Judges are increasingly using “implicit bias” instructions in jury trials in an effort to reduce the influence of jurors' biases on judgment. In this article, we report on findings from a large-scale mock jury study that tests the impact of implicit bias instructions on judgment in a case where defendant race was varied (Black or White). Using an experimental design, we collected and analyzed quantitative and qualitative data at the individual and group levels obtained from 120 small groups who viewed a simulated federal drug conspiracy trial and then deliberated to determine a verdict. We find that while participants were sensitized to the importance of being unbiased, implicit bias instructions had no measurable impact on verdict outcomes relative to the standard instructions. Our analysis of the deliberations, however, reveals that those who heard the implicit bias instructions were more likely to discuss the issue of bias, potentially with both ameliorative and harmful effects on the defendant. Most significantly, we identified multiple instances where, in an effort to avoid bias, participants who heard the implicit bias instructions interfered with their own or other participants' appropriate assessments of witness credibility.