Enrolling and tracking cohorts for HIV vaccine efficacy trials requires that participants disclose behaviors that place them at risk for exposure to HIV. Brief screening procedures have been suggested for this purpose. In a previous study gay and bisexual men in three U.S. cities reported unprotected anal intercourse on a brief screening instrument. Screen reports were compared to subsequent in-depth, face-to-face interview data; 29% of the men who reported unprotected anal intercourse during the interview failed to disclose this behavior during screening. For recruitment into an HIV vaccine feasibility study at the same study sites, screening procedures were modified to encourage accurate reporting: to lessen stigma, low risk as well as high risk sexual behaviors were assessed, and screens were administered by trained study staff who presented it as a tool for understanding the gay community. Failures to disclose risk decreased to 18%, a rate that, while lower than in the previous study, remains high. Men less likely to disclose unprotected sex during the screen engaged in fewer high risk sexual behaviors, had more stringent norms regarding sexual safety, and were less identified with the gay community than were men who disclosed unprotected sex. Failure to disclose risk may have significant implications for participant selection and behavior tracking during vaccine trials. More systematic assessments that are sensitive to target communities may facilitate disclosure.