Voters Share Polls That Say What They Want to Hear: Experimental Evidence From Spain and the USA

Abstract

Abstract Pre-election polls are published continuously, particularly in the run-up to elections. Increasingly, polls are commissioned by the media and gain traction through their propagation on social media. But polls differ along many dimensions, including in their estimated vote shares and the corresponding portrayal of the state of public opinion. We rely on recent evidence to argue that these factors are likely to affect people’s willingness to share polls with their networks. To test this claim, we conduct two studies. In the first, we expose Spanish voters to a random selection of recent, real, 2023 general election vote intention polls. In the second, a conjoint analysis presents US voters with abstract, hypothetical 2024 presidential election polls. Across both studies, we find no evidence of any effect of polling firms, fieldwork dates, or sample sizes on intentions to share polls. Above all, our results suggest that the main factor consistently affecting voters’ proclivity to share polls is the result of the poll itself. Our findings are consistent with the literature that suggests that voters’ reception of poll results is usually driven by directional goals, and may have significant implications in light of the unique epistemic value of reposting on social media.