The study of public opinion has become increasingly central to our understanding of American politics. What the American public believes, why it holds those beliefs, and whether or not those beliefs matter have become essential questions that guide our understanding of how American democracy functions. In order to answer these questions, however, it is important to consider the tools we use to measure public opinion accurately and reliably and to understand the substantive applications and limitations of those tools. This dissertation is composed of three essays that consider important questions in public opinion measurement today. The first considers how the technique of multilevel regression with poststratification (MRP) performs on polling data collected using area-based cluster sampling techniques. While MRP has been a boon to researchers with limited resources, it must still be examined to understand its strengths and shortcomings. The second paper uses two datasets to look at the measurement of scales of political values over time, focusing on both individual and state-level measures, and discusses implications of these results for larger debates around the measurement of partisan sorting and polarization. The third paper turns to the question of social desirability bias in polling. Specifically, it uses list experiments to look at whether survey respondents answer truthfully when asked about support for same-sex rights. These papers all aim to shed light on recent innovations in the measurement of public opinion and illustrate how we can use these innovations to improve our understanding of American public opinion.
| Identifer | oai:union.ndltd.org:columbia.edu/oai:academiccommons.columbia.edu:10.7916/D8VT24DV |
| Date | January 2017 |
| Creators | Stollwerk, Alissa Florence |
| Source Sets | Columbia University |
| Language | English |
| Detected Language | English |
| Type | Theses |
Page generated in 0.002 seconds