. This analysis investigated Hypothesis 7 in order to deter- mine if participant race moderated the linear relationship between defendant race

826 DEAN ETAL.

Table 4

Cross-Level Effects of Participant Gender and Minority Status on Likelihood of Guilt

~

Predictor r t ratio P

Intercept 3.39 33.66 .oo** Race of victim (intercept)b -0.10 -2.42 .03 *

Genderb 0.26 2.80 .01**

Minorityb -0.3 1 -2.84 .01** Race of defendant (intercept)a 0.06 1.67 1 .oo

Genderb -0.29 -3.73 .oo** Minorityb 0.09 0.94 2 5

Gender of victim (intercept)h 0.15 2.79 .o I ** Genderb -0.09 -0.79 .29 Minorityb 0.24 1.71 .09

Gender of defendant (intercept)a -0.17 -3.00 .01**

Gendera -0.26 2.0 1 .05* Minorityb -0.13 -0.84 .28

Type of crime (intercept)h -0.09 -1.84 .os Genderb 0.04 0.37 3 7 Minorityb 0.13 0.94 .25

aHypothesized a priori. bPost hoc * p 5 . 0 5 . * * p ~ . O l .

and likelihood of guilt. The results reveal that Hypothesis 7 was not supported, r = .088, t(38) = .998, ns. Other analyses revealed, however, that participant gen- der was a significant moderator of the relationship between defendant race and likelihood of guilt.

As discussed previously, results indicate that the intercept value for the linear relationship between race of defendant and likelihood was not statistically signif- icant at the a priori established level of .05, r = .058, ((38) = 1.67, p = . lo. How- ever, the results indicate that the impact of participant gender on the relationship between race of the defendant and likelihood was statistically significant and negative, r = -.29, t(38) = - 3 . 9 3 , ~ < . O l . Together, these two parameters associ- ated with the within-group slope between race of the defendant and likelihood can be interpreted as follows: There was not a significant linear relationship

JUDGMENTS OF GUILT 827

between race of the defendant and likelihood ratings. However, for male partici- pants, the slope of this line became more positive. Thus, for male participants, there was a stronger relationship between being a minority defendant and the likelihood of his or her guilt.

Race of victim. Although victim effects were not explicitly hypothesized a priori, this analysis examined whether participant gender or minority status mod- erated the significant negative relationship between race of the victim and likeli- hood ratings. Both gender and minority status were significant moderators of this relationship. Specifically, participant gender, r = .257, t(38) = 2.81, p < .05; and minority status, r = -.312, t(38) = -2.84, p < .05, were significantly related to the variance in this parameter across participants.

The significant prediction of variance by participant gender indicated that participant gender positively affected the strength of the relationship between race of the victim and likelihood of guilt. Thus, the negative relationship between race of the victim and likelihood of guilt was stronger for female participants. Female participants were more likely than were male participants to rate defen- dants as guilty of a crime if the victim was White.

Similarly, the significant prediction of variance by participant minority status indicated that the negative relationship between race of the victim and likelihood of guilt was stronger for White participants. Conversely, for minority partici- pants, the relationship between race of the victim and likelihood of guilt was weaker. Thus, the effect of a victim’s race had a greater influence on perceptions of likelihood of guilt for White participants than for minority participants. The significant individual and cross-level effects, both a priori and post hoc, are suni- marized in Figure 2.

Discussion

In summary, the results indicate that for the crimes used in the present study (assault and theft), gender stereotypes are often used in judgments about a defendant’s likelihood of guilt, resulting in a bias against male defendants. In addition, the race and gender of the victim are important determinants of guilt ratings. That is, in cases with female or White victims, defendants were rated as more likely to be guilty. Contrary to our hypotheses, race of the defendant was not an important determinant in judgments of guilt. In fact, it was the only inde- pendent variable that did not have a linear effect on guilt judgments. However, given that this study was conducted during the “trial of the century” (the 0. J. Simpson trial), it may be true that participants were primed toward the race of the defendant and may have been responding in socially desirable ways. Thus, the socially desirable responding may have made findings based on race-based stereotypes implausible. This question is only speculative and indicates the need for further empirical investigation. Finally, the results did not support the

828 DEAN ET AL.

Likelihood of Guilt

SD Sex of Defendant RV Race of Vcbm SV Sex of Vrtim SP Sex of Panicipant RP Race of Participant

a apriori b post-hoc

Figure 2. Summary of individual and cross-level effects on likelihood of guilt ratings.

hypothesis that dispositional happiness would systematically affect juror judg- ments. Thus, dispositional happiness did not demonstrate the same effect on deci- sion making as did being in a positive mood. However, given that this is the sole study to examine the effect of disposition on stereotyping, further research is needed.

