New Directions in Media Measurement and Computational Social Science

Abstract

Communication is central to a functional, representative democracy. In an ideal world, processes such as group deliberation, media consumption, and political persuasion allow members of the public to form coalitions based on common interests and participate in elections with healthy, informed beliefs. This dissertation presents three novel media datasets to further our understanding of three important forms of political communication that enable these processes in the present era: (i) election news coverage on television, (ii) public relations from corporations, and (iii) public meetings in local government.

In the first paper (co-authored with David Rothschild), I describe how local television news covers candidates for federal office using broadcast transcripts from nearly 2,000 races across 3 electoral cycles. Moreover, I causally demonstrate that television news coverage (earned media) can, in many cases, deliver greater electoral gains than television advertisements (paid media) that are conventionally studied in campaign media effects. In the second paper, I systematically describe how the most recognized consumer brands in America use political cues in public speech and measure whether and how these cues represent their stakeholders' preferences as well as their parent firms' ideological priorities. In the third paper (co-authored with Tyler Simko), I argue that public meetings are the central political institution of local government and introduce LocalView, a new dataset of over 100,000 real-time public meeting transcripts from more than 2,000 local governments in the U.S, as a means of studying this institution at scale. In two applications, I harness this database to demonstrate map how local policy agendas - for example, COVID-19 public health responses - vary across partisan geographies.

Together, this body of work leverages modern computational social science tools to advance our knowledge of political communication with implications for theories of elections, local government, political economy, and partisan polarization.