March 19, 2025 | Noon to 1:00 PM EST
Kirill Zhirkov, University of Virginia
Work with Jazmin Brown-Iannuzzi and Jeremy Cone
Although the root causes of partisan affective polarization in the United States — increasing aversion toward members of the out-party among Democrats and Republicans — are numerous, biased mental representations of those in people’s own and the opposing political party may contribute to affective polarization. Importantly, existing studies measure perceptions about the parties using either self-reports or methods that allow capturing only one stereotype dimension at a time. We address the same topic using a new approach to directly measure mental representations called reverse correlation. This approach presents participants with randomly distorted images and, over several trials, asks participants to select which images are most representative of a social category. The selected images are aggregated to create the average mental representation of the category. We collected a sample of White participants who resided in the United States. Participants were selecting images of a typical Democrat or a typical Republican. A separate sample of participants rated these images. The findings suggest that participants tend to hold an ingroup bias with respect to valence: partisans imagine more positive and feminine mental representations of in-party members and more negative and masculine representations of out-party members. However, perceptions of race/ethnicity seem to be tied to which party participants are imagining: images of Democrats were rated as less representative of White Americans by all participants independently of their own partisanship.