Résumé : Whether people make the right choice when they vote for a given candidate or party and what factors affect the capacity to vote correctly have been recurrent questions in the political science literature. This paper contributes to this debate by looking at how the complexity of the electoral context affects voters’ capacity to vote correctly. Correct voting is defined as a vote that maximizes one’s payoffs in lab elections with monetary incentives. We examine two aspects of the electoral context: district magnitude and the distribution of preferences within the electorate. The main finding is that the frequency of correct voting is much higher in single-member than in multi-member district elections. As soon as there is more than one single seat to be allocated, voters have more difficulty figuring out whether they should vote sincerely for their preferred party or opt strategically for another party in order to maximize their payoffs. By contrast, the distribution of preferences within the electorate has no significant effect.