Artificial Judgment: AI Models and the Evaluation of LGBTQ Candidates

This study examines whether issue ownership moderates evaluative penalties faced by LGBTQ political candidates. Drawing on theories of stereotype congruity, leadership schemas, and political viability heuristics, the analysis investigates whether LGBTQ candidates receive different evaluations when associated with distinct policy domains, including economy, education, policing/crime, and national security. Regression analyses reveal a consistent electability penalty for LGBTQ candidates across issue domains, even after controlling for issue area and AI model effects. LGBTQ candidates were generally viewed as less electable, while competence penalties were smaller and leadership evaluations showed no significant differences by candidate identity.