People Can Infer Which Politicians Are Corrupt From Their Faces

This is a photo of two men exchanging cash

When people are shown photos of politicians they’re not familiar with, they can make better-than-chance judgments about whether those politicians have been convicted of corruption, according to research published in Psychological Science, a journal of the Association for Psychological Science.

People can make these judgments even without knowing anything about the politicians or their careers. And one thing they seem to be picking up on is how wide the politicians’ faces are.

The individual people making these judgments do only very slightly better than if they were making a random guess (although the difference is statistically significant). However, the judgments become much more accurate when they are combined across a group of people.

Face wideness—technically, the facial width-to-height ratio—has been shown in previous research to be correlated with aggressive behavior in men. That is, men with wider faces have a greater tendency to be aggressive and threatening toward others than do men with thinner faces. Studies have also shown that wide-faced men are perceived by others to be more threatening than men with thinner faces.

The research is the first to show that observers have a knack for picking out corrupt politicians based on just a portrait and that observers perceive politicians with wider faces as more corruptible.

“It might be difficult to understand why you can look at others’ faces and tell something about them,” says Chujun Lin, study coauthor and Caltech graduate student. “But there is no doubt that people form first impressions from faces all the time. For example, on dating sites people often reject potential matches based on pictures without reading the profile.”

However, it’s important to note that the researchers are not claiming that politicians who look corruptible are inherently more corrupt than those who look honest. Their research shows a connection between facial appearance and corruption, but, they say, there could be many explanations. One possibility is that if a face conveys a sense of dishonesty, the politician might be offered bribes more often. Another possibility is that corruptible-looking politicians are not any more corruptible than honest-looking politicians, but because of their looks they are more often suspected of, investigated for, and convicted of corruption.

“If a jury is deciding whether or not a politician is guilty, having a corruptible-looking face might create a negative impression, which might influence the jury’s decision,” says Lin, who adds that the “clean” politicians used in the study might not actually be clean. “Maybe they just haven’t been caught.”

The study consisted of four parts.

Part 1: The researchers collected pictures of 72 politicians who held office at the state or federal level. Half had been convicted of corruption and half had clean records. For consistency, all of the politicians included were male and Caucasian. All of the photos were black-and-white; cropped to the same size; and featured a frontal, smiling portrait. The images were presented randomly to 100 volunteers, who were asked to rate each politician on how corruptible, dishonest, selfish, trustworthy, and generous they looked.

An analysis of the data collected showed that the volunteers as a group were able to correctly differentiate the corrupt politicians from the clean politicians nearly 70 percent of the time based on their faces alone.

Part 2: The second part of the study replicated part 1, but used photos of 80 politicians elected to state and local offices in California. Half had violated the California Political Reform Act—a law that regulates campaign finance, lobbying, and politicians’ conflicts of interest—and half had clean records. As before, the data showed that the volunteers could correctly differentiate the corrupt politicians from the clean politicians nearly 70 percent of the time.

Part 3: The third part of the study used the images from part 1, but asked the volunteers to judge the politicians on a new set of criteria: corruptibility, aggressiveness, masculinity, competence, and ambitiousness. The data from this study showed that only corruptibility-related trait inferences (inferences of corruptibility, dishonesty, selfishness, aggressiveness, generosity, and trustworthiness) differentiated corrupt politicians from the clean politicians. Inferences of competence, ambitiousness, or masculinity did not predict the politicians’ records.

Part 4a: In the fourth section of the study, the researchers examined which of the politicians’ facial structures the volunteers associated with dishonesty and corruption. The faces were characterized by eight measures that described things like distance between the eyes, size of the cheekbones, nose length, and face width. By comparing the data from those measures against the judgments made by the volunteers and the records of corruption convictions, the researchers found that politicians with higher facial-width ratios were more likely to be perceived as corruptible.

Part 4b: To double check that face width was truly the characteristic driving negative perceptions about a given politician, the researchers gathered photos of 150 politicians and digitally manipulated each into a wide-faced version and a narrow-faced version. The 450 resulting photos, including the 150 unaltered originals, were shown to 100 participants who were asked, as in the previous studies, to rate each image according to how corruptible the politician seemed. And again, face width made the difference. The volunteers judged the wide-faced versions of the politicians to be more corruptible than their thin-faced counterparts.

“These findings raise many interesting questions for future research,” says Lin. “For example, what is the underlying causal mechanism of the correlation between perceived corruptibility and politicians’ records found in our study? Are politicians who look more corruptible more likely to be suspected, investigated, and even convicted?”

The findings might make a person wonder why corrupt politicians get elected in the first place if people can tell they’re corrupt just by looking at them. Co-author Ralph Adolphs, Bren Professor of Psychology, Neuroscience, and Biology, and director and Allen V. C. Davis and Lenabelle Davis Leadership Chair of the Caltech Brain Imaging Center, says that a lot more than just a face goes into how you feel about a person.

“In the real world, you’re not just seeing a photo of a politician. You’re seeing them talk and move,” says Adolphs, who is also on the leadership team of the Tianqiao and Chrissy Chen Institute for Neuroscience. “Their face might make a first impression on you, but there are other factors that can come in and override that.”

Before conducting this study, Lin researched how politicians’ facial characteristics relate to their success in elections across different cultures. She plans to continue her research on the judgments people make based on facial structures. Her next study asks volunteers to judge people from all walks of life on 100 different characteristics such as helpfulness, meanness, and healthiness, based on their faces.

Michael Alvarez, a Caltech political science professor, is also a coauthor on the study.

Funding for the research was provided by the Caltech Conte Center for Neurosciences and the National Institute of Mental Health.

Comments

Trump has an extremely wide face.

Of course another alternative explanation is that the study will not replicate because of some as yet recognized bias in stimuli selection. Happy replication!

I have same feeling as Thomas Hilton! The one absolutely critical but simple experimental control missing from this study was to present only proven corrupted politician images to the participants. If the authors’ theorizing is correct, the vast majority of faces would be judged correctly as ‘corruptible’. But, if the participants are in some way adjusting their responses based upon some kind of expectation that not all faces would be of ‘corrupted’ individuals (perhaps thinking that who in their right mind as an experimenter would present them with everyone who looks corrupted?), then we would detect that biasing effect – which would moderate any conclusion we might draw about such judgements. E.g. when all pictures are of corrupted politicians, and respondents individually give 60% accuracy, and an aggregate of 70% accuracy .. well, you see the problem!

I have same uneasy feeling as Thomas Hilton! The one absolutely critical but simple experimental control missing from this study was to present only proven corrupted politician images to the participants. If the authors’ theorizing is correct, the vast majority of faces would be judged correctly as ‘corruptible’. But, if the participants are in some way adjusting their responses based upon some kind of expectation that not all faces would be of ‘corrupted’ individuals (perhaps thinking that who in their right mind as an experimenter would present them with everyone who looks corrupted?), then we would detect that biasing effect – which would moderate any conclusion we might draw about such judgements. E.g. when all pictures are of corrupted politicians, and respondents individually give 60% accuracy, and an aggregate of 70% accuracy .. well, you see the problem!


APS regularly opens certain online articles for discussion on our website. Effective February 2021, you must be a logged-in APS member to post comments. By posting a comment, you agree to our Community Guidelines and the display of your profile information, including your name and affiliation. Any opinions, findings, conclusions, or recommendations present in article comments are those of the writers and do not necessarily reflect the views of APS or the article’s author. For more information, please see our Community Guidelines.

Please login with your APS account to comment.