These results suggest that victim and juror characteristics are very important to juror judgments. Specifically, certain types ofjurors weighed victim character- istics (race and gender) more heavily in their decision making. To reduce the effects of stereotyping on juror judgments, researchers need to further investigate these moderating factors and attempt to overcome them. These results also have important implications for attorneys trying criminal and civil cases. For example (and based only on the results of this single study), in a case where the defendant is female, defense lamyers should try to select as many female jurors as possible, given that female participants were less likely to perceive a female defendant as guilty of a crime. Similarly, these results suggest that if a defendant is a minority, male jurors may be harsher in judgments against minority defendants than against White defendants.

There are limitations to the present study. As with other types of laboratory research, we must concern ourselves with the external validity of our findings. There are several characteristics of this study that limit the generalizability of the results. First, the methodology employed here was a paper-and-pencil measure. Participants were presented with brief “police reports” and were asked to judge the defendant’s likelihood of guilt. The results indicate that participants were

JUDGMENTS OF GUILT 829

using certain characteristics of the victim (i,e., gender and race) and defendant (i.e., gender) to determine whether the defendant was guilty of a crime. Usually, the goal of conducting juror research is to be able to generalize those results to the courtroom.

Reactions to defendants may be different in actual court cases when jurors are presented with actual people and are provided with more contextual information. I t is possible that bias might be even stronger against certain racial or gender groups when minority defendants and victims are visually present before jurors. The race and gender of the defendant and victim in these scenarios were pre- sented but not highlighted; that is, the participants were not made blatantly aware of these characteristics, yet they were still important to their decisions. These characteristics may be more salient in an actual courtroom when the characteristics are visible and could have even more of an impact on decision making.

Or, it may be that the scarcity of information in the police reports caused par- ticipants to rely more heavily on the information that they did have and conse- quently to rely on stereotypes. A future research question that should be addressed is whether participants would arrive at the same conclusions if more information was given (e.g., judicial instructions) and if participants actually saw the defendant and victim. To overcome this limitation in future studies, research- ers could use videotaped scenarios with more information about the crime and provide judicial instructions to increase the realism of the scenarios.

A second limitation is the use of college students with limited experience in the role ofjurors. In a real jury process, jurors are given instructions by the judge for use in decision making, which participants in this study were not given. Thus. one must consider whether these same results would be obtained using an actual juror sample.

Another limitation of this study was the presence of only 4 Black participants in this sample. While a greater number of minority participants would have been desired, it is interesting to note that the minority status of participants did signifi- cantly affect the relationship between victim’s race and participants’ evaluations of defendant guilt. In addition, participants’ minority status also came close to moderating the relationship between victim gender and evaluations of guilt (Table 4). Including a greater number of Black participants in future studies may provide more powerhl tests of the role of modem racism and participant race as moderators. In addition, since the race of the defendant was defined as White, Black, or Hispanic, the lack of Hispanic participants hampered the ability to find additional participant race effects. Future studies should be conducted on other groups of participants to determine whether these results have population validity (Christensen, 1994). However, even if these results were not replicated with an older group of participants, it is important to recognize that the sample used here reflects a segment of the population.

830 DEAN ET AL.

Potential problems with the DHS and MR measures may represent another limitation to the study. The lack of moderation by happiness on defendant guilt may have been a result of a restriction of range on the happiness measure. In the overall sample, the majority of respondents were at the high end of the happiness scale (i.e., more high happiness scores). Thus, not enough variability was present to find differences among participants. Interestingly, the restriction of range on the happiness measure was not present on the minority sample. Additionally. for the minority sample, there was a restriction of range on the MR variable; the majority of the scores indicated low levels of modern racism. This may be a result of more experience with racism on the part of minorities.

The MR scale may be somewhat dated i n that the participants i n this sample may be growing increasingly sophisticated in their knowledge of issues such as racism, particularly given that the study was conducted during the 0. J . Simpson trial. As such, the reactivity of the items on the MR scale may have led to more socially desirable responding on the part of the majority group respon- dents.

There are several future research ideas suggested by this study. First, as men- tioned, researchers should examine whether results in scenario research differ when videotapes are used, rather than paper-and-pencil measures. Second. researchers should also consider the role that attorney characteristics (i.e.. race and gender) have on juror decisions. Third, the independent variables used here should be studied again using other types of crime. If you recall, only assault and theft were used here. Future research should also include other types of critne, including white-collar crime, sexual assault, and other crimes. I t may be that defendant and victim minority characteristics may be greater (or less) when dif- ferent categories of crimes are used. Finally, the role of disposition in stereo- typing behavior in court cases must be explored further.

Despite the limitations of this study, the results strongly suggest that the par- ticipants were using extralegal factors to determine guilt. The goal of ourjudicial system is to provide defendants with a fair and impartial jury. How is that possi- ble if jurors bring their inherent biases into the jury box? Future research should continue in this direction to identify when stereotypes against defendants occur and how to reduce them.

