2014 Know Thy Avatar: The Unintended Effect of Virtual-Self Representation on Behavior

The online version of this article can be found at:
http://pss.sagepub.com/content/25/4/1043

PDF here: http://rpgresearch.com/documents/other-projects/Know-Thy-Avatar-Psychological%20Science-2014-Yoon-1043-5.pdf/view

Virtual environments enable people to experience extra­
ordinary identities or circumstances. People can take on
superhero or super-villain roles using digital avatars in
virtual space. By acting as these avatars, individuals may
learn new behaviors and model their own, real-life
behaviors after them (Bandura, 1977; Bem, 1972). The
virtual environment is, thus, a vehicle for observation,
imitation, and modeling; players’ avatars may fuel these
processes.
Recent empirical research confirms that the behavior
of players’ avatars can affect players’ self-concepts, cogni­
tions, and feelings (Gentile et al., 2009; Greitemeyer &
Osswald, 2010). Thus, concepts related to avatar behav­
iors in general (e.g., fighting against evil) or to particular
avatars (e.g., Superman) may affect subsequent behavior
(e.g., good deeds). Identification with an avatar is corre­
lated with avatar-consistent behavior in the real world
(Rosenberg, Baughman, & Bailenson, 2013). In the exper­
iments reported here, we investigated whether certain
types of avatars and avatars’ behaviors could promote
pro- or antisocial actions in everyday behavior.
Experiment 1
One hundred ninety-four undergraduates participated in
the experiment (95 males, 99 females; mean age = 20.34
years, SD = 2.10). Participants were told that they were
involved in two separate studies: a test of game usability
and a blind tasting test. After signing a consent form, par­
ticipants were randomly assigned to heroic (Superman),
villainous (Voldemort), and neutral geometric-shaped (cir­
cle) avatars. They then played a game for 5 min in which
they battled enemies as their avatar (see the Supplemental
Material available online for further descriptions of the
stimuli and the game). Participants’ identification with their
avatars was measured using four items (e.g., “While you
were playing the game, how much did you identify with
your avatar?”; Cronbach’s α = .75), each of which they
responded to on a 7-point scale ranging from 1, not at all,
to 7, very much. Then they were told that the first study
was over.
Participants were then informed that a blind taste test
of food additives would take place and were asked to sign
another consent form.1 We manipulated good and bad
action by asking participants to first taste and then give
either chocolate or chili sauce, respectively (Fischer,
Kastenmüller, & Greitemeyer, 2010), to a (fictional) future
participant. Participants were instructed to pour an
unspecified amount of food into a plastic dish (“to allow
the experimenter to be blind to experimental conditions”)
and were told that the future participant would consume
all of the food provided. The total amount poured was
measured in grams. As a manipulation check, we asked
participants to rate the valence of giving chili sauce or
chocolate to a subsequent participant on scales ranging
from 1, bad, to 7, good; 1, unpleasant, to 7, pleasant; and
1, unfavorable, to 7, favorable (Cronbach’s α = .97).
Participants’ identification with their avatars did not
differ significantly among conditions, F(2, 191) = 1.83,
p = .16, η 2 = .02 (hero: M = 3.90, SD = 1.30; villain: M =
p3.54, SD = 1.19; circle: M = 3.88, SD = 1.38). The bad
action (giving chili sauce; M = 1.82, SD = 0.70) was rated
more negatively than the good action (giving chocolate;
M = 5.33, SD = 0.89), t(192) = 30.42, p < .001, d = 4.40.
More important, a 3 (avatar: heroic vs. villainous vs.
neutral) × 2 (behavior: good vs. bad) analysis of variance
(ANOVA) revealed the predicted interaction effect, F(2,
188) = 35.91, p < .001, η 2 = .28, but no main effects, Fs <
p.37, ps > .16. Participants who played the heroic avatar
gave more chocolate than those who played the villain­
ous or neutral avatars (see Fig. 1a). Conversely, partici­
pants who played villains poured more chili sauce than
Corresponding Author:
Gunwoo Yoon, Institute of Communications Research, College of
Media, University of Illinois at Urbana-Champaign, 810 South Wright
St., Urbana, IL 61801
E-mail: gyoon3@illinois.edu

Psychological Science
http://pss.sagepub.com/
KnowThy Avatar: The Unintended Effect of Virtual-Self Representation on Behavior
Gunwoo Yoon and Patrick T. Vargas
Psychological Science 2014 25: 1043 originally published online 5 February 2014
DOI: 10.1177/0956797613519271
The online version of this article can be found at:
http://pss.sagepub.com/content/25/4/1043
Published by:
http://www.sagepublications.com
On behalf of:
Association for Psychological Science
Additional services and information for Psychological Science can be found at:
Email Alerts: http://pss.sagepub.com/cgi/alerts
Subscriptions: http://pss.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Apr 9, 2014
OnlineFirst Version of Record - Feb 5, 2014
What is This?
