Talk: Alexander Wuttke (LMU)

Title:

Citizens and Democracy: Studying Mass Support for Democracy

Abstract:

Democracy is under pressure. Many aspiring autocrats who undermine democracy have come into office by democratic vote. Why do citizens vote away their democracies? This talk surveys current research on ordinary citizens’ attitudes towards democracy. It shows that generic support for democracy in the abstract is strong but manifest support for democracy as it exists is weak. Based on this finding, I argue that key to understanding whether citizens are committed to democracy is whether their support for democracy is meaningful, i.e. whether citizens possess an interrelated and coherent web of personally relevant attitudes towards democracy that are bolstered with underlying beliefs and cognitions. Yet, micro-level democracy research heavily relies on standardized surveys, which allow large sample sizes but are inept to assess whether citizens have meaningful attitudes. Therefore, we know very little about who genuinely supports democracy. I advance the idea of complementing standardized surveys with in-depth interviews that allow to map citizens’ democratic belief systems. Traditionally, in-depths interviews were conducted in-person which dramatically limits sample size. The technological revolution sparked by the emergence of large language models (LLMs) promises to resolve this trade-off between scale and depth. The grant proposal demonstrated here develops a technological infrastructure for scalable in-depth AI Interviewing, which allows for automated in-depth conversations in great numbers. During these dialogues about democracy, the AI interviewer poses open-ended questions and provides prompts for further clarification or expansion on participants' initial responses. This approach empowers individuals to express their perspectives freely and in their own words. Preliminary evidence from a small pilot survey demonstrates the challenges and promises of this method.