Abstract
This study was conducted to investigate the relationship between expert witness likeability and jurors' judgments of credibility and tendencies in sentencing. Two actors playing expert witnesses were trained to present themselves as high and low in likeability in a standard testimony scenario in the sentencing phase of a capital murder trial. The effects of extraversion and gender of the 210 psychology undergraduates serving as mock jurors attending to expert testimony were also examined. The dependent variables were the jurors' perceptions of the witnesses' credibility and their agreement with the testimony. The likeability of the expert witnesses was found to be significantly related to the jurors' perception of their trustworthiness, but not to their displays of confidence or knowledge or to the mock jurors' sentencing decisions. The women rated highly likeable experts as more credible than the less likeable ones, but the men did not differentiate between the two types. As extraversion increased, the male jurors' agreement with testimony increased, but the female jurors' agreement decreased. The results suggest that likeability can be an important element in the credibility of the source and that attorneys and trial consultants now have an empirical foundation for addressing likeability as part of witness preparation.
A growing body of literature on expert testimony1,2 has described the need for study of the behavioral components associated with effective testimony. The purpose of the present study was to investigate one component, expert witness likeability, utilizing a theoretically derived framework for credibility. We begin by reviewing source credibility and the literature about perceiver variables related to source likeability.
Source Credibility
The topic of source credibility has been substantially discussed in the psychological literature. McCroskey and Young3,4 established much of the conceptual and empirical groundwork on the subject. In their 1981 discussion of the state of source credibility theory and research,4 they identified eight factor analysis-supported components of credibility: sociability as reflected in ratings of pleasantness, friendliness and warmth; extroversion as reflected in ratings of talkativeness, boldness and aggression; calmness as reflected in ratings of poise, relaxation, and calm; competence as reflected in ratings of expertise and intelligence; character as reflected in ratings of honesty and trustworthiness; and additional ratings of size (large-small), time (early-late), and weight (skinny-fat). The authors point out that size, time, and weight are part of general ways in which individuals perceive other people. They concluded that these eight components could be collapsed into two overarching domains of credibility: competence and character. These domains are similar to the more recent conceptions of knowledge and trustworthiness reported by Brodsky1 in the context of expert witness credibility.
Griffin and colleagues5 have identified four empirically supported domains of courtroom credibility: trustworthiness, knowledge, confidence, and likeability. The components of credibility in the courtroom may be more specific than the components of general credibility, as conceptualized by McCroskey and Young,3,4 due to the specific dynamics involved in courtroom testimony. Existing research has already demonstrated the significant and curvilinear relation of expert witness confidence to credibility.6 Medium levels of witness confidence were rated as most credible, followed by high and then low levels of confidence. In this study, we sought to examine the relation between expert witness likeability and mock juror judgments of credibility and their sentencing decisions.
Extraversion as a Moderator ofPerceived Likeability
Juror characteristics (e.g., gender, personality) are potentially useful in understanding their perceptions of witnesses. Juror extraversion is one such construct, and it has been linked to the likeability of expert witnesses.7–12 Just as extraverted jurors are more likely than introverted jurors to like expert witnesses, parallel findings have been found in the ratings of extraverted target persons. Extraverts have been found to be rated as more likeable than their introverted counterparts.7,10 Oltmanns and colleagues10 found in a rating of thin-slice behaviors that extraversion was positively related to ratings of likeability.
The relation between extraversion and likeability has been shown to be different between the genders.11,12 In 1986, Riggio and Friedman12 used three measures of extraversion: one measure from the Personality Research Form, one measure from the Eysenck Personality Inventory, and one subscale of the Self-Monitoring Scale. The intercorrelations (computed separately for women and men) of these scales were all statistically significant and were all related to perceived likeability. The extraverted men tended to display outwardly focused and fluid expressive behavior and were in turn judged more likeable than were the men who scored low on expressiveness and extraversion. The women who displayed more facial expressiveness drew more favorable initial impressions as rated by others. In a 1984 study, Riggio11 found that the most frequently chosen women in a mock video-dating service (that is, the most likeable) were those who were less extraverted and expressive than their counterparts. The collective drawback of these studies is that the extraversion-likeability link was limited to the same person, rather than looking at how extraversion on the part of the perceiver influences their judgment of the likeability of others.
