When each of us gets up in the morning, we start messing with what might as well be avatar customization tools to change our appearance. We decide what clothes and jewelry to wear. We decide which hairs to shave and which hairs to style. Some of us occasionally make more radical alterations, such as getting tattoos, piercing various dangly bits with metal, or even going in for cosmetic surgery. In real life, though, we’re often limited in the changes we can make to appear taller or more prosperous. Videogames and virtual realities, on the other hand, are more flexible.

Researchers have been studying the effects our appearance has on how other people react to us for a long time, but they’ve also started to seriously study the psychology of our video game avatars. At first they used models of human behaviour relevant to appearances in real space, but they have gradually built up new concepts to understand how people behave when they adopt different types of in-game form. Why do we choose the avatars that we do? How do different avatars change our behavior in games? And how does the experience affect us when we select ‘quit game’ and re-enter the real world?

Sure, explaining why we adopt the avatars we do is sometimes easy: we decide to look like an elf because elves get +5 Intelligence and we want to max out our mage build. Put that one in your thesis and smoke it. But what about virtual playgrounds where we have options that aren’t constrained by the game’s mechanics? An emerging line of research says that when the choice is ours, it’s often about building a better version of ourselves.

Hey, I’ve got an entire chapter about the psychology of game avatars in my book, Getting Gamers: The Psychology of Video Games and Their Impact On Those Who Play Them. Click above for more info.

“Studies have shown that, in general, people create slightly idealized avatars based on their actual selves,” says Nick Yee, who used to work as a research scientist at the Palo Alto Research Center but who now works at Ubisoft. He should know: before joining Ubisoft Yee has spent years studying the effects of avatars on human behavior in settings such as Second Life and World Of Warcraft. “But a compensation effect has been observed. People with a higher body mass index – likely overweight or obese – create more physically idealized avatars, [which are] taller or thinner. And people who are depressed or have low self-esteem create avatars with more idealized traits, [such as being] more gregarious and conscientious.”

Other researchers have found that the ability to create idealized versions of ourselves is strongly connected to how much we enjoy the game, how immersed we become, and how much we identify with the avatar. Assistant professor Seung-A ‘Annie’ Jin, who works at Emerson College’s Marketing Communication Department, did a series of experiments with Nintendo Miis and Wii Fit.1 She found that players who were able to create a Mii that was approximately their ideal body shape generally felt more connected to that avatar and also felt more capable of changing their virtual self’s behavior – a fancy way of saying that the game felt more interactive and immersive. This link was strongest, in fact, when there was a big discrepancy between participants’ perceptions of their ideal and actual selves.

“I would definitely recommend that developers allow players to design and don whatever kinds of avatars they like,” states Jim Blascovich, a professor of psychology at the University Of California in Santa Barbara, and co-author of the book Infinite Reality: Avatars, Eternal Life, New Worlds, And The Dawn Of The Virtual Revolution.2 Doing so tends to make the game more appealing and lets us connect more with our avatar and the world he or she inhabits. But what then? Once we’ve adopted an avatar, how does its appearance affect how we play games and interact with other players?

This research has its roots in what’s called self-perception theory, a watershed concept in social psychology pioneered by physicist-turned-psychologist Daryl Bem in the 1960s. Essentially, the theory says that we observe ourselves and use that information to make inferences about our attitudes or moods, as opposed to assuming our attitudes affect our behaviors. For example, someone who hurls themselves out of an airplane with a parachute might think, “I’m skydiving, so I’m the kind of person who seeks out thrills.”

In one clever study of this theory by Fritz Strack and his colleagues3, subjects were given a ballpoint pen and told to hold it in their mouth in one of two ways. Some were asked to use pursed lips and others were told to hold it between their front teeth, with their lips drawn up and back. The former approach tricked the subjects into frowning, while the latter got them to smile. When asked to rate the amusement value of a cartoon, those who were being made to smile thought it was far funnier than those who were forced to frown. Their appearance was affecting their mood.

My own avatar on Xbox Live

This kind of “first behavior, then attitude” effect has been widely replicated in other studies. In one, researchers hooked male participants up to a monitor that beeped in time with their heart rates while they perused centerfolds from Playboy magazine.4 When the researchers used their control over the machine to fake an accelerated heartbeat, subjects decided that they must have a thing for the particular model they were viewing. The effect was even still there two months later when subjects were invited back.

So first we perceive what we look like or what we’re doing, and then we draw conclusions about our attitudes and identity. And it turns out that we may continue to act in line with that presumed identity. In fact, Yee started his career by taking the precepts of social identity theory and using them to understand how people behave depending on the virtual avatars they assume. In one of his earliest experiments,5 Yee had subjects don a head-mounted display that let them perceive and move around in a simple virtual environment. There was just a virtual room, another person controlled by someone else, and a virtual mirror. The mirror was important, because it obviously wasn’t a real mirror and the researcher could use it to show whatever ‘reflection’ of the subjects’ avatars he wanted. In fact, Yee randomly showed subjects one of three types of avatar reflection: ugly, normal, and attractive.