Downloaded from pss.sagepub.com at EASTERN WASHINGTON UNIV on May 26, 2014
research-article2014
519271
PSSXXX10.1177/0956797613519271Yoon, VargasKnowThyAvatar
Short Report
Psychological Science
Know Thy Avatar: The Unintended Effect of
© 2014, The Vol. Author(s) 25(4) 1043­–1045
2014
Virtual-Self Representation Behavior
sagepub.com/journalsPermissions.nav
DOI: Reprints 10.1177/0956797613519271
and permissions:
onpss.sagepub.com
Gunwoo Yoon1 and Patrick T. Vargas2
1
Institute of Communications Research, College of Media, and 2Charles H. Sandage Department
of Advertising, College of Media, University of Illinois at Urbana-Champaign
Received 10/26/13; Revision accepted 12/11/13
Virtual environments enable people to experience extra­
ordinary identities or circumstances. People can take on
superhero or super-villain roles using digital avatars in
virtual space. By acting as these avatars, individuals may
learn new behaviors and model their own, real-life
behaviors after them (Bandura, 1977; Bem, 1972). The
virtual environment is, thus, a vehicle for observation,
imitation, and modeling; players’ avatars may fuel these
processes.
Recent empirical research confirms that the behavior
of players’ avatars can affect players’ self-concepts, cogni­
tions, and feelings (Gentile et al., 2009; Greitemeyer &
Osswald, 2010). Thus, concepts related to avatar behav­
iors in general (e.g., fighting against evil) or to particular
avatars (e.g., Superman) may affect subsequent behavior
(e.g., good deeds). Identification with an avatar is corre­
lated with avatar-consistent behavior in the real world
(Rosenberg, Baughman, & Bailenson, 2013). In the exper­
iments reported here, we investigated whether certain
types of avatars and avatars’ behaviors could promote
pro- or antisocial actions in everyday behavior.
Experiment 1
One hundred ninety-four undergraduates participated in
the experiment (95 males, 99 females; mean age = 20.34
years, SD = 2.10). Participants were told that they were
involved in two separate studies: a test of game usability
and a blind tasting test. After signing a consent form, par­
ticipants were randomly assigned to heroic (Superman),
villainous (Voldemort), and neutral geometric-shaped (cir­
cle) avatars. They then played a game for 5 min in which
they battled enemies as their avatar (see the Supplemental
Material available online for further descriptions of the
stimuli and the game). Participants’ identification with their
avatars was measured using four items (e.g., “While you
were playing the game, how much did you identify with
your avatar?”; Cronbach’s α = .75), each of which they
responded to on a 7-point scale ranging from 1, not at all,
to 7, very much. Then they were told that the first study
was over.
Participants were then informed that a blind taste test
of food additives would take place and were asked to sign
another consent form.1 We manipulated good and bad
action by asking participants to first taste and then give
either chocolate or chili sauce, respectively (Fischer,
Kastenmüller, & Greitemeyer, 2010), to a (fictional) future
participant. Participants were instructed to pour an
unspecified amount of food into a plastic dish (“to allow
the experimenter to be blind to experimental conditions”)
and were told that the future participant would consume
all of the food provided. The total amount poured was
measured in grams. As a manipulation check, we asked
participants to rate the valence of giving chili sauce or
chocolate to a subsequent participant on scales ranging
from 1, bad, to 7, good; 1, unpleasant, to 7, pleasant; and
1, unfavorable, to 7, favorable (Cronbach’s α = .97).
Participants’ identification with their avatars did not
differ significantly among conditions, F(2, 191) = 1.83,
p = .16, η 2 = .02 (hero: M = 3.90, SD = 1.30; villain: M =
p3.54, SD = 1.19; circle: M = 3.88, SD = 1.38). The bad
action (giving chili sauce; M = 1.82, SD = 0.70) was rated
more negatively than the good action (giving chocolate;
M = 5.33, SD = 0.89), t(192) = 30.42, p < .001, d = 4.40.
More important, a 3 (avatar: heroic vs. villainous vs.
neutral) × 2 (behavior: good vs. bad) analysis of variance
(ANOVA) revealed the predicted interaction effect, F(2,
188) = 35.91, p < .001, η 2 = .28, but no main effects, Fs <
p.37, ps > .16. Participants who played the heroic avatar
gave more chocolate than those who played the villain­
ous or neutral avatars (see Fig. 1a). Conversely, partici­
pants who played villains poured more chili sauce than
Corresponding Author:
Gunwoo Yoon, Institute of Communications Research, College of
Media, University of Illinois at Urbana-Champaign, 810 South Wright
St., Urbana, IL 61801
E-mail: gyoon3@illinois.edu
Downloaded from pss.sagepub.com at EASTERN WASHINGTON UNIV on May 26, 2014
1044    
Yoon, Vargas
a
20
Chocolate
Chili Sauce
b
16
Heroic Avatar
Villainous Avatar
)
g(detacollAtnuomA16
12
8
4
0
Heroic Avatar
Neutral Avatar
Villainous Avatar
Virtual-Self Representation
)
g(d 12
etacollA 8
ecuaSl i 4
ihC0
Player
Role
Observer
Fig. 1. Results from Experiment 1 (a) and Experiment 2 (b). The graph in (a) shows the mean amount of chocolate and chili sauce given by
participants as a function of the avatar they used. The graph in (b) shows the mean amount of chili sauce given as a function of participants’
role and the avatar they used. Error bars show standard errors of the mean.
did participants who played heroes and neutral avatars
(see the Supplemental Material for further details).
