The limitations of volunteer panels of Internet users are well documented (e.g., self-selection bias due to the fact that participants recruit themselves into such panels). To compensate these biases in online samples, for instance with the aid of propensity weighting or propensity score adjustments, a serious academic discussion about their pros and cons is currently under way.
As an alternative to volunteer panels of Internet-users, pre-recruited panels of Internet users seem most promising to overcome some limitations. In contrast to the volunteer panels, these pre-recruited panels are probability-based, because the panel members are initially recruited by using probability sampling methods such as RDD telephone surveys. However, a major disadvantage here is the attrition in the respondent´s pool, because potential respondents go through different stages before being sampled to participate in a specific online survey. In general, at least four preceding stages can be identified: (1) participation in a telephone survey aimed at recruiting potential respondends, (2) acceptance given to sign up for the panel, (3) willingness to become an active panel member, and (4) willingness to respondent in a specific survey. In view of the usual attrition rates on each of these different stages, the overall response rate can be less than 10 percent, taking the validity of the results from pre-recruited panels of Internet users into question.