What the researchers were interested in was how this would affect how subjects interacted with the other person in the virtual room. After following directions to inspect their avatars in the mirror, subjects were asked to approach the room’s other occupant and chat with him or her. This other person was controlled by a research assistant and followed a simple script to get the conversation going, saying something like: “Tell me a bit about yourself.”

The study revealed that an avatar’s attractiveness affected how its owner behaved. Relative to those with ugly avatars, people assigned attractive looks both stood closer to the other person and disclosed more personal details about themselves to this stranger. Then, in a follow-up study using the same setup, Yee found that people using taller avatars were more assertive and confident when they engaged in a simple negotiation exercise. So, generally speaking, people with prettier and taller avatars were more confident and outgoing than those with ugly and stumpy virtual representations. Like in the real world, we first make an observation about our avatar, infer something about our character, and then continue to act according to our perceived expectations. We needn’t make a conscious decision to do it.

Video game Avatar. See what I did there? Eh? Eh?

“Studies have shown that people unconsciously conform to the expectations of their avatar’s appearances,” said Yee when I contacted him to talk about this study. “We’ve termed this phenomenon the Proteus effect, after the Greek god who could change his physical form at will. These studies in virtual environments parallel older studies in psychology showing that people conform to uniforms given to them.”

The Proteus effect, then, describes the phenomenon where people will change their in-game behavior based on how they think others expect them to behave. “In our studies at Stanford, we have demonstrated that avatars shape their owners,” agrees Jeremy Bailenson, an associate professor at Stanford University and Infinite Reality’s other author. “Avatars are not just ornaments – they alter the identity of the people who use them.” Subsequent research by Yee, Bailenson and others has even revealed that there doesn’t even have to be an audience for us to feel the need to conform to our avatar’s appearance – an assumed one is sufficient.

But what about after we quit? Well, our avatars’ power extends beyond the game, and perhaps unsurprisingly, there’s an angle to this that involves selling you stuff. Imagine, for example, that you’re in the Xbox dashboard and you notice that your avatar is holding up a branded soft drink and grinning like some kind of moron. Do you think you’d be more likely to remember that brand and pick some up the next time you’re at the shops? Research by Bailenson and his colleague Sun Joo Ahn suggests you would.6 In their study, the team altered photos of people to show them holding up fictitious brands of fizzy drinks like “Cassina” or “Nanaco.” Even though the participants knew the photo was doctored, they tended to express a preference for the fake brand, simply because they’d seen a representation of themselves holding it.

Drink Ternio brand soda, you witless consumer lemming. Taken from Ahn & Bailenson (2011).

Other researchers have found similar results when they showed people pictures of themselves in a certain brand of clothing, and one study by Rachel Bailey, Kevin Wise and Paul Bolls at the University Of Missouri in Columbia looked at how kids reacted to advertisements for sweets and junk food that were thinly disguised as Web games. If the ‘advergames’ allowed players to customize their avatars, the kids remembered the snacks better and said that they enjoyed the game more.7

It’s not all scary news, though. For example, psychiatrists use mental visualization as a technique for treating phobias and social disorders. Someone deathly afraid of swimming, for instance, might be coaxed into imagining themselves at a pool. Through this kind of repeated imaginary exposure, the person might eventually seize control of their phobia.

And along those same lines, a body of work around social learning theory has shown that we can be encouraged to adopt new and beneficial behaviours by watching others perform them. The more similar the other person is to us, the more likely it is to work. Today, the technology exists to take our likeness and show it exercising and eating vegetables instead of chugging soft drinks. In fact, some researchers are experimenting with such approaches. Jesse Fox and Bailenson at Stanford University recently published a paper in which they examined this exact possibility.8

In the study, the researchers outfitted participants with a head-mounted display and set of controls that let them experience and navigate a simple virtual environment. Some people saw avatars with photo-realistic images of their faces attached, while others saw no avatar, or an avatar with an unfamiliar face. Everyone was then told about the importance of physical activity, asked to practise some simple exercises, and invited to keep exercising for as long as they wanted. Through a series of experiments based on this setup, Fox and Bailenson found that when people saw avatars that looked like them mirroring the exercises they tended to work out for longer. The effect was even greater when they saw the avatar slim down in the process. When asked later, people who saw their face on happy avatars also reported hitting the gym after being dismissed.

So while you needn’t have a panic attack the next time you see a character-creation screen full of choices, keep in mind that whatever you pick not only says something about you, but it can unconsciously affect how you behave on both sides of the screen as well.

A version of this article first ran in Edge Magazine.