Since the early days of Web surveying it was recognized as a potential advantage of computer mediated surveying to enrich studies with multimedia stimuli such as graphics, pictures, spoken word or other sounds. But in fact, these possibilities have only seldom put into action. Just in recent years Web surveys have been enriched by graphics and pictures – some of which are content bearing. Methodological evaluations have shown that these pictures can have a serious impact on the perceived question meaning and thus on the responses provided. In this paper we will report preliminary results from an assessment of the next step towards a full audiovisual Web survey. In spring of 2007 we have conducted a field experimental randomized comparison of a traditional text-based Web surveys to a Web survey containing videos of an interviewer reading the questions to the respondents. In the paper, unit nonresponse, social desirability and social presence are evaluated. In addition, we assess response times as well as the social setting and the location in which the survey was filled in. Finally we will consider ideas on how the underlying question answer process can be examined in greater detail.