Epidemiologic studies exploring risks for sexually transmitted diseases, including human immunodeficiency virus infection, typically rely on self-report of sexual behaviors. Estimates of the incidence and prevalence of sexual practices are important measures for assessment of behavioral interventions as well as for examining disease transmission. This study examined the degree of agreement within heterosexual couples reporting frequency and type of sexual behaviors, including condom use. Self-reports were obtained from 71 couples attending Baltimore sexually transmitted disease clinics in 1989-1990 regarding the number of days and number of episodes of three specific sexual practices (any type of sexual activity, vaginal sex, and vaginal sex with condom use) over a 30-day period. Paired t test analysis revealed both sexes to be very consistent in their reporting of recent sexual experiences. Multivariate analysis showed that agreement did not vary by socioeconomic status, by whether the partners were married to each other, or by age. These findings suggest that reliable information regarding sexual behaviors may be obtained from men or women.