Computer-assisted interview methods are increasingly popular in the assessment of sensitive behaviors (e.g., substance abuse and sexual behaviors). It has been suggested that the effect of social desirability is diminished when answering via computer, as compared to an interviewer-administered face-to-face (FTF) interview, although studies exploring this hypothesis among adolescents are rare and yield inconsistent findings. This study compared two interview modes among a sample of urban, ethnic-minority, perinatally HIV-exposed U.S. youth (baseline = 148 HIV+, 126 HIV-, ages 9-16 years; follow-up = 120 HIV+, 110 HIV-, ages 10-19 years). Participants were randomly assigned to receive a sexual behavior interview via either Audio Computer-Assisted Self-Interview (ACASI) or FTF interview. The prevalence of several sexual behaviors and participants' reactions to the interviews were compared. Although higher rates of sexual behaviors were typically reported in the ACASI condition, the differences rarely reached statistical significance, even when limited to demographic subgroups--except for gender. Boys were significantly more likely to report several sexual behaviors in the ACASI condition compared to FTF, whereas among girls no significant differences were found between the two conditions. ACASI-assigned youth rated the interview process as easier and more enjoyable than did FTF-assigned youth, and this was fairly consistent across subgroup analyses as well. We conclude that these more positive reactions to the ACASI interview give that methodology a slight advantage, and boys may disclose more sexual behavior when using computer-assisted interviews.