Attractiveness vs. Partisan stereotypes
Carmelo Licata,Pierre-Guillaume Méon
DOI: https://doi.org/10.1016/j.jebo.2024.01.014
IF: 2
2024-02-02
Journal of Economic Behavior & Organization
Abstract:Using two surveys, we study how respondents process visual cues to identify the political orientation (left- vs. right-wing) of members of the French National Assembly (referred to as "deputies"), based on official photographs only. We first confirm that respondents outperform random guesses. Second, we find that their categorizations correlate with observable characteristics (tie color, facial hair) and subjective assessments of deputies' personality traits (attractiveness, competence, trustworthiness). Third, the objective visual cues that respondents use are consistent with the actual characteristics of left- and right-wing deputies, and respondents mistakenly react to subjective personality traits that differ little across the two groups of deputies. Fourth, left- and right-wing respondents use the same cues in the same way. However, attractiveness stands out because respondents categorize attractive deputies as corresponding to their own orientation. Fifth, the magnitude of the marginal impact of a characteristic on the probability of a respondent categorizing a photograph as left- or right-wing increases strictly with the representativeness of that characteristic. Finally, we find evidence that some characteristics correlate with categorization errors. Findings 1, 2, 4, and the finding that respondents use cues in the correct way are consistent with both Bayesian behavior and the representativeness heuristic. Findings 5, 6, and the finding that respondents react to subjective cues that do not differ across groups are at odds with Bayesian inference but consistent with the representativeness heuristic. Finding that attractiveness trumps partisan stereotypes suggests the existence a partisan halo effect.
economics