Studies of extraversion on the part of the perceiver have implications for the present study: extraverted and introverted jurors may differentially perceive witnesses as likeable and credible. For instance, Nass and Lee13 investigated computer-synthesized speech and personality. In a two-part study, they looked at participants (extraverts or introverts) who heard a synthesized voice (extraverted or introverted) on a book-selling Web site. They found that participants accurately assessed personality cues in the synthesized voice and showed similarity of attraction in their evaluation of the computer voice, the book reviews, and the reviewer. The second part of the study added an element of personality to the previous design (e.g., “It is guaranteed to be in very excellent condition!” versus “It is in like-new condition.”), and the findings replicated those in part one. The authors concluded that to maximize likableness and trust, a computer personality should be created to be consistent with the user and with the content being presented. The results of this study suggest that an interaction between extraversion and likeability may exist between the personality of the juror and that of the expert witness.
Extraverted people, compared with their introverted counterparts, have been found to rate the likeability of target persons differently.14 Extraverted college students, compared with introverted ones, rated target persons described by unfavorable traits as less likable and target persons described by favorable traits as more likable. The differences in social responsiveness between the extraverts and the introverts could be related to two mechanisms. First, extraverts have a stronger need for stimulation and are thus more likely than introverts to interact with other persons.15 Second, as a result of this interaction, the extravert learns to be more responsive to the positive and/or negative reinforcement potential of other persons.14
Gender, Extraversion, and Perceptionsof Likeability
The gender of the juror appears to be another pertinent trait. There are gender-related differences in extraversion, such that women generally show higher levels of extraversion than do men.16–20 This finding is consistent across a variety of different personality measures, including the NEO (Neuroticism-Extraversion-Openness) Five-Factor Personality Inventory,16,20 the Eysenck Personality Questionnaire,19 and items from the International Personality Item Pool.18 That females are more extraverted than males has been shown in both adolescents and adults, and the difference seems to be stable across the lifespan.16,20 From a meta-analytic perspective, Feingold17 examined the norms from 13 personality inventories that included 36 independent normative samples and found that the women tended to display higher levels of gregariousness (a facet of extraversion) than did the men in 19 of the 36 samples (Cohen's d, 0.09–0.76). Given the gender differences in extraversion, part of our objective was to study how the juror's gender affects his or her perceptions.
No studies were located in which differences were examined between men and women in perception of source likeability. Thus, the present study may address a new area for investigation. However, women have rated child witnesses as more credible than men have.21,22 If this pattern holds with adult witnesses, the juror's gender may influence perceptions of expert witness credibility.
The Present Study
Credibility research has shown that the confidence displayed by expert witnesses is a key factor in jurors' perceptions of credibility and in their decision-making.6 In the present research, we sought to further the understanding of associations between the components of credibility described by Brodsky1 as they relate to each other, the juror's personality, and his or her decision-making. The degree of witness likeability may influence jurors' decisions, especially jurors who are highly extraverted. The scant literature on gender differences in perceptions of likeability raises the question of whether this juror characteristic is worthy of sustained empirical attention.
Hypotheses
-
Manipulated expert likeability will show a linear association with overall credibility, as well as with credibility subscales of perceived confidence, trust, and knowledge.
-
Manipulated expert likeability will show a linear association with mock juror ratings of sentencing recommendations, in that higher likeability will be associated with higher agreement with expert witness conclusions.
-
The extraversion of the juror will moderate the perception of expert witness credibility and the juror's sentencing decisions.
-
The gender of the juror will moderate the perception of expert witness credibility and the juror's sentencing decisions.
