Die Erforschung der Wirkung politischer Meinungsumfragen: Lehren vom ‚Sonderfall‘ Schweiz

This paper summarizes findings from a Swiss project about the impact of polling. Using data from experimental designs, cross-section and panel surveys as well as a meta-analysis, we present data which are of interest for comparative research as well. First, we address the question of the nature of effects and specifically distinguish between direct and indirect effects. Second, we show that the framework of bandwagon and underdog effects is lacking precision in the multi-party system. Third, it is shown that the theoretical basis of the various stages of information processing is helpful and quite powerful for analysis. Fourth, due to the empirical profile of people relying on polls, and our findings regarding the issue predisposition regulatory authorities might be eased. Nevertheless, for scientists some processes still need to be disentangled and indirect effects seem to be substantial enough to warrant further attention.