John Bolton, former U.S. ambassador to the United Nations, center, stands in the elevator at Trump Tower in New York. (Justin Lane/Bloomberg)

It’s not just women whose looks matter to President-elect Donald Trump, according to recent news reports. Apparently, John Bolton’s mustache is partly why Trump decided against nominating him for secretary of state. Sen. Bob Corker’s short height (and insider trading allegations) hurt him for the same job. If it’s true that Bolton’s and Corker’s looks kept them from the job, then Trump’s common offensive tactic against political rivals may genuinely be how he thinks about other people’s worth.

Of course, hiring — or failing to hire — based on looks is hardly new. In the past, Trump has told his resorts to hire only good-looking women. He’s urged managers to fire women he thought were not attractive enough. Nor is Trump the first high-level politician to conflate appearance with competence: President George W. Bush’s dissatisfaction with his National Economic Council director, Lawrence B. Lindsey, reportedly involved both Lindsey’s managerial skills and his “failure to exercise.”

Many people recoil from looks-based hiring standards, seeing them as ethically questionable at best. No federal law prohibits looks-based discrimination, but Michigan, the District of Columbia and several cities, including San Francisco, have banned it.

Still, how different are Trump’s evaluations of politicians’ appearance from those of ordinary citizens? According to our research, not much.

Here’s how we did our research

In our study, Peter Loewen and I asked Canadian undergraduates to rate official photos of incumbent politicians from Belgium and Israel. After briefly seeing a pair of photos, each student was asked to tell us which of the two had more of one of these three traits: dominance, trustworthiness and intelligence. This was repeated multiple times per student with different pairs of photos.

After collecting thousands of such pairwise comparisons, we computed overall impression scores on the three traits. And for both Belgian and Israeli politicians, mustached politicians were consistently and substantially evaluated as less trustworthy and less intelligent.

Mustaches weren’t the only superficial feature that changed how someone was evaluated. Glasses-wearing politicians were more likely to be judged as intelligent — but also as being less dominant. And while politicians who smile in photos score points on trustworthiness, they lose on intelligence. Both women and men were deemed less dominant and less intelligent when smiling. But smiling men were deemed more trustworthy, while women politicians were seen as less so.

It’s not clear whether voters are more or less likely to elect someone based on looks

So what? Do these judgments matter for electoral success? The evidence is mixed. Several studies find a statistical correlation between certain facial traits — especially those interpreted as signaling competence — and winning elections. Such studies have been done in the United States, Brazil, Mexico, France and Japan. At the same time, evidence suggests that voters don’t actually vote differently because of some facial features usually thought to hurt politicians — being bald, say, or having a “baby face.”

Perhaps more important, it’s not clear whether facial traits make the difference in getting elected once important factors affecting vote choice are accounted for, such as partisanship, ideology and political knowledge.

But political parties pay attention to looks when picking a candidate. Sometimes. 

However, parties do seem to have picked candidates based in part on how attractive they are. In a 2009 study, Matthew Atkinson, Ryan Enos and Seth Hill asked students to evaluate the attractiveness of almost 1,000 congressional candidates and found evidence that visually appealing candidates were strategically deployed. Then they analyzed what difference looks made in Senate races.

In a competitive race, the out-party was much more likely to put attractive challengers up against the incumbent, but not when the race was one-sided. The researchers estimated that, all else being equal, being more visually attractive can give a Senate candidate an edge of up to four points.

Does such a strategy pay off? In the 99 Senate races they examined, the estimated “facial gap” between candidates in a race was never enough to pass the victory margin of the winner. That is, they tested whether a race’s outcome would have changed if a losing candidate whose face respondents evaluated as being of below median attractiveness were replaced with an identical candidate with a face ranked as more attractive.

In no case did the increased votes make up for the margin of loss.

Why? It might be precisely because parties were more likely to deploy attractive candidates in hard-fought races. Those are exactly the races in which voters pay closer attention, have more information and are more likely to be mobilized by partisan groups. The marginal value of appearance might well make less difference in such a race.

Of course, none of this suggest that making cabinet appointments based on height or facial hair is smart or ethical. Looks-based evaluations of politicians have a strong gender bias. And beyond that, they are usually an unreliable cue for how decisions will be made once in office.

Do looks accurately predict how a candidate will behave?

As part of a different project, we interviewed the same Belgian and Israeli politicians whose faces we evaluated to assess to what degree they sought out risks in their decision making. Other fields’ evidence suggests that when individuals evaluate such things as whether someone looks “dominant” or “trustworthy,” they are picking up on facial evidence of testosterone levels — and testosterone levels predict aggressive and risky behavior.

But how “dominant” or “trustworthy” students thought an incumbent looked had no correlation with whether they made riskier choices. By contrast, politicians who were judged as more intelligent were, on average, more risk-averse.

But nothing about a mustache predicts someone’s decision-making.

Lior Sheffer is a PhD candidate in political science at the University of Toronto.