A web survey with a self-selected sample is compared with a face-to-face survey with a probability sample. Both of these surveys consist of a rather similar questionnaire that requests a vote on a TV program. The findings show that a big concern with the web survey is item non-responses while the face-to-face survey has less problems in this regards. The respondents of the web survey consist of male single and young persons. Most participants in this survey are interested watchers of the TV program. In addition, advertisements appreciably have a positive effect on participation in the web survey.