In Study 1, participants listened to an audiotaped description of a burglary (Stage 1). After a 10-minute filler task, they completed a questionnaire containing several critical questions (Stage 2). Some questions contained misleading information that is typical of burglary, and some contained misleading information that is atypical of burglary; some contained no misleading information. After another 10-minute filler task, participants completed a questionnaire that measured the effects of the misleading information on their memory reports about the crime. Participants were 91 introductory psychology students at a Midwestern research university. Study 2 explored the frequency and typicality of memory intrusion errors. Participants listened to the audiotaped burglary from Study 1, did an unrelated 10-minute filler task, and then completed the Stage 3 questionnaire, which assessed the accuracy of their memory reports for the critical items. They did not complete Stage 2 of the previous experiment and therefore did not receive any misleading information. Participants were 18 introductory psychology students at a Midwestern research university. Study 3 had some variations from the first two studies. Consistent with the expectancy hypothesis, Study 1 showed that people are more susceptible to misleading information that is perceived to be typical of the crime in question than misleading information that is atypical. Studies 2 and 3 showed that people's intrusion errors also contain more typical than atypical information. These results indicate that people's prior knowledge of crime categories does provide an expectancy for fact-finding. Thus, prior knowledge plays a broader role than previously appreciated. 2 tables and 18 references