Value Representation

Abstract

Public opinion research shows that core political values play a central role in structuring voters' beliefs, which makes them a natural site for the study of substantive representation. However, existing approaches to measuring representation focus instead on voters' issue preferences and ideological positions. This dissertation offers the first study of value representation. In the first paper, I evaluate the centrality of political values to the development of issue stances. I employ panel data from the ANES and GSS to show that values drive the formation of issue stances over time to a greater extent than left-right ideology or partisanship does. As voters' values shift, their issue stances change to be in line with the value. Additionally, when new policy issues come to the fore, voters' values help to drive the formation of their attitudes on these issues. In the second paper, I measure legislators' value communications with constituents using over 100,000 electronic newsletters. I construct a novel dictionary of values-speech using a supervised learning algorithm. I find that legislators use values-language about as often as partisan or policy language and more often than ideological language. Legislators' value speech is stable over time, across legislative topics, and across mediums. Finally, extreme and electorally safe legislators are more likely to use speech on values owned by their party, while moderates and electorally vulnerable legislators are more likely to discuss concrete policy benefits. In the third paper, I develop a framework for value representation and test whether legislators' actions reflect their constituents' traditionalist and egalitarian values. I estimate voters' traditionalist and egalitarian values at the district and state level using multilevel regression and poststratification with the 2012, 2016, and 2020 ANES. On the legislative side, I classify the values-content of hundreds of roll call votes, over 550,000 bill cosponsorships, and the newsletters described above. Across all three types of activities, I find that legislators are responsive to their constituents' values. Value responsiveness is also higher if the legislator's party "owns" the value in question. I conclude that legislators are responsive to voters' values in contemporary American politics.