In this article, we analyse the impact of intra-party procedures of candidate selection for national elections on the representativeness of parties towards their voters. With regard to candidate selection we distinguish between two dimensions: inclusion and centralization. While the first identifies the type of selectorate for candidate nominations (members, delegates or committees), the second captures the territorial unit in which the nomination is decided (local, regional or national). Based on data for 53 parties in 9 Western European countries for the period 1970 to 1990, the analysis points to the relevance of the inclusion dimension. Parties in which party elites decide the nomination of candidates show slightly higher degrees of representation than parties with more inclusive selectorates. We conduct our analysis separately for two frequently used but theoretically different concepts of representation: cross-sectional representation (at one point in time) and dynamic representation (over time). Our analysis shows that candidate-selection procedures only matter for the first concept. The empirically inconsistent results between the two concepts are due to deficiencies in the way dynamic representation is currently operationalized.