Candidates' 'face value' could influence voter behavior

Taking politicians purely at "face value" can frequently predict their success in elections, according to a study by Princeton researchers published in the June 10 issue of Science.

Participants asked to choose which political candidate in a race seemed more competent — based solely on the candidates' photos — accurately predicted the outcome of 71.6 percent of U.S. Senate races in 2000, 2002 and 2004.

The findings suggest that fast, unreflective decisions can contribute to voting choices, which are widely assumed to be based primarily on rational and deliberate considerations, the researchers said.

"The findings are striking — I didn't believe them at first," said Alexander Todorov, assistant professor of psychology and public affairs . "I think that a lot of inferences that we make about other people are fairly automatic and can even occur outside of conscious awareness. The catch is that these inferences can influence important deliberate decisions."

The evaluations of the candidates were derived solely from facial appearance. Participants were shown black-and-white headshots of two candidates in 95 Senate races. If a participant recognized either candidate, the data were excluded.

Races involving highly familiar candidates such as Hillary Clinton and Richard Gephardt also were excluded. Across all studies, participants were 843 undergraduate and graduate students at Princeton. However, judgments from as few as 40 participants were sufficient to reliably predict the outcomes of the Senate races.

The study also asked participants to look at photos of candidates in 600 U.S. House races in 2002 and 2004. In those races, the candidates who were deemed more competent won the election 66.8 percent of the time.

In addition, the researchers asked participants to make judgments based on the photos on a variety of other traits, including attractiveness, honesty, trustworthiness, charisma, likeability, extroversion and agreeableness. Only their judgments on competence accurately predicted the outcome of the elections, the study found.

Todorov, who has been a professor at Princeton since 2002, studies social cognition, judgment and decision making. He conducted the study with Anesu Mandisodza, a former research assistant, and Princeton graduate students Amir Goren and Crystal Hall.