Abstract

Research on affective polarisation has established that partisans perceive their copartisans more favourably across a range of traits (a phenomenon known as the partisan halo effect). Two questions about this effect remain largely unexplored: whether it extends to visual traits, and whether it operates bidirectionally, such that positive traits lead people to infer partisan ingroup membership. This project addresses both by investigating the mental representations Americans hold of the faces of Democratic and Republican voters.

Using a noise-based reverse correlation design, we first elicit mental representations of faces of Democratic and Republican voters from a national sample of 1,000 partisan Americans, then have a separate sample of 2,725 raters evaluate those images across ten dimensions, including attractiveness, trustworthiness, dominance, masculinity, smiling, and age. Our results reveal strong and consistent ingroup projection: both Republicans and Democrats hold similarly positive mental representations of individuals who share their political affiliation. Evidence for shared partisan stereotypes is more limited and indirect: while Republican faces are perceived as older, less smiling, and more dominant than Democratic faces, regardless of the participant’s own partisanship, partisans do not converge on a common representation of what a typical Democrat or Republican looks like.

Our findings suggest that the partisan halo effect can be bidirectional and includes visual traits, with implications for our understanding of affective polarisation, appearance-based biases, and the formation of stereotypes.

This talk will also present ideas for future steps of the project, and feedback from the audience would be very welcome!

Practical information