A Psychologist's Guide to Online Dating

Can we predict romantic prospects just from looking at a face?

Mike Blake / Reuters

Edward Royzman, a psychology professor at the University of Pennsylvania, asks me to list four qualities on a piece of paper: physical attractiveness, income, kindness, and fidelity. Then he gives me 200 virtual “date points” that I’m to distribute among the four traits. The more I allocate to each attribute, the more highly I supposedly value that quality in a mate.

This experiment, which Royzman sometimes runs with his college classes, is meant to inject scarcity into hypothetical dating decisions in order to force people to prioritize.

I think for a second, and then I write equal amounts (70) next to both hotness and kindness, then 40 next to income and 20 next to fidelity.

“Oh wow,” he says.

“What?”

“Your response is somewhat atypical for a female. Usually women allocate more to fidelity and less to physical attractiveness. Maybe you think fidelity is something people can cultivate over time?”

(Sure, but I mean, who would want an ugly, broke jerk sticking faithfully by their side?)

Royzman said that among his students (not in a clinical condition), men tend to spend much more on physical attractiveness, and women spend more on social attractiveness traits like kindness and intelligence.

This trait game, along with Royzman’s review of the literature on attraction, hints at some of the endless quirks of the online dating marketplace. You might like someone online, but they put 100 on income, and unfortunately you’re about a 10.

Men and women make mating decisions very differently, he speculates. Men tend to act like single-issue voters: If a prospect is not attractive enough, he or she usually doesn’t qualify for a first date, period.

For women, however, "It's a more complex choice,” he said. “What tends to matter for females is that the overall package is good," meaning that women might accept a less-attractive mate if he was outstanding in some other way. "Online, this might result in males restricting their potential mates.”

Match.com is two decades old, but new, fast-growing apps such as Tinder have shifted the online-matching emphasis back to looks. Tinder dispenses with the idea that it takes a mutual love of pho or Fleet Foxes to create a spark; instead, users of the phone app swipe through the photos of potential mates and message the ones they like. As one columnist who used the service put it, “There’s a short bio, age, and mutual friends listed, but who’s really paying attention to that stuff when your Tinder flame is wearing next to nothing on the beach?”

Then there’s Hinge, which uses a similar interface, but is backed by recommendations from the user’s “social graph,” such as their school or career field. Grindr serves up a mosaic of gay bachelors’ head and body shots. There are also a raft of appearance-based spin-off sites, such as Facemate, a service that aims to match people who look physically similar and thus, the company’s founder claims, are more likely to have chemistry.

This more superficial breed of dating sites is capitalizing on a clear trend. Only 36 percent of adults say marriage is one of the most important things in life, according to a 2010 Pew study, and only 28 percent say there is one true love for every person (men are more likely to say so than women). Rather than attempting to hitch people for life based on a complex array of intrinsic qualities, why not just offer daters a gaggle of visually appealing admirers?

Recent research has examined what makes people desire each other digitally, as well as whether our first impressions of online photos ultimately matter. Here, then, is how to date online like a social scientist.

Does the photo matter?

Tinder offers a one-sentence tagline and a selection of five photos, including the all-important first photo, or “calling card,” as the writer Amanda Lewis put it. She points out a few other tips in her “Tinder glossary:” “Most players reflexively swipe left [reject] at the sight of a toddler or baby,” but posing with your adorable Lab can be an “effective misdirection.” And then there’s the iron law that “95 percent of players who choose a calling card that does not include a clear shot of their face are unattractive.”

It’s not the first time in history that a face plays such an important role in one’s fate. Physiognomy, or the bogus theory that we can predict a person’s character from their features, was once a widespread doctrine. Charles Darwin first began to develop his theory of natural selection while journeying on the HMS Beagle as a “gentleman companion” to its captain, Robert Fitzroy, but only after nearly being turned down from the job because Fitzroy thought “no man with such a nose could have the energy" required for an arduous voyage.

There has been some evidence that strangers can accurately predict qualities like extraversion, emotional stability, and self-esteem based on photos. Hockey players with wider faces, considered a sign of aggression, spend more time in the penalty box.

Tinder

It takes longer, more meaningful interactions, however, to pinpoint other traits, like if the prospective mate is open, agreeable, or neurotic. It seems people might only be able to determine the extremes of a personality from a photo, rather than its nuances. (One study found that the owner of an "honest" face is not any more likely to be trustworthy, for example.)

It’s true that attractive people generally are treated more nicely by others, and they might have better-adjusted personalities as a result. But Royzman said looks can deceive. In relationships, personality eventually overtakes attractiveness—or at the very least, we tend to find people more attractive when we think they have good personalities. So perhaps you should make that Tinder tagline all about how you volunteer at an animal shelter every weekend.

Swiping through endless Tinder photos in search of the most alluring possible one might not be fruitful, either. Most people end up with someone who’s about as good-looking as they are.

“People might prefer attractive people, but they often end up pairing off with people who are similar in attractiveness,” Leslie Zebrowitz, a psychology professor at Brandeis University and an expert on face perception, said. “You might shoot for the moon, but you take what you can get.”

Should I date someone who looks like me?

Twenty years ago, Christina Bloom was in a committed relationship when she met someone who “knocked me off my heels.” The two embarked on a fiery romance, during which she noticed that friends and strangers were always telling them they looked alike.

She launched FaceMate in 2011, drawing on her opinion that people in happy relationships tend to resemble each other. The site matches the photos of its users based on their faces’ bone structure using face-scanning techniques and a computer algorithm. The service is free, for now, and currently has 100,000 users.

