Abstract
One important limitation of previous research involving measures of bias is that they rely heavily on brief methods of conveying trial information (case summaries) and participants typically do not deliberate as juries before rendering verdicts. In the present study, 183 jury-eligible college students completed the Pretrial Juror Attitudes Questionnaire (PJAQ) and the Juror Bias Scale, watched a videotaped simulated armed robbery trial, and rendered verdicts both before and after deliberating as juries. The results indicate that the PJAQ, and especially the subscales of System confidence and Conviction proneness, significantly predicts pre-deliberation verdicts as well as verdict change (i.e. the shift in verdicts from pre- to post-deliberation). Moreover, the PJAQ provides incremental predictive validity over a more commonly used measure of pretrial bias (the Juror Bias Scale) for pre- and post-deliberation verdicts, as well as providing incremental predictive validity with respect to a juror's tendency to change his/her verdict as a function of the deliberation process. Specifically, bias is associated with a greater probability of verdict change, regardless as to whether that bias reflects a pro-prosecution or pro-defense position. The implications these results have for the importance of pretrial attitude on juror judgments and the deliberation process are discussed.
Notes
1. We would like to thank Larry Wrightsman for providing this videotape.
2. Obviously, post-deliberation verdicts are not independent and it is reasonable to statistically control for their non-independence. In the present study, as is the case in any deliberating jury, participants are encouraged to reach a unanimous verdict and this by definition results in non-independence. In fact, in the present research, the effect of the jury (i.e. group) was extremely large (F=8.17, p<0.0001) for post-deliberation verdicts, thereby making it impractical to predict any variance after removing the effects of the group. (Note: There was no significant effect for group when examining the verdict change variable.) However, the goal of a real deliberating jury is to achieve non-independence of judgment. So, comparing the predictive validity of two bias measures without controlling for group effects is justified as long as the analysis is confined to a comparison of the JBS and PJAQ. Because each participant received both the JBS and PJAQ, comparing the scales examines the effects of deliberation in a way that does not capitalize on reduced within-group variation. The rationale here is that regardless of the size of the effect of group, it would necessarily have to be equal for both measures of bias. This allows us to directly compare the predictive validity of the two instruments (PJAQ and JBS) for post-deliberation verdicts, without interpreting the significance of the absolute r 2 value. Importantly, the predictive ability of the PJAQ with respect to post-deliberation verdicts (i.e. the absolute r 2 value) was not interpreted due to the fact that post-deliberation verdicts are not statistically independent.
3. The PJAQ subscales, with the exception of Conviction proneness, did not predict guilt judgments at the jury level. This is to be expected as the PJAQ was designed to measure generalized juror bias. Guilt judgments at the jury level incorporate a host of other factors (e.g. propensity toward compliance and acquiescence on the part of jurors to conform to normative pressures, willingness to express their attitudes, etc.) that are not directly measured by the PJAQ, or any other measures of pretrial bias. Other research has investigated the deliberation process itself with a focus on the factors that predict a willingness to influence (Casteel & Lecci, Citation2008).