An experimental test of the pivotal voter model under plurality and PR elections

•We perform experiments on turnout to replicate the findings of Duffy and Tavits (2008).•We find a strong link between perception of being decisive over the outcome and turnout.•Overestimation of the probability of being decisive partly explains the high turnout rates.•We find no difference between PR and plurality elections.•Risk aversion increases the impact of perceptions.

This paper examines the decision to vote or not vote in experimental elections. We replicate the important findings of Duffy and Tavits (2008), though with a different design. Our results support their finding, that is, turnout is affected by the belief that one's vote counts and that overestimation in the perception that one's vote counts does not disappear completely over time. Going beyond previous research, we also find that turnout is not higher under a proportional system than under a plurality system, as well as that beliefs about being in a pivotal position have a greater impact on turnout among the risk averse.