This methodological investigation set out to examine the use of the Internet as a data collection tool: its potential for sampling gay and bisexual men and gathering data on sexual behaviour. The research took place in the context of the Internet and HIV study, an examination of high risk sexual behaviour among gay and bisexual men seeking sex on the Internet. The Internet and HIV study collected quantitative data using selfadministered questionnaires completed online or by pen-and-paper, and qualitative data using one-to-one interviews conducted face-to-face or online. Its design provides an invaluable opportunity for investigating Internet-based research methods. Methodological issues were explored through collection of primary data from respondents who had participated in the Internet and HIV study and secondary analysis of data collected for the study. The first area of exploration was the use of web surveys for sampling gay and bisexual men. The results indicated that participation in the Internet and HIV web survey was driven by an altruistic desire to contribute to research into sexual health promotion and HIV prevention. Drop out was most likely to occur in the early stages of the survey and minor differences in sub-group drop out were identified. Men who participated in the web survey were broadly similar to the MSM drawn from a national probability sample, but were more likely to participate in high risk sexual activity. Differences were more pronounced when the samples were restricted to London men. The second area of exploration was the effect of the Internet mode on data collection. The web survey reduced item nonresponse in comparison to the pen-and-paper survey but did not improve response to sensitive questions over and above this general mode effect. The results indicated that online synchronous interviews may be suited to a more structured format, given the difficulty of expressing complex ideas through typing in real time. It was concluded that the Internet has excellent potential for researching sexual behaviour among gay and bisexual men. It provides effective access to this population and more complete survey data than the traditional alternative. The application of online interviewing requires further exploration.