The (Un)Representative Preferences of Party Members. Do party members compile representative candidate lists?

Almost all over the world, women are underrepresented in the legislature. To explain the underrepresentation of women, scholars often study political recruitment: are there specific hurdles in the recruitment process that are harder for women to overcome than for men (or vice versa)? One of the most important stages in this recruitment process is the candidate selection procedure, in which political parties select which applicants will become candidates. This stage is of utmost importance, since in most political systems this selection often equals election. Therefore, in order to understand the underrepresentation of women, we need to study the candidate selection procedure. In this paper we focus on a crucial actor within this process: party members. Although there is a wide variety in candidate selection procedures of political parties, there is one common trend of democratization of intra-party democracy. This means that the selection procedure has become more inclusive: party members get a more influential role in determining which candidates will run for office. Party members not only can accept or reject the candidate list, but increasingly have the opportunity to alter the order of candidate lists. Theory, however, suggests that a more inclusive selection procedure results in a less representative candidate list due to two reasons. First, the representational problem: since party members are unrepresentative of society itself, they will not create a representative list either. The second mechanism is the coordination problem: because every party member compiles their preferred candidate list in isolation, they cannot coordinate their vote to compose a representative list. These mechanisms suggest that an inclusive candidate selection procedure makes it harder for women to become politicians. Previous studies have found mixed results when testing the relation between the inclusiveness of the candidate selection procedure and the representativeness of the candidate list. To get a better understanding of the relation between inclusiveness and representative candidate lists, this study will administer a survey among party members with a rank conjoint experiment. Party members need to rank candidates on a candidate list that are different regarding their gender, age, party activism and political experience. Consequently, we can test whether party members either prefer candidates that look like them (triggering the representation problem) and/or whether and how party members vote strategically (the coordination problem). The experiment will be conducted in the Netherlands: a country in which only one third of the MP’s is a woman. Although voters increasingly cast preference votes for women, and thus seem to signal to political parties that they are favorable towards women, political parties have until now not presented gender balanced candidate lists. In the Netherlands the majority of parties grant their members an influential role in the candidate selection process. It is thus crucial to study the way these party members use these opportunities and decide on the composition of the candidate list.