Studies interested in the effects of electoral systems have to contend with a difficult problem. Ideally, to properly estimate these effects, one would wish to compare elections that are completely identical except for their electoral system. This ideal situation, however, never occurs in real life. To circumvent this limitation, we created electoral simulations that approximate this ideal situation. In the context of a general election, we asked respondents to take part in multiple simulated votes, each of them using different electoral rules. By studying the different choices of voters and different outcomes of these elections, we gain the ability to identify how electoral systems determine voting behavior. This design was used to identify the relative importance of psychological and mechanical effects on the vote as well as voting preferences for female political candidates.