This analysis examined the importance of differential exposure to infected partners in epidemiologic studies of latex condom effectiveness for prevention of sexually transmitted infections. Cross-sectional, enrollment visit data were analyzed from Project RESPECT, a trial of counseling interventions conducted at five publicly funded US sexually transmitted disease clinics between 1993 and 1997. The association between consistent condom use in the previous 3 months and prevalent gonorrhea and chlamydia (Gc/Ct) was compared between participants known to have infected partners and participants whose partner infection status was unknown. Among 429 participants with known Gc/Ct exposure, consistent condom use was associated with a significant reduction in prevalent gonorrhea and chlamydia (30% vs. 43%; adjusted prevalence odds ratio = 0.42, 95% confidence interval: 0.18, 0.99). Among 4,314 participants with unknown Gc/Ct exposure, consistent condom use was associated with a lower reduction in prevalent gonorrhea and chlamydia (24% vs. 25%; adjusted prevalence odds ratio = 0.82, 95% confidence interval: 0.66, 1.01). The number of unprotected sex acts was significantly associated with infection when exposure was known (p for trend <0.01) but not when exposure was unknown (p for trend = 0.73). Restricting analyses to participants with known exposure to infected partners provides a feasible and efficient mechanism for reducing confounding from differential exposure to infected partners in condom effectiveness studies.