The children assigned to the participant group played with a man described as a babysitter and were interviewed about their experience 11 days later. Those in the observer group watched a videotape of a child in the participant condition and were also interviewed about their observations after 11 days. Children in the police condition played with the babysitter, but before their followup interview, they were told by a police officer that the babysitter may have done something bad. The results showed that, compared to the observer group, children in the participant group had better free recall accuracy concerning the events that had occurred and exhibited lowered suggestibility to misleading questions. Forensic context, as measured by the police condition subjects, led to increased error in free recall and elaborating comments to misleading questions. However, the forensic setting did lead to higher accuracy on an age- identification task and did not affect the children's accuracy in terms of answering abuse-related questions. Further research should focus on identifying and using ecologically valid procedures, the contexts that elicit the most accurate children's testimony. 3 tables and 33 references