Experiment 2
The design of Experiment 2 was the same as that of
Experiment 1, except that there was an additional set of
conditions to test whether our role-taking manipulation
(i.e., playing as a superhero or villain) generated stronger
real-world outcomes than common behavioral-priming
(e.g., Dijksterhuis et al., 1998) and perspective-taking
(e.g., Galinsky, Wang, & Ku, 2008) manipulations, both of
which lead people to behave in ways consistent with the
target. In order to simplify the experiment, we dropped
the neutral-avatar condition and focused on how much
chili sauce players allotted. Thus, we tested whether
game players (who “were” the heroic or villainous avatar)
would show stronger behavioral effects than observers
(who “were primed with” or “took the perspective of” the
heroic or villainous avatar). Observers were asked to put
themselves in a heroic or villainous avatar’s “shoes” and
to watch a game demonstration for 5 min. All other pro­
cedures and measures were identical to those used in
Experiment 1.
One hundred twenty-five undergraduates partici­
pated in the experiment (44 males, 81 females; mean
age = 19.42 years, SD = 1.37). A 2 (avatar: heroic vs.
villainous) × 2 (role: player vs. observer) ANOVA on the
amount of chili sauce served yielded a significant main
effect of avatar, F(1, 121) = 48.35, p < .001, η 2 = .29, and
pno effect of role, F(1, 121) < 1, p = .57. As predicted,
villains administered a greater amount of hot chili sauce
than heroes (see the Supplemental Material for further
details). There was a significant interaction effect,
F(1, 121) = 24.17, p < .001, η 2 = .17. Indeed, partici­
ppants who played heroes served significantly less chili
sauce than participants who observed heroes, and par­
ticipants who played villains served more chili sauce
than participants who observed villains (Fig. 1b; see
the Supplemental Material for further details). AsExperiment 1, “being” an avatar caused participantsbehave in ways that conformed to their avatars, which
caused stronger effects on subsequent behavior than
did priming or perspective taking.
in
to
Discussion
This research not only demonstrates that acting as a hero
or villain causes people to perform coincident behaviors,
but also highlights that role taking facilitates behavior
consistent with the actions of a target above and beyond
the behavior facilitated by priming (Nelson & Norton,
2005). A 5-min gaming experience with certain avatars
is enough to reverse a potential pattern of behavior.
One likely explanation is that immersion (Weinstein,
Przybylski, & Ryan, 2009) or arousal (Berger, 2011)
derived from the gaming experience imbues people with
agency. In Experiment 2, the perspective-taking manipu­
lation was almost certainly less arousing than active game
play; perhaps arousal mediates the effect of avatars on
behavior (e.g., arousal facilitates action).2
Human social responses can be altered by how virtual-
self representations are implemented, and those can play
a role in shaping the way people interact with others. In
Downloaded from pss.sagepub.com at EASTERN WASHINGTON UNIV on May 26, 2014
Know Thy Avatar    1045
everyday gaming, players choose their own avatars, but
creating games with more heroic avatars could encour­
age more prosocial behavior. By exploring the important
outcome of virtual experiences, this study broadens the
potential of unintended influence of self-representation
derived from role taking on human behavior.
Author Contributions
G. Yoon developed the study concept. G. Yoon and P. T.
Vargas equally contributed to the study design. Data collection
and statistical analyses were performed by G. Yoon and P. T.
Vargas. G. Yoon analyzed and interpreted the data under the
supervision of P. T. Vargas. G. Yoon drafted the manuscript,
and P. T. Vargas provided critical revisions. All authors approved
the final version of the manuscript for submission.
Acknowledgments
The authors thank Jesse Preston and members of the Psychology
of Religion, Agency, and Morality Laboratory for their useful
comments.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with
respect to their authorship or the publication of this article.
Supplemental Material    
Additional supporting information may be found at http://pss
.sagepub.com/content/by/supplemental-data
Notes
1. We handed out different consent forms to make the idea of
“two separate studies” more believable. No participants noticed
a connection between the two consecutive studies.
2. We tested mediation effects linking role taking to behavior
through avatar identification. Results from bootstrap analyses
did not support this avatar-identification account. In addition,
there were no significant correlations between avatar identifica­
tion and behavior.
References
Bandura, A. (1977). Social learning theory. Englewood Cliffs,
NