Judging Guilt: Implicit Evaluations of Defendants Predict Verdicts

Abstract

Cognitive jurisprudence research that has used indirect measures has mostly focused on how people’s implicit biases (e.g., race) predict people’s verdicts for an individual belonging to an out-group. We aim to expand on this work by investigating implicit evaluations’ predictive strength for verdicts using an impression formation approach. In three preregistered studies, we presented mock jurors with a fictional murder trial in which the evidence against a defendant was mixed: some witnesses provided testimony suggesting guilt, whereas others sowed doubt. In all studies, implicit evaluations of the defendant, operationalized by the Affect Misattribution Procedure (AMP) scores, uniquely predicted verdicts above and beyond explicit evaluations (Studies 1–3), the reason for evidence exclusion (Study 2), and demographics of the defendant (Study 3). These findings advance our understanding of implicit social cognition by demonstrating that implicit evaluations, operationalized by the AMP scores, can have predictive power in complex, ecologically rich contexts.