OB Research Seminar - John Antonakis, HEC Lausanne

Studies show that voters are biased by the facial appearance of political candidates and can predict reasonably well who will win an election after rating candidates only on how they look (i.e., on competence, intelligence, leadership). This effect is robust, whether using data across cultures or even when children perform the ratings. We replicate these results using adults but also extend this work by showing that what a candidate potentially says matters too. In experiment 1 (n=4,972), we show that inferences using appearances predict outcomes in 2024 U.S. congressional elections. In experiment 2 (n=1,862 raters; 11,172 ratings) we obtain absolute ratings on all the candidates--these scores correlate with the relative differential the candidates receive in Experiment 1. In Experiment 3 (n =10,438), we randomly assign to each face a brief written speech that is either high or low on charismatic signaling. We show that the charismatic speech changes what inferences raters make; however, the non-charismatic speech does not. The effect of charisma is equal whether looking at the average treatment effect across men or women or for various ethnic groups.