Some types of judicial decisions requiring an evaluation of the prejudicial effect of evidence are (1) prospective determinations regarding the admission of evidence and the effect of inadmissible evidence, (2) retrospective determinations regarding the effect of evidence, and (3) the appellate review of judgments. The accuracy of such judicial judgments depends on the validity of three underlying assumptions. First, jurors must share a common view of the significance of matters brought to their attention; second, judges must generally accurately perceive the jurors' evaluation of information and arguments that arise during trial; and third, judges themselves must hold a common view of the jurors' perceptions of prejudicial evidence and argument. The validity of these assumptions was assessed with a pretest designed to measure the prejudicial significance of various items of stricken evidence in a particular case. Members of the community, the bar, and the bench were asked to evaluate, from a summarized trial, the prejudice associated with 40 evidentiary items. Four measures of agreement within and across respondent groups indicate a level of disagreement about the prejudicial effects of evidence that undermines the assumptions underlying the use of judicial discretion in this area. The study concludes with a discussion of the implications of the findings for the harmless-error doctrine, which holds that reversal of a trial verdict is proper only when the trial error has affected a 'substantial' right of the litigant. A total of 123 footnotes are provided.