The subjects were 116 randomly selected citizens who were eligible for jury duty. Each viewed a simulated voir dire conducted by an actor posing as a judge or an attorney and using a personal or formal interpersonal style. The subjects were told that they would be taking part in a mock trial, but they were led to believe that the judge and the attorneys were authentic. They completed an attitude scale containing 29 statements regarding legal issues related to voir dire and designed especially for the present study. The questionnaire used a 10-point Likert-type scale. Results from the questionnaires were compared with the subjects' verbal replies during the voir dire. The subjects changed their answers almost twice as much when questioned by a judge as when interviewed by an attorney. The judge's interpersonal style did not improve the results. Females distorted their replies to a greater degree than males did regardless of who conducted the voir dire or how they behaved. The distortion of replies occurred in all age, income, and occupational groups. Thus, jurors are not as candid as sometimes presumed. Data tables and 30 references.