Research

An emerging agenda using large language models and vision-language models (e.g., GPT-4o, Claude) to evaluate political stimuli and candidates — comparing machine and human judgments, probing what models "see," and using AI to scale image- and text-based measurement in political science.

Affective and perceived polarization among Korean voters — distinguishing actual from perceived polarization, tracking it over time, and its consequences for democratic accountability.

Attitudes and behavior across social-group lines — gender and candidate evaluation, immigration and intergroup contact, and generational differences in political participation.

How voters decide and what moves turnout and vote choice — the opportunity costs of voting and election-day weather, heuristics in low-information elections, incumbency, and candidate appearance.

How wealth and economic inequality shape perceptions of social mobility, subjective well-being, and political behavior.