Online Sexual Activity (OSA) is an important and growing phenomenon. Prior research in this area has been criticized on methodological grounds. This study examines the reliability of Internet research regarding online sexual activities by comparing a selected random sample to a convenience sample. Participation in the selected random sample was limited to every 1,000th visitor to the MSNBC website in June 2000. Participation in the convenience sample was available to anyone with access to the Internet during the same time period. Most differences between these samples indicated that, relative to a selected random sample, a significantly greater proportion of Internet users in the convenience sample had a heavier involvement with OSA, including online sexual difficulties. We discuss the methodological and clinical implications of this finding.