The growth of eHealth research has been both rapid and sustained over the past several years. As part of this growth, the Internet is increasingly being used, both to deliver online interventions, and to collect data from a wide variety of groups of people. This paper reviews some of the issues around the latter use of the Internet, and in particular, the use of web surveys to produce results intended to be generalizable to a broader population. The paper provides an overview of issues of sample representation and the challenges faced in using Internet surveys. Various methods of sample selection, and their strengths and weaknesses are reviewed, and the threats to valid inference are discussed.