Recent studies with nonclinical populations have yielded widely varying estimates of the prevalence of child sexual abuse. This paper focuses on four representative studies, describing how differences in methodology and sample characteristics may contribute to the variation in prevalence rates. Two aspects of data collection appear to be the most significant factors accounting for discrepant findings. First, the use of face-to-face interviews is associated with much higher prevalence rates than the use of self-administered questionnaires. Second, higher prevalence rates are reported in studies that use multiple questions to ask about specific types of abusive sexual behavior. Variation in the age range of subjects is an additional factor that may also have an effect on prevalence rates. The evidence reviewed points to several aspects of methodology and sample characteristics that do not seem to influence prevalence findings. These are the use of random sampling techniques, the area of the country in which the study is conducted, the educational level of subjects, and the ethnic composition of the sample.