“It all starts with the face,” she said. “People say, ‘From the first time I met him, I knew.’ There’s a sense of recognition. That's what they're seeing, is their own image. That's what we call chemistry.”

Psychologists tend to disagree with that theory. In another experimental mock speed-dating event, subjects who thought they were similar to one another were more likely to be attracted to each other, but that wasn’t the case for those who were actually similar to one another.

“People are not romantically attracted to people who look like them,” Zebrowitz said. “That has to do with the disadvantages of mating with your brother, for example.”

Indeed, Lisa DeBruine, a psychologist at the University of Glasgow in the U.K., has found that people find self-resembling, opposite-sex faces to be trustworthy, but not sexy, and they can even be repulsive for a short-term relationship.

An example of a photo being modified
to look more like the individual's partner,
from the study on Norwegian couples. (PLoS)

But George Michael and Maeby might be relieved to know that while excessive genetic overlap between two people results in poor reproductive prospects, a small amount can be acceptable. That might be why 20 young Norwegian couples rated their partner’s photograph as more attractive when it was digitally “morphed” to look ever so slightly more like themselves. The magic number was a 22 percent resemblance—any more similar was deemed gross.

And, by the way, you really should call the whole thing off if one of you says potato and the other “po-tah-to” (because after all, who says it like that?). Couples with similar speech styles were more likely to stay together than those who speak differently.

DeBruine points out that though we’re programmed to avoid dating our relatives, we also have a certain, subconscious affinity for our own parents.

“The scientific evidence reflects complexity and suggests that there may be a ‘happy medium,’” DeBruine told me. “But, ultimately, other factors are much more important in successful relationships.”

Will my online dating attempts lead to a relationship?

We may have more options for potential mates than ever before, but unfortunately people have trouble determining what they really want in their lovers. One 2008 study by Eli Finkel and Paul Eastwick at Northwestern University found, for example, that though men and women tend to say they prioritize different things in their mates (men are more likely to emphasize looks and women money), there’s no difference in the types of mates the two sexes actually choose in a real-life setting—which the authors gauged using a speed-dating exercise.

What’s more, there was little association between the traits participants said they wanted in a partner on paper and what they actually liked about the mates at the speed dating event. In other words, you may flaunt your Rolex in your Tinder photo, but that might not stop your date from heading home with a scruffy artist once you’re at the bar.

This is in part because the way people pair with one another on dating sites is different from the way they will then later evaluate the relationship, according to Finkel and Eastwick. People browse online profiles in what’s known as “joint evaluation mode,” comparing multiple suitors against one another on the basis of attractiveness, income, and other factors. But they make relationship decisions in what’s called “separate evaluation mode,” judging just that person and thinking, “Is this person right for me?” Even if you pick out the prospect with the most striking jawline, and you may overlook the one who will willingly spend hours watching Cake Boss with you, sans judgement.

“The joint evaluation model ... is likely to cause users to focus on certain qualities they think are important in a potential partner, perhaps to the neglect of qualities that actually are important,” Finkel wrote in a paper published last year in the journal Psychological Science.

“Certain qualities are easy to focus on in a joint evaluation mode (e.g., height, income, physical appearance),” Finkel later told me in an email. “But the truth is that those qualities aren’t the important ones that predict relationship well-being. What we really want is information about rapport, compatibility of sense of humor, sexual compatibility” and the like.

And computers simply aren’t able to convey information about people the way people can about themselves, Finkel says.

“There is something that people must assess face-to-face before a romantic relationship can begin—the myriad factors such as sense of humor, rapport, interaction style, holistic impressions, and nonconscious mimicry that determine how comfortably two people interact. You can assess compatibility better in 10 minutes of face-to-face time than in 100 hours of profile browsing.”

Finkel and Eastwick wrote that while online dating services greatly expand the dating pool for their users, they don’t necessarily foster better relationships: The sites “do not always improve romantic outcomes; indeed, they sometimes undermine such outcomes.”

At the same time, though, apps like Tinder remain remarkably popular. A little over a year after its launch, two million Tinder “matches” happen each day.

I asked Finkel which online dating site he’d use, if he had to use one. He said it depended on what he was looking for.

“If I were an Evangelical Christian looking for marriage, I might start with eHarmony. If I were looking for an extramarital affair, I might start with AshleyMadison. If I were in my 20s and looking for fun, casual dating, I might start with Tinder,” he said. “The whole point is that you can’t tell much from a profile, anyway, so using some complex algorithm to assess whether the partner is as kind as Mother Teresa or as smart as Einstein is a fool’s errand. Find somebody who seems cute or sexy, and then get face-to-face to assess whether there’s actual compatibility there.”

I also asked him if he’d use online dating at all, as opposed to some other matchmaking mechanism, knowing what he knows about it academically.

“Hell yes,” he said. “It’s probably a bit worse than meeting people organically through one’s existing social network, but, outside of that option, it’s probably as good an approach as any. But it’s important to realize what online dating can and can’t do. It can expand the pool of potential partners, making available a whole slew of people who otherwise would have been unavailable. That’s a huge, huge benefit. But, at least thus far, it can’t figure out who’s compatible with you. That’s your job.”

Olga Khazan is a staff writer at The Atlantic. She is the author of Weird: The Power of Being an Outsider in an Insider World. She has also written for The New York Times, the Los Angeles Times, The Washington Post, and other publications. She writes a Substack on personality change.