We examined the interaction of testimonial consistency and witness group identity on mock jurors' judgments of witness effectiveness, probability that the defendant committed the crime, and verdict. In a 3 × 2 (Witness Group Identity × Testimonial Consistency) between-groups design, 180 mock jurors heard a trial of a person charged with assault. Although both variables affected judgments, group-identity effects were weak when testimony was characterized by inconsistencies, and they were stronger when testimony was internally consistent but ambiguous. The judgment patterns were consistent with predictions from Chaiken, Liberman, and Eagly's (1989) heuristic-systematic processing theory, suggesting that heuristic processing would bias systematic processing when the evidence was not decisive.