While vast literatures explored the role of political representation in terms of isolated identities such as gender or race, many questions remain about how these identities interact. We develop an experimental framework to explore why certain types of representation are more important for different social groups and how converging social identities may mobilize voters. We test our theory based on a preregistered survey experiment conducted as part of the 2022 Cooperative Election Study. Shortly before the U.S. midterm elections, our respondents participated in a visual conjoint experiment that matched them with various (fictional) lawmakers from different states. Respondents were asked to evaluate how well they would feel represented in each legislature based on the lawmaker's photo, their individual policy overlap, as well as the state legislatures' composition with respect to party, race, and gender. As part of the conjoint, we manipulated whether policy overlap with each lawmaker focuses on issues specifically related to gender, race, or general policies. This design allows us to identify diverging preferences for different types of descriptive and substantive representation among historically underrepresented groups. We further contextualize these findings with data from a post-election wave after the midterms where respondents are asked to evaluate how represented they feel by their own state legislatures. In addition to contrasting individual preferences for different types of representation, our approach allows us to examine how differently positioned groups perceive politicians that share (or do not share) various components of their identity, and how it motivates subsequent political participation.