Effects of Avatar Physique and Online Environment on Avatar Personality

To be presented at SWPA 2008 by Scharff, L., Sylvester, R., Dyess, J. and Lowe, O.

Introduction: Multi-player online interactions allow new approaches for psychological research in both fantasy environments such as World of Warcraft and more realistic environments such as Second Life.  Such interactions often require an avatar, which may or may not resemble the human user controlling it. Does the choice of avatar characteristics influence the personality characteristics subsequently assumed for the avatar?  If so, such assumptions may influence how the human controller interacts with others in the environment or how other players interact with the avatar.  Further, different online environments have different objectives for players.  Will the online environment influence the personality assumptions about the avatar?

Method: The current study measured how participant gender, avatar physique (muscular or average) and type of online environment (social or gaming) affected the extroversion, physique anxiety, and assertiveness scores (subscale of extroversion score) assigned to a male avatar. For each online environment, participants (22 males, 26 females) were told the objectives of interactions within the environment, to assume the avatar represented them in the environment, and to answer the questions assuming they were the avatar.  They then viewed static images of different views (front, side or back) of either the muscular or average build avatar superimposed on backgrounds from the appropriate online environment (e.g. a street scene for the social environment or a castle scene for the gaming environment). Interleaved with the static pictures were the randomly ordered NEO-PI-R extroversion scale and Physique anxiety scale questions (60 total).  After answering all questions for the first environment, participants repeated the procedure for the second environment (counterbalanced across participants). 

example muscular avatar gaming average avatar social environment
Figure 1 (left): Muscular avatar example shown in a gaming online environment; (right): Average build avatar example shown in a social online environment.

Results: Experience with neither online social nor gaming environments correlated with responses, so they were not used as covariates.  2 x 2 x 2 mixed ANOVAs showed no significant effects or interactions for either the assertiveness or extraversion scores, although there was a tendency for the muscular avatar to be assigned greater levels of assertiveness. The ANOVA indicated significantly greater physique anxiety for the avatar with average build, with tendencies for females to assume greater physique anxiety, especially in the gaming environment.

Discussion: Avatar physique does seem to influence the personality characteristics assigned to it, especially with respect to physique anxiety.  Future research should investigate if such assumptions also influence online behavior of the human avatar controller, as well as other online players. The trend for females to assign greater levels of physique anxiety to the avatar of average build, especially in the gaming environment suggests that females relied more on stereotypes when evaluating the avatar. 

We would like to acknowledge the rest of the class members who also helped collect data and join in discussion of the project: Robin Adams, Matthew Akers, Blair Elder, Janette England, Stephanie Gerber, Laura Kaufman, Tracy Latham, Kevin Mahoney, Jeannae McCray, Joy McMullen, Joshua Pace, Lucy Partales, Jessica Runnels, Kari Shird, and Shawn Spears.

Return to:  Experimental Psychology front page || Dr. Scharff's front page.