Content area
Abstract
This dissertation collects several essays on democratic opinion and speech.
First, however, I present a methodological study investigating how to elicit longer responses to open-ended questions in web surveys. Advances in computational methods, together with the advent of online surveys, have made it much more feasible to analyze large numbers of items where participants respond in their own words. This is an appealing prospect because it allows us to collect a wider range of opinions than those accomodated by closed-ended questions. However, this type of question is effortful, and respondents are incentivized to be terse. This study tests several methods of eliciting longer responses, and finds two that have a significant effect: a lottery scheme where each extra written word results in an additional entry, and simply asking people for a long response (while providing a word count as they type). The effects are substantially significant, with treatments resulting in around 20 additional words, and do not seem to result in lower-quality responses. These findings inform my data collection efforts for the next two studies.
The first substantive paper deals with the definitions of democracy held by ordinary Americans. This survey has two distinguishing features. First, I collect open-ended responses of unprecedented length. Second, I use closed-ended questions to probe for a uniquely diverse range of possible definitions. Closed-ended questions reveal a diverse landscape of views, though liberal conceptions of democracy predominate, as they do in other countries. The open-ended responses indicated that Americans, when thinking about democracy, consider a number of factors Academics have usually not thought to ask about. For example, only work on populism (e.g. Akkerman, Mudde, and Zaslove 2014) probes for the importance of popular sovereignty, which in my survey comes up in more responses than just about any other aspect of democracy. In addition, a small but significant set of respondents focus on the mechanics of decisionmaking, emphasizing majoritarian and consensual approaches in turn.
The next paper investigates the relationship between group threat and views on democracy. Other work has shown that white Americans who feel their group status threatened are more likely to support violence against political opponents (Thompson et al. 2023), but there is no causally-identified work connecting threat to broad views about democracy. I conduct a survey experiment on a sample of white Americans with the aim of establishing whether such a link exists. I find, first, that the threat treatment only worked on Republicans, joining recent research that suggests this once-universal threat is becoming a partisan issue (Brown et al. 2022). Second, even among Republicans, the threat had no effect on support for democracy or democratic norms. This surprising result may stem from the specific treatment used, as I discuss in the paper.
The final paper turns from citizens to politicians. I consider whether politicians in three anglophone countries --- the United States, the United Kingdom, and South Africa --- speak about democracy in distincive ways. Using word embedding approaches on party manifestos and transcripts from legislatures, I find not only distinctive national conversations about democracy, but also consistent themes on the party level. Democracy is clearly not a monolithic concept, nor is it a mere rhetorical device devoid of substantive content.
In this dissertation I make contributions in terms of methodology --- showing a way to get longer responses without an appreciable decrease in quality --- and data, introducing a dataset of South African parliamentary speech from 2008 to 2017. I show that, without prompting, a significant minority of Americans define democracy by its decision rules, while a preponderance of respondents mentions the importance of popular sovereignty. I show that, for white Americans, at least one group threat treatment does not change responses to broad questions about the democratic system, but also that group threat in response to demographic and cultural threats is felt by Republicans, not Democrats --- an emerging partisan response that deserves further study. Finally, I show significant differences across countries, and within party systems, in how party elites talk about democracy. Elected officials, when speaking about democracy, consistently mention the same issues, in a matter befitting their parties' histories and goals.
Details
