A total of 244 female college undergraduates participated in this study. The trial summary presented to the participants described a criminal case of child sexual abuse in which the 23-year-old complainant alleged abuse by a parent when the complainant was 10 and 12. The summary described the allegation as "memories of five specific incidents of sexual assault." The parent/defendant denied the allegation. The summary varied three factors among study participants: the complainant's testimony (recovered versus remembered but not previously told), the complainant's gender, and the defendant's gender. Between 27 and 40 participants were exposed to each version of the summary. The memory of the abuse was revealed in the course of the complainant's therapy for depression and bulimia. In the "recovered" testimony condition, memory of the abuse was described as having been repressed and first recovered during therapy. In the "remembered" testimony condition, the abuse was described as having been always remembered but not admitted to anyone until therapy. For each of these conditions, the complainant's and parent/defendant's genders were varied. Findings show that compared to a case in which the memory of the abuse had always been remembered, recovered memory testimony yielded lower perceptions of the defendant's culpability and higher perceptions of the defendant's credibility. In addition, the complainant with recovered memory was viewed as less credible and less likely to be telling the truth. These effects of testimony type were qualified by an interaction with complainant and defendant gender, such that testimony type exerted an effect in cases of alleged heterosexual but not homosexual abuse. These findings suggest that mock jurors' judgments in sexual abuse cases reflect their stereotypes about sexual abuse and expectations regarding the relative likelihood of repression of events in various circumstances. 2 tables, 2 figures, and 29 references