References

Allport, G. W. ( 1958). The natirre ofprejudice. Reading, MA: Addison-Wesley. Beck, A. T., Rial, W. Y., & Rickels, K. ( 1 974). Short form of depression inven-

tory: Cross validation. Psychological Reports, 34(3. pt. 2) , 1184-1 186. Bless, H., Bohner, G., Schwarz, N., & Strack, F. (1990). Mood and persuasion: A

cognitive response analysis. Personality arid Social Psychology Biilletin, 16(2), 331-345.

JUDGMENTS OF GUILT 831

Bodenhausen, G. V., Kramer, G . P., & Suesser, K. (1 994). Happiness and stereo- typic thinking in social judgment. Journal of Personality and Social Psychol-

Bodenhausen, G. V., Sheppard, L. A., & Kramer, G. P. (1994). Negative affect and social judgment: The differential impact of anger and sadness. European Journal of Social Psychology, 24,45-62.

Bodenhausen, G. V., & Wyer, R. S. (1985). Effects of stereotypes on decision making and information processing strategies. Journal of Personality and Social Psychology, 48,267-282.

Bohner, G., Crow, K., Erb, H., & Schwarz, N. (1 992). Affect and persuasion: Mood effects on the processing of message content and context cues and on subsequent behaviour. European Journal of Social Psychology, 22,5 I 1-530.

Brown, J. (1984). Effects of induced mood on causal attributions for success and failure. Motivation and Emotion, 8, 343-353.

Christensen, L. B. (1994). Experimental methodology (6th ed.). Boston, MA: Allyn & Bacon.

Ehrlich, H. J. ( 1 973). The social psychology of prejudice. New York, N Y John Wiley & Sons.

Hamilton, D. L., & Trolier, T. K. (1986). Stereotypes and stereotyping: An over- view of the cognitive approach. In J. F. Dovidio & s. L. Gaertner (Eds.), Prej- udice, discrimination, and racism (pp. 127- 163). Orlando, FL: Academic Press.

Isen, A. M., & Daubman, K. A. (1984). The influence of affect on categorization. Journal of Personality and Social Psychology, 47, 1206- 12 17.

Katz, I. (198 I ) . Stigma: A socialpsychological analysis. Hillsdale, NJ: Lawrence Erlbaum.

Lewis, C. A. (1996). Convergent validity of the Depression-Happiness scale with the Crown-Crisp Experiential Index. Psychological Reports, 78,497-498.

Lewis, C. A., & Joseph, S. (1995). Convergent validity of the Depression-Happi- ness scale with measures of happiness and satisfaction with life. Psychologi- cal Reports, 76,876-878.

Lippman, W. (1922). Public opinion. Forth Worth, TX: Harcourt, Brace, and Company.

Mackie, D. M., & Worth, L. T. (1989). Processing deficits and the mediation of positive affect in persuasion. Journal of Personality and Social Psychology,

McConahay, J. B. (1986). Modem racism, ambivalence, and the Modem Racism scale. In J. F. Dovidio & S. L. Gaertner (Eds.), Prejudice, discrimination, and racism (pp. 91-125). New York, NY: Academic Press.

McConahay, J. B., Hardee, B. B., & Batts, V. (1981). Has racism declined in America? It depends on who is asking and what is asked. Journal of Conflict Resolution, 25(4), 137-138.

OQ, 66, 62 1-632.

57, 27-40.

832 DEAN ET AL.

McConahay, J. B., & Hough, J . C. (1976). Symbolic racism. Journal oj'Social Issues, 32,23-45.

McGreal, R., & Joseph, S. ( 1993). The Depression-Happiness scale. Psychologi- cal Reporrs, 73, 1279- 1282.

Pettigrew, T. G., & Martin, J. (1987). Shaping the organizational context for Black-American inclusion. Journal of Social Issues, 43,4 1-78.

Rossi, P. H., & Anderson, A. B. (1982). The factorial survey approach: An intro- duction. In P. H. Rossi & S. L. Nock (Eds.), Measuring socialjudgments: The factorial survey approach (pp. 15-67). Beverly Hills, CA: Sage.

Sherif, M., & Sherif, C . W. (1953). Groups in harmony and tension: An integration of studies on intergroup relations. New York, NY: Harper & Brothers.

Tajfel, H. (1969). Cognitive aspects of prejudice. Journal of Social ISSZICY, 25, 79-97.

Taylor, S. E. ( 198 1). A categorization approach to stereotyping. In D. L. Hamil- ton (Ed.), Cognitive processes in stereotyping and intergroup behavior (pp. 83-1 12). Hillsdale, NJ: Lawrence Erlbaum.

Walsh, J., Joseph, S., & Lewis, C. A. (1995). Internal reliability and conl'ergent validity of the Depression-Happiness scale with the General Health Ques- tionnaire in an employed adult sample. Psychologicul Reports, 76, 137- 138.