Publication Type: Symposium Paper
Abstract: The present paper extends research on witness testimony by investigating an outcome measure for witness preparation. This two phase study assesses psychometric properties of the Witness Self-Efficacy Scale (WSES; Cramer, Brodsky, DeCoster, & Neal, 2008). Phase I of the study featured mock defendants completing the WSES and construct validity measures, and then testifying about false accusations. Results largely supported expected construct validity relations and showed excellent reliability of the WSES. Phase II of the study will include mock jurors making observer ratings of videotaped testimony scenarios to obtain predictive validity for the WSES. Theoretical and practical implications are discussed.

Publication Type: Symposium Paper
Abstract: The Witness Credibility Scale (WCS) was designed to serve as an outcome measure of expert witness believability. No published measure exists for assessment of the credibility of expert witnesses. The current study addressed this gap by developing and cross-validating a measure of perceived witness effectiveness. We hypothesized that credibility was a product of four factors: “likeability”, “believability, “trustworthiness”, and “intelligence.” A 41-item measure initially was constructed based on successive iterations of ratings by a panel of judges, and was administered to 264 undergraduates. A factor analysis of the data yielded four factors: “knowledge”, “likeability”, “trustworthiness”, and “confidence.” The final version of the WCS used 20 adjectives with four subscales of five items, each subscale reflecting high loadings on the four factors. The scale was then tested in four additional studies in which the scale successfully differentiated between groups of experts in manipulated conditions. The empirical data from these studies permit a scholarly foundation for comparative outcome data in research investigations.

Publication Type: Poster
Review Method: Peer Reviewed
Abstract: Previous research suggests that jurors do not consider the witnessing conditions when evaluating the reliability of eyewitness identification evidence. We propose that this is due to backward reasoning. Specifically, if the witness makes an identification, and the witnessing conditions were poor, the juror may assume that the witness must possess unusually good powers of observation and memory. Further, we predict that this effect will be stronger when the witness has higher status or is similar to the juror. We examined these predictions in a mock-juror study.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Very little research in eyewitness memory has focused on intoxicated witnesses – a group that law enforcement interacts with frequently – and none has examined their confidence-accuracy relationship. We manipulated intoxication between participants: Participants were either given cranberry juice (control), cranberry juice with little alcohol (low-dose placebo) and cranberry juice with vodka (alcohol). Repeated intoxilyzer® measurements ensured that participants’ breath alcohol levels were between .06 and .09 while presented with a staged theft, a witness interview, and subsequently providing confidence ratings. Preliminary results suggest that intoxicated witnesses may be more confident than sober witnesses in their witness reports.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: When statistical evidence is offered in a litigation context, the result can be bad law and bad statistics. This paper analyzes how statistical evidence is used at the class certification phase of employment discrimination litigation, in cases challenging decentralized and highly discretionary management practices as discriminatory. Courts are reaching radically different conclusions about whether these cases involve core and relatively straightforward applications of prior caselaw, or whether they instead test the limits of the legal concepts of what an appropriate class action is. Disputes over the application of statistical evidence in these cases are deeply intertwined in the larger theoretical and legal debate about decentralized decisionmaking and subtle bias as sources of common “systemic” injury. Further, the institutional context of litigation determines how statistical evidence is presented by plaintiffs and defendants and evaluated by the courts. Through the use of simulated data, we show how defendants can and do introduce statistical evidence that has absolutely no relevance to the issues before the Court, yet is often persuasive in defeating class certification. We also address problems in how plaintiffs present their expert evidence in support of certification. Finally, we propose alternative ways for courts to handle these problems that better reflect the relevant legal framework and statistical principles.