Web surveys have the adventage that many visual possibilities can be incosporated in the survey. One possibility is to add pictures to the survey questions. These pictures may increase respondent′s motivation and could help clarify questions to define the subject of interest. Question text and accompanying pictures may be processed in parallel and interference effects can occur if the verbal and visual language is inconsistent with each other.

Couper, Tourangeau, and Kenyon (PQO 2004) show that presenting a picture of a high frequency instance enhances the retrieval of such instances and increase the total number of instances reported. For example, respondents who got an intimate restaurant picture reported they went out to eat less often compared to respondents who received a fast food picture. The effects of the pictures were also apparent in follow-up questions. Pictures can suggest either a broader or narrower scope than the text of the question conveys, especially in relatively broad, poorly defined questions.

We will report on an experiment designed to further explore the use of pictures (visual language). This experiment will be fielded in the LISS panel (JUNE/JULY 2009). Besides replicating earlier results results by Couper et al. (POQ 2004), we want to check whether effective questions, the effect of pictures should diminish if verbal language is more powerful than visual language. To check this, we also investigate the effect of counteracting visual and verbal language.