Your avatar and you: Personality and gender predictors of avatar-self discrepancy (Dunn & Guadagno, 2012)

Verne Troyer and his World of Warcraft avatar in an advertisement.

One of my goals in this blog is to review a broad range of topics related to videogames disallowing myself from narrowing in to one topic. Neuropsychological studies are the most difficult to read and review, other studies have lost my interest, some turned out to be more complex to understand, and some don’t need blogging as study authors took a step further beyond the university press release, then I don’t need to repeat the study. I guess my role is to give a voice to some studies, but I am just one fool.

Robert Dunn (East Tennessee State University) and Rosanno Guadagno (University of Alabama) have published in Computers in Human Behavior, a personality-related videogame study on avatar creation.


This study examined the influence of gender, the Big 5 personality factors, and self-esteem on virtual self-representation in the form of avatar-self discrepancy. To examine this, participants designed characters to play in a video game, spent 20 min playing the video game, and then had their actual pictures taken. Our results indicated that, consistent with predictions, men and women generally selected self-representations consistent with ideal male and female bodies. This finding was pronounced for men and women high in agreeableness. Conversely, some results contradicted the normative prescriptions often associated with self-presentation. For instance, men did not build taller avatars than did women. Men who were high in openness to new experiences were more likely to select avatars with skin tone variations. Introverts – both male and female – and women high in neuroticism were more likely to build attractive avatars. Moreover, those with low self-esteem were more likely to select lighter skin tones than those with high self-esteem. Thus, the effects of gender and personality have considerable implications for online self-presentation and self-representation.

It’s really hot out in Columbus, but that’s irrelevant as I am comfortable in my air conditioned office.

The expression of self is more or less in our control in the physical realm. We can control to a certain extent in the way we behave to others in various social contexts, we control our appearances by using combs, brushes, clothing, make-up. But there are certain expressions of self that we have no control over, such as height, ethnicity, baseline physical appearance barring plastic surgery, voice, among others. In the virtual world, a greater degree of control in expression of our self is afforded to us. The kind of control to our expression of self is called impression management that is we manage the impressions of our self in relation to others. A line of research has come up about the different aspect of the self: “actual self”, “true self”, “ideal self”, “para-authentic self”, “alter self”, “real me” among many others.

The authors were interested to ask how much of a discrepancy there are between the person and the avatar in terms of the person’s perception, others’ perceptions and some objective measures. Furthermore, they argued that certain personality characteristics may have predictive factors towards self-discrepancies.

Gender role is one of the personality factors that would influence avatar creation choices. The prescriptive norms of masculinity lead men to present themselves by competence through achievement, competition, and physical strength. This pea-cocking of strength ascribe to the muscularity norms of men. The authors hypothesized that men would create taller and larger avatars. The prescriptive norms of femininity lead women to present themselves by attraction through physical appearances and social warmth in relationships. Their pea-cocking of physical appearances would manifest, as the authors hypothesized, through thinness.

The Big Five personality traits are examined as factors that would influence avatar creation choices. The authors had hypotheses for each of the five personality traits: they hypothesized that more extraverted individuals would create avatars that are least discrepant from them. They hypothesized that more neurotic individuals would create more discrepant avatars due to being worrisome about appearances. More open to experience individuals would create avatars that are least discrepant from themselves. As for agreeableness and conscientiousness, they do not have any predictions as to what they might influence avatar creation choices.

Finally, they investigated the role of self-esteem in avatar choices. They hypothesized that individuals with low self-esteem are more likely to create avatars that are more discrepant from themselves. This is likely because they are dissatisfied with their real self and having control on their virtual self allows them to “uplift” their esteem through impression management.


Participants: 174 undergraduate students, 64 men and 110 women. Mostly white, followed by African American, Hispanic and such.


Self-Esteem: The Rosenberg Self-Esteem scale using 4-point scale. No further information.

Big Five Personality: The Big Five Personality Scale using an 8-point scale. No further information.

Avatar-self discrepancy: participants were asked four questions on how they relate to their created avatars. Two items used a 7-point scale and two used 3-point scale.


Participant coding: two undergraduate coders rated participants’ photographs based on skin tone, height, girth, and attractiveness on a 7-point scale. They went further by measuring chest size, waist size, hip size and height using fabric rules.

Avatar coding: two other undergraduate coders rated participants’ avatars based on skin tone, height, girth and attractiveness on a 7-point for each. Once again, using fabric rules to measure chest size, waist size, hip size and height on the avatars’ screenshots.

Videogame used: Neverwinter Nights 2.