Methods
Procedure
Two actors were trained and then videotaped while they demonstrated high and low levels of expert likeability. Rehearsal feedback was given to shape successful manipulation of the likeability variable. Pilot studies were conducted to ensure successful manipulation of likeability and clarity of procedures and to avoid revealing the hypotheses to the participants. For the primary data collection, participants were apprised of their rights as research participants and then watched a randomly assigned condition of testimony. They then completed the questionnaire packet outlined herein. The procedures and protection of participants were approved by the Non-Medical Institutional Review Board, Office of Research Compliance, The University of Alabama, Tuscaloosa.
Defining Likeability
Expert witness likeability may be defined as the degree to which an expert is friendly, respectful, kind, well-mannered, and pleasant.5 However, to assess the effect of expert likeability empirically, we sought to define it behaviorally. Drawing on literature from a variety of sources, we identified the following list of verbal and nonverbal components associated with high likeability: a pleasant, smiling facial expression23; use of “we” or “us” when referring to groups24; demonstration of a less controlling attitude24; physical attractiveness25; use of deferential speech and considerate disagreement as opposed to aggressive, defiant contradiction26; a low degree of arrogance exhibited in verbal responses, such as acknowledging limited certainty of findings or a potential for error27; use of informal speech, such as referring to an individual by name and use of less technical jargon28; direct eye contact29; and truthfulness30 (suspicion of lying was negatively associated with likeability).
We concentrated our efforts on the variables most reflective of likeability that could be readily manipulated in the context of testifying. Therefore, we operationally defined likeability according to degree of smiling, use of we or us in reference to groups, absence of arrogant responses, and maintenance of good eye contact. The following criteria were used in manipulated conditions of high and low likeability:
High likeability: consistent use of we or us when discussing members of the scientific community or humanity as a whole, moderate levels of smiling, modest statements and conclusions (e.g. “relatively certain” or “we do not know everything there is to know in psychology”), consistent eye contact with lawyer and jury, and use of informal speech (i.e., little use of technical jargon and referral to parties in the courtroom by surname).
Low likeability: no use of we or us, no smiling, excessive display of certainty of conclusions, inconsistent eye contact, highly technical jargon, and frequent formal references to persons in the courtroom (e.g., “the client,” “the defendant”).
Pilot Study
Results of analysis of the pilot data showed generally that the manipulations were successful. Four conditions (low and high likeability with two different expert witnesses) were assessed to ensure differential ratings between conditions and equity in perceived likeability between actors. Results of an ANOVA (n = 44) showed that the overall model was significant (F(3,41) = 20.53, p < .001). Least significant difference post hoc analyses indicated that the manipulation was successful, as each low-likeability condition was rated significantly lower than each high-likeability condition (p < .001). When the actors were compared, neither the low (p = .46) nor the high (p = .71) conditions were significantly different from each other in likeability ratings. All but one participant indicated that the instructions were clear. (That person failed to answer the question.) In addition, mock juror ratings of other witness characteristics were collected to confirm the manipulation of likeability. Adjectives conceptually linked to likeability were selected based on a list adapted from earlier work.31 Results showed appropriate correlations of mock juror ratings of likeability with other constructs as follows: charm (r = 0.73, p < .001), friendliness (r = 0.73, p < .001), and conceit (r = −0.71, p < .001).
Participants
In the study proper, we enlisted 225 students in introductory psychology from a large public southeastern university. The stimulus materials involved expert testimony about dangerousness in a capital murder sentencing simulation. In accordance with the death-qualification criteria in Witherspoon v. State of Illinois,32 those who reported an absolute inability to assign the death penalty were excluded from data analysis to pursue verisimilitude. A total of 210 participants satisfied death penalty qualification criteria based on opinions expressed on a 10-point Likert scale, with higher values denoting increased support for the death penalty.
