Do Gender Cues from Images Supersede Partisan Cues Conveyed via Text? Eye Movements Reveal Political Stereotyping in Multimodal Information Environments

Abstract

ABSTRACT An important feature of the information environment is its multimodal nature. In politics, people encounter representations of political candidates that combine images and text. Among the most prominent pieces of information people encounter are a candidate’s gender, obtained from images of the candidate’s face, and the candidate’s partisan identification, often represented as text. This feature of the environment is important given previous work showing that individuals infer gender categories from faces rapidly and effortlessly. In this study (N = 113), we use eye movements to determine how individuals assign stereotypical policy positions to candidates in an environment in which photos of candidates’ faces are paired with labels of their partisan IDs. We find that politically-knowledgeable individuals are more likely to use partisan- than gender-based stereotypes. In contrast, political novices did not prioritize one type of stereotype over the other. Our findings have implications for understanding how individuals make political evaluations in multimodal settings and show the advantages of measuring eye movements when studying stereotyping in multimodal environments.