Discussing women’s numerical underrepresentation involves a kind of blame game in regard to the electoral phase: Is it the nominating political parties or the participating voters who are discriminating against female candidates? Empirical research has typically dealt with the question by comparing net effects of the two processes. The effects, however, are hard to disentangle empirically and findings are generally inconclusive. In this paper we present a new take on the question about voter discrimination by studying individual level gendered preferential voting. We claim that the minor net effects often observed empirically hide more substantial differences between individuals and districts. These differences are important to uncover in order to answer the question about when and who we should expect to discriminate against female candidates. The research thus aims to enhance our knowledge of voter discrimination and under what conditions this process is particularly important.