The mean age of the participants was 19.06 years (SD = 2.09). There were 59 men and 149 women; 2 did not identify their gender. A total of 97 participants viewed the low-likeability condition and 113 the high-likeability condition. The participants reported their religion as Christian (mostly Southern Baptist) (n = 114), Catholic (n = 41), Protestant (n = 31), Jewish (n = 3), agnostic (n = 5), atheist (n = 3), and other (n = 12); one person chose not to specify religion. Only four participants had served on a jury, and so this variable was not analyzed.
Materials
Demographics
The participants completed a demographic form inquiring about age, sex, ethnicity, religious orientation, attitudes toward the death penalty (10-point Likert scale), and previous experience as a juror.
Manipulated Likeability
Two levels (low and high) of likeability were manipulated with a scenario based on the scripts in Krauss and Sales33 that depict a state-hired expert witness testifying under direct and cross-examination about the recidivism potential of a convicted murderer. The only different content between the testimonies were the manipulated verbal and nonverbal likeability behaviors defined earlier. These conditions were presented in videotaped format.
Two male actors of similar age and credentials were used in the videotaped scenarios. Both actors presented themselves as tall, bearded, male professors at a major university. All scripts held the following psychologist credentials constant: licensed clinical psychologist, an established private psychotherapy practice, 14 years of experience in psycho-legal evaluations (more than 100 risk prediction assessments), and testimony in over 50 cases.
Expert Credibility
The Witness Credibility Scale5 was used to assess credibility. The scale consists of 20 bipolar adjectives on a 10-point Likert scale, in which higher values denote increasing agreement with the adjectives. Each of the four subscales comprises five items. The α coefficients have been reported for each subscale as follows: confidence (0.88), likeability (0.86), trustworthiness (0.93), knowledge (0.86), and overall credibility (0.95).5 The likeability subscale was eliminated from the analyses to avoid conceptual overlap between the independent variables of behavioral likeability and the criterion measure of credibility.
Sentencing Recommendation
Items scored on a 10-point Likert scale were used for the mock jurors' ratings of the likelihood of assigning the death penalty or life without parole. Because the mock expert provided testimony suggesting that the convicted criminal posed a continuing danger to society, higher values reflected agreement with the observed testimony. Thus, higher likelihood of assigning the death penalty reflected agreement with testimony, while life without parole did not.
Juror Extraversion
Extraversion was assessed with Goldberg's Five-Factor Items.34,35 The scale comprises 50 statements, each of which is rated on a five-point Likert scale. Each of the Five-Factor Model domains, neuroticism, extraversion, openness, agreeableness, and conscientiousness, is assessed with 10 items from the scale. Cronbach's α for the reliability of the domains has been reported: neuroticism (0.86), extraversion (0.87), openness (0.84), agreeableness (0.82), and conscientiousness (0.79).34,35 Extraversion was the only subscale of interest in the present study.
Results
Effects of Expert Witness Likeability
Independent samples t tests were used to assess the impact of expert likeability on dependent measures. The first hypothesis was that expert likeability influences juror perceptions of credibility in a linear manner, and it was supported by the results. The findings showed that highly likeable witnesses (mean [M] = 120.12, SD = 20.61) were rated higher in overall credibility than their less likeable counterparts (M = 112.11, SD = 21.54; t(205) = −2.74, p = .007). The impact of likeability on perceptions of credibility is clarified by examining differences in the subscale responses. Highly likeable experts (M = 37.60, SD = 8.41) were rated as more trustworthy than their less likeable counterparts (M = 29.79, SD = 10.14; t(206) = −6.06, p < .001). There was no significant main effect of expert likeability on juror perceptions of knowledge (t(207) = −0.72, p = NS) or confidence (t(206) = 0.14, p = NS).
The second hypothesis, which predicted that expert likeability would directly affect juror sentencing decisions, was not supported. Independent samples t tests showed no significant main effect of expert likeability on assignment of the death penalty (t(208) = −0.88, p = NS) or life without parole (t(208) = 0.72, p = NS).
