Objectives: To compare smokers recruited by mail or through the Internet. 
Methods: A questionnaire was mailed to 19,352 inhabitants of Switzerland in 1998, in an effort to enroll them in a smoking cessation trial. The same questionnaire was also available on the Internet. Furthermore, we mailed a survey to a representative sample (n = 1000) of the population of Geneva, Switzerland, in 1996. In this study, we compare three groups: 1027 smokers recruited through the Internet, 2961 volunteer trial participants recruited by mail (response rate 16%), and 211 smokers in the representative sample also recruited by mail (response rate 75%). 
Results: Smokers self-recruited through the Internet were younger, more educated, more motivated to quit smoking and smoked more cigarettes per day than smokers in the other samples. Compared to trial participants, Internet participants had more negative attitudes towards smoking, higher self-efficacy scores, and were more addicted to tobacco. The strength of associations between smoking-related variables was similar in Internet and trial participants. 
Conclusion:> As expected, the three groups of smokers differed on several characteristics. However, bias in distributions of variables did not imply bias in associations between variables. Thus, Internet recruitment is a potentially useful method for analytical studies that focus on associations between variables