Incorporating legislators from historically underrepresented groups into the legislature should be associated with the introduction of new perspectives to the legislative process. Achieving an ideal form of political representation is not so simple. The institutional structure of the political system shapes legislative choices and defines the process of political representation. Structured legislative choices can have moderating effects on the presence of unique perspectives during the policymaking process. This means that the incorporation of historically underrepresented groups is not always sufficient for group representation. I develop a contextual theory of political representation that isolates when we should-and should not-observe unique patterns of political representation. I argue that ballot type and party affiliation are two distinct factors that shape legislative choices and define to whom legislators are accountable.

The theoretical argument synthesizes previous literature on gender and ethnic descriptive representation to develop an integrated theory of political representation. It leverages the uniqueness of group identity and cross-cutting factors to isolate where descriptive representatives should express unique patterns of political representation and the extent to which the political context conditions the legislative behavior of descriptive representatives. I develop a new measure of political representation using automated content analysis of legislative debates to empirically explore patterns in speech communication across different types of descriptive representatives. This measure makes it possible to empirically determine the strength of the divide that separates types of descriptive representatives. Unique perspectives should be apparent in the way legislators frame the justification and explanation of public policy to those who hold them accountable. This helps us identify the extent to which incorporating legislators from historically underrepresented groups has an influence on a broadly-defined set of issues.

Bolivia provides a unique opportunity to explore patterns of representation. Indigenous and female descriptive representatives have been historically underrepresented in Bolivia and possess interests that are relatively uncrystallized in the legislative assembly. The historical absence of these interests in the legislative assembly leaves a void in group representation. Indigenous legislators are expected to possess broadly-defined unique perspectives associated with group identity. Party affiliation, however, should structure the primary dimension of conflict for Indigenous representatives, whose interests overlap with partisan affiliation. Unique group perspectives should be observed within party. Female representation is different. The nature of cross-cutting factors should suppress the uniqueness of female perspectives within party, but unique collective perspectives should be obvious between parties.

Overall the empirical evidence supports theoretical expectations. Indigenous representation is structured by party when explored at the chamber-level. The most distinct patterns of Indigenous representation are within party. These differences are moderated when we explore the strength of the divide for those legislators elected on party lists compared to those elected in plurality districts. The most distinct speech patterns of female representation are at the chamber-level. Cross-cutting interests make it possible for women to speak with a collective voice. Unlike Indigenous representation, distinct patterns of female representation are moderated within party, where there is broader agreement on partisan issues among copartisans. I interpret these results as preliminary support for a theoretical argument that simultaneously explains ethnic and female descriptive representation.

Understanding when interests intersect and isolating the uniqueness of those interests can help us strengthen our broader understanding of gender and ethnic representation. We need to know where to look and how to find these unique patterns of political representation. Plenary debates provide an opportunity for different types of descriptive representatives to frame messages in order to simultaneously strengthen a collective partisan and individual reputation. This is particularly valuable in systems where political parties are unified. This project finds that the political context does indeed condition the behavior of descriptive representatives and moderate observed legislative behavior. Unique patterns of representation of historically underrepresented groups, fortunately, do exist under favorable conditions and fill a void of representation that satisfies normative values of democracy.