Under Federal and State laws, a defendant who is charged with more than one offense may be tried for all the offenses in a 'joined trial.' A review of legal theories, research on jury decisionmaking, and current theories of memory led investigators at the University of Wisconsin to the hypothesis that joinder would increase the likelihood that a defendant would be convicted. The subjects were 68 male and 84 female undergraduates who, acting as mock jurors, read trial summaries and rendered verdicts on three types of cases: rape, sexual assault (touching the victim's breast), and trespass. Some subjects judged single offenses while others judged trials with joined offenses. The subjects were asked to make judgments about each of the cases: the probability that the defendant committed the offense; the verdict; confidence in the verdict; years of prison or probation appropriate for guilty defendants; and the seriousness of the criminal acts. In addition, some subjects were asked to recall facts about their cases. The results supported the hypothesis. Instructions to consider cases separately were not effective in removing the bias produced by the joining of the cases. In addition, recall errors occurred more frequently in joined than in separate trials, and these were in the form of intrusion of facts from other charges, suggesting that the confusion of evidence is at least one explanation for the results. It is recommended that a defendant required by the court to 'prove' prejudice resulting from joinder might want to consult a psychological researcher. Related literature is reviewed. Data tables and charts and a 35-item reference list are included. (Author abstract modified)