There are many empirical studies that show lower response rates for Web surveys compared to other data collection methods. One explanation for these differences in response rates may be due to the mode of invitation, in particular invitation by email to do Web surveys: More and more people are deleting emails without opening them. The reasons for this could be the increasing number of emails people receive, and the increasing number of spam emails. One possible solution for this problem may be to use an invitation letter by mail, because mail is better for gaining attention and legitimizing the survey. On the other hand, an email invitation is more user-friendly for delivering the URL, providing a password, and getting invited people to the Web site with the survey. Against this backdrop in 2011 an experiment was conducted to test which mode of invitation may be most effective to increase Web survey response. There were two factors: invitation mode (mail vs. email) and prenotification by a mailed letter (prenotification vs. no prenotification). Both factors were fully crossed, yielding a 2 x2 experimental design. In the study we used the German General Social Survey (ALLBUS) as a sampling frame. Those ALLBUS respondents with Internet access who were willing to participate in a follow-up survey and gave us their email address (N = 667) were randomly assigned to one of the four different experimental groups. The dependent variable was participation in the online follow-up survey. In addition, we invited a random subsample of those Internet users who were not willing to provide their email address (N = 400) to do the online follow-up survey. These people were invited only by mail, one half with prenotification and the other half without a prenotification. We will report first results and discuss their implications for sampling procedures in Web surveys.