Participants were told to create an avatar to play in the game. They were not given any specific instructions as to how to create them, which is a good thing since it would avoid influencing their avatar choices. After they finished creating their avatars, participants were asked to step out while the researcher takes a screenshot of the avatar. Of course, the participants were not made aware of that. Afterwards, participants played the game for 20 minutes. After play, participants proceeded to complete the questionnaires. When the questionnaires are complete, the researchers took participants’ head-to-toe photograph with a marker on a whiteboard indicating their height. Participants were led to believe that the photograph was meant to assist in further research in creating more realistic characters. At the end of the experiment, they were debriefed.


Through a series of t-tests, they found that men created avatars with larger chest relative to themselves than were women. Women created their avatars that were taller and thinner. Interestingly, the women were more likely to thing their avatars looked like them more so than the men, although this was cautioned by the authors due to a statistical assumption being broken.

I must ask whether the results were due to the limitations of the avatar creation system in Neverwinter Nights 2 or is truly due to personal choice. Case in point by Sociological Images is APB’s avatar creation options for the “heaviest characters”. Therefore in videogames, obese heroes don’t exist; flat-chest heroines don’t exist.  A gender-related consideration is how much the individual man or woman conforms to their respective gender norms. Men and women who highly endorsed traditional or even media-imposed gender norms are likely to create avatars that reflective of these norms as their “ideal selves”. Some self-report measures like drive to muscularity, thinness, among others might shed some light on this connection.

Their analyses for personality traits were conducted through regression analyses where one trait is the predictor variable with the physical characteristics as the outcome variables. For extraversion, they found that more introverted participants created more attractive avatars controlling for gender. This is much in line with prior research where introverts turn to computers to reach out and attract more social interaction than through face-to-face. The authors argued as for this personality trait and the others is that real life circumstances pretty much influence their impression management in virtual life.

For neuroticism, they found that highly neurotic women created more attractive avatars than themselves whereas highly neurotic men created less attractive avatars than themselves. The authors argued for future research into examining how neurotic women (and men) derive some value out of their control of their virtual (and real through photoshop) self-representation in computer-mediated environments.

For openness to experience, very open men created darker (or should I say tanned based on the numbers) avatars than themselves whereas less open men created lighter avatars than themselves. I sense a hint of connections between ethnic openness, ethnic diversity, social dominance orientation, right-wing authoritarianism, low openness to experience and impression management. I’ll muse about it later, if ever. Interestingly, openness to experience is predictive of participants’ self-avatar discrepancy in that higher scores is correlated to greater relatedness with their avatars. The authors discussed this as an open way for experimentation among men who would be less concerned about ostracism. As for women’s lack of skin tone variation as a function of openness, they discussed that other factors might have a role that predicts skin tone selection. At the top of my head, it might have to do with attractiveness, lighter skin tones is prescribed by Western culture to be more attractive for men. But then again, a guido wanna-be would very much like to have a tanned avatar.

For conscientiousness, less conscientious participants created avatars with smaller hips and waists relative to themselves. On the other hand, more conscientious individual is likely to regard their avatars as more physically alike to themselves. No comment.

For agreeableness, more agreeable individuals thought their avatars as more likely to themselves, but this was statistically marginal. On the other hand, they found an interaction effect in that highly agreeable men created avatars with larger chests and hips whereas highly agreeable women tend to created avatar with smaller chests and hips. The authors argued that this goes in line with conforming to gender ideals as creating these avatars served a purpose of engendering likeability. But again, there is no explicit link made in this study between gender ideal conformity, agreeableness and impression management. This is especially important in considering online social networking environments, such as Second Life, MMORPGs among others.

Finally for self-esteem, they found that individuals with low self-esteem are likely to create darker (or tanned) avatars relative to themselves and found that they regard their avatars more discrepant relative to individuals with higher self-esteem. The authors reasoned that low self-esteem individuals created darker/tanned avatars to compensate for their pale skin, especially since the majority of their sample is Caucasian.

The take home message the authors offered is that people self-represent in the virtual world as themselves rather than an invented persona. Participants’ avatar impression management strategy is that their avatar should be at their most positive light according to their personality traits and gender conformity norms. This is reflected by individuals with low self-esteem or highly neurotic creating more discrepant avatars in order to compensate their real life shortcomings.

Dunn, R. A., & Guadagno, R. E. (2012). My avatar and me – gender and personality predictors of avatar-self discrepancy. Computers in Human Behavior, 28 (1), 97-106. DOI:10.1016/j.chb.2011.08.015

5 thoughts on “Your avatar and you: Personality and gender predictors of avatar-self discrepancy (Dunn & Guadagno, 2012)

  1. Pingback: Designer Weekly July6 » Live. Play. Design

  2. Interesting. Did the study say anything about people playing as avatars not consistent with their own gender?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s