A field experiment tested the effect of an Arizona civil jury reform that allows jurors to discuss evidence among themselves during the trial. Judges, jurors, attorneys, and litigants completed questionnaires in trials randomly assigned to either a Trial Discussions condition, in which jurors were permitted to discuss the evidence during trial, or a No Discussions condition, in which jurors were prohibited from discussing the evidence during trial according to traditional admonitions. Judicial agreement with jury verdicts did not differ between conditions. Permitting jurors to discuss the evidence did affect the degree of certainty that jurors reported about their preferences at the start of jury deliberations, the level of conflict on the jury, and the likelihood of reaching unanimity.