Greg Laun pointed me to this paper by Alexander Todorov, Anesu Mandisodza, Amir Goren, and Crystal Hall, whose abstract states:
Inferences of competence based solely on facial appearance predicted the outcomes of U.S. congressional elections better than chance (e.g., 68.8% of the Senate races in 2004) and also were linearly related to the margin of victory. These inferences were specific to competence and occurred within a 1-second exposure to the faces of the candidates. The findings suggest that rapid, unreflective trait inferences can contribute to voting choices, which are widely assumed to be based primarily on rational and deliberative considerations.
It looks pretty interesting. I’d also like to see things broken down by elections that were and were not seriously contested. Even without appearance, we can predict something on the order of 90% of races just based on incumbency and the partisan preferences of the voters in the states and districts. This is not at all to dismiss the finding but rather to place it in the context of other factors affecting voting.
You have to be careful in interpreting the results, however. Todorov et al. seem to be saying that individual voters’ visual “inferences of competence” are affecting votes. Another story, perhaps more plausible, is that the more competent-looking people are the ones who rose to political success.
The accompanying editorial in the journal (by Leslie Zebrowitz and Joann Montepare) associates the appearance of competence with not looking baby-faced. But I’m amused by the part of the editorial that puzzles over why appearance isn’t the only thing that matters: “When does perceived competence fail to predict election outcomes? Todorov et al. found that more competent-looking candidates were defeated in 30% of races. One possible explanation is that face biases could have favored babyfaced candidates in those particular contests. . . .” I mean, sure the findings are interesting, but chill out! Can’t you be satisfied with predicting 70% of the time? That seems pretty good to me!
Finally, that “68.8%” in the abstract is funny. People just don’t know about rounding. (And, of course, Table 1 should be a time series graph. But I do give them credit for doing a secret-weapon-style display.)