Public opinion polls have become important and increasingly visible elements of representative democracies. Especially during election campaigns, opinion polls reflecting citizens' vote intentions attract substantial media attention. In this regard, previous research has frequently shown that polls can influence both citizens' vote intentions as well as political parties' campaign strategies.

At the same time, opinion polls are fraught with great uncertainty. One way to reflect this uncertainty is to report confidence intervals. This paper investigates how citizens change their vote intention dependent on whether polling estimates are presented with or without confidence intervals. Using a vignette experiment (N=3000), we examine this question based on a real-world example in which different election polls were shown to respondents ahead of the 2021 federal election in Germany. We manipulated the display of confidence intervals and journalists' interpretation of the polls. In this way, we estimate the average causal effect of confidence intervals on citizens' vote intentions.

The findings of this study are important for two main reasons. First, they will help understand the extent to which confidence intervals in media reports can help citizens make more informed (strategic) vote choices. They will thus shed light on whether depicting opinion polls’ uncertainty can affect key features of representative democracy, such as democratic accountability. Second, this paper's findings will also contribute to the broader debate on the potential benefits of confidence intervals for communicating scientific research results to the broader public.