PIP: It is not easy to collect valid and reliable data on sexual behavior, and despite the development of more sophisticated survey instruments since the 1980s, largely in response to the HIV/AIDS pandemic, many problems persist. Since the advent of AIDS, there is a greater need to obtain accurate data on sexual behavior. Findings are presented from a comparison of data collected in a questionnaire survey with those collected through in-depth interviews. 58 young men from Glasgow were interviewed twice at a year's interval, first as part of a general health and lifestyle survey and then in an in-depth interview focusing upon sex. Considerable discrepancies were observed between the 2 studies in reported sexual behavior and knowledge. Poor concordance was observed between the 2 studies on age and contraception at first intercourse and upon the number of sex partners at age 18 years. Moreover, there was almost no consistency between the 2 studies regarding participants' understanding of the notion of safer sex. The most likely explanations for the discrepancies between the studies are poor recall, inadequate understanding of the question, and differing presentational concerns, especially related to embarrassment. Poor recall and embarrassment are particularly possible explanations because they can account for discrepancies in both directions. While the evidence suggests that the in-depth interview data may have greater validity than the survey data, the analysis nonetheless identifies ways of improving survey methods on sex.