Moderation Analyses
Custom general linear modeling (GLM) was used for all moderation analyses. All continuous predictor variables were standardized. Participant support for the death penalty was included as a covariate in all moderation models, to obtain the most comprehensive predictive model of dependent measures. Thus, each predictive model featured support for the death penalty, likeability conditions, juror gender, and juror extraversion. All two- and three-way interaction terms were included in moderation analyses to clarify any main effects of moderators. One model was run for each dependent measure: total expert witness credibility, likelihood of assigning the death penalty, and likelihood of assigning life without parole.
The overall model predicting total credibility was significant (F(8,192) = 2.03, p = .05; adjusted R2 = 0.04). No significant main effects emerged; however, there was a significant two-way interaction between level of likeability and juror gender (F(1,192) = 5.22, p = 0.02). While the men showed stable ratings of expert credibility, the women rated high-likeability witnesses as more credible than low-likeability witnesses. Figure 1 depicts this moderation.
Two-way interaction (expert likeability by juror gender) on overall credibility. (In all graphs, constructs on the x-axis are categorical variables. Where gender is represented on the x-axis, gender categories (i.e., male, female) are discrete but do not increase.)
The overall model predicting likelihood of assigning the death penalty was significant (F(8,195) = 6.19, p < .001; adjusted R2 = 0.17). The only significant main effect that emerged was support for the death penalty (F(1,195) = 39.05, p < .001). A significant two-way interaction between juror gender and extraversion was found (F(1,195) = 5.11, p = .03). For male jurors, the likelihood of assigning the death penalty increased as extraversion increased. However, for female jurors, the likelihood of assigning the death penalty decreased as extraversion increased. Figure 2 depicts this interaction. Also, a significant trend emerged for the three-way interaction between level of behavioral likeability, juror gender, and juror extraversion (F(1,195) = 3.54, p = .06). At low levels of juror extraversion, the women consistently showed a higher likelihood of assigning the death penalty than did their male counterparts in response to low-likeability testimony. However, at high levels of juror extraversion, the men were more likely than the women to assign the death penalty based on low-likeability testimony. This pattern evened out with high-likeability testimony.
Two-way interaction (juror gender by juror extraversion) on likelihood of assigning the death penalty.
The overall model for likelihood of assigning life without parole was significant (F(8,195) = 2.35, p = .02; adjusted R2 = .05). The only significant main effect was support of the death penalty (F(1,195) = 12.88, p < .001). A significant trend emerged for the interaction between level of expert likeability and juror gender (F(1,195) = 3.57, p = .06). The male jurors were more likely to assign life without parole in the low-likeability than in the high-likeability condition. The women showed a stable probability of assigning life without parole. Figure 3 shows this trend.
Two-way interaction trend (expert likeability by juror gender) on likelihood of assigning life without parole.
Discussion
The primary purpose of this study was to continue the investigation of the relation of the four components of witness credibility in the courtroom as conceptualized by Brodsky1: confidence, likeability, trustworthiness, and knowledge. Prior research has shown confidence to be a key factor associated with witness credibility and juror decision-making.6 The present study extended the line of research, finding that the likeability of the expert witness is positively associated with credibility. The results of the study revealed no significant main effect of witness likeability on knowledge or confidence; however, there was a main effect of likeability on trustworthiness, in which likeability was positively associated with trustworthiness. The likeability of the expert witness was not associated with juror decision-making in this study. It may be that likeability influences verdicts in non-death penalty cases, but is less important in jury decisions pertaining to death sentencing. The seriousness of capital murder charges and the possible sentences may well demand a greater focus by actual and mock jurors alike, so that the central processing of probative content is more common and compelling than the peripheral processing in which likeability would play a role.
From a theoretical standpoint, the credibility and likeability constructs are partially clarified in the specific setting of testimony. Perceptions of likeability directly influenced the jurors' trust, but not their decisions. This finding appears to be a new contribution to understanding believability of expert witnesses. Overall, likeability is a construct dependent on the influence of other individual difference factors (e.g., gender) in relation to decision-making. When credibility is examined, the present results combined with those of Cramer and colleagues6 highlight trustworthiness as a pivotal facet of expert witness credibility. Juror perceptions of both confidence and likeability have been shown to influence the trustworthiness component. Extrapolation of these findings suggests that being perceived as likeable and/or confident engenders trust. A possible implication is that trust is a factor that influences behavioral outcomes such as decision-making in the courtroom context. The role of trustworthiness as a determinant of trial outcomes should be further investigated.
Gender and extraversion were shown to be important individual difference factors in jury decision-making. For example, results from the model predicting overall credibility found that while the men showed stable ratings of expert credibility across the conditions of expert likeability, the women rated witnesses with high likeability as significantly more credible than witnesses who were not likeable. This finding may be interpreted through Tannen's conceptualization of distinctive gender communication patterns.36 Tannen argued that men and women may differentially perceive the same verbal behavior due to different communication structures and purposes. Her theory holds that, in general, men use fact-based conversation to uphold a hierarchical social world order, to avoid failure and preserve independence. In contrast, women's communication generally serves the primary purposes of building connections, providing support, and ultimately warding off social isolation.
A 1988 survey of American adults by Kroeger and Thuesen,37 utilizing the Myers-Briggs Type Indicator, lends some support to Tannen's assertions. The researchers found that roughly two-thirds of American men prefer making decisions based on logic and rational thought, and that about two-thirds of American women prefer making emotion-based decisions. The finding that the women's, but not the men's, rating of credibility was affected by witness likeability can be interpreted through the more emotional purposes of communication for the genders, as presented by Tannen.36
Apart from witness likeability, juror gender and extraversion interacted to predict decisions. Extraversion and agreement with testimony correlated positively in the male jurors who assigned the death penalty; the female jurors exhibited the opposite pattern. A three-way trend emerged between these factors and the degree of expert witness likeability. Although the level of behavioral likeability offers a situational explanation, there is more to the picture on gender, extraversion, and subsequent decisions. Previous research shows consistent gender disparities in extraversion, in that women are more extraverted than men.16–18 Our findings begin to illuminate how such personality differences manifest in other behaviors and attitudes. From a five-factor perspective,38 highly extraverted persons seek out excitement and positive emotions, engage the social world often, and are gregarious. High degrees of this trait in men appear to result in an agreeable stance toward expert testimony, at least on the backdrop of a capital murder trial. For female jurors, less extraversion may lead to less agreement, based on the present results. Our basic two-way interaction finding allows only for speculation as to the cause of differential agreement, because we did not measure traditional subcomponents of extraversion such as excitement-seeking and gregariousness. A future step in ferreting out the link between gender differences in personality and resulting agreement with testimony would be to examine how various components of extraversion may mitigate decisions by gender.
There are limitations to the current study, including that the testimony was presented via videotape rather than live in a courtroom and that jury deliberations were not included in the study. Moreover, we used only men as mock experts. The findings of the present investigation should be examined with women acting as witnesses as well. In addition, our sample was composed of primarily Caucasian undergraduate students. However, it should be noted that research has found few differences between various trial media or mock juror samples. Bornstein39 and Dunn40 separately concluded that using students as mock jurors in jury-simulation research is not necessarily a cause for concern. Finally, the collection of behaviors conceived of as likeable may be a weakness, because the present design fails to identify which particular behaviors are related to credibility. At the same time the study did feature a comprehensive reflection of behavioral likeability.
The implications of the results in the present study may apply to trial consultation practice. For instance, juror gender and extraversion can be used in the jury selection process as markers for voir dire questions and questionnaire items. Furthermore, trial consultants and attorneys may seek to shape behaviors of experts during witness preparation explicitly to promote trust in the expert's testimony.
- American Academy of Psychiatry and the Law
