Revising the Factor Structure of the Juror Bias Scale: A Method for the Empirical Validation of Theoretical Constructs

The application of factor analytic techniques to explore the construct and predictive validity of a popular scale used for the identification of pretrial juror bias is herein reported. A confirmatory factor analysis (CFA) was employed on the Juror Bias Scale (JBS) scores of 301 participants, but empirical findings did not support the theoretically derived single-factor scales of Probability of Commission and Reasonable Doubt. Empirically driven alternative models were generated using exploratory factor analysis. The JBS scores of an additional 301 participants were then employed to cross-validate the initial findings using nested modeling CFA. The empirical model achieved a significantly improved fit over the theoretical model and resulted in the elimination of approximately 30% of the original items with no attenuation in the scale's ability to predict juror verdicts. Moreover, a theoretical reorganization of the items was consistent with the empirically derived model and provided a rationale for altering the scoring of the JBS which, in turn, maximized its predictive validity. The use of CFA techniques to aid in the development of scales assessing jury attitudes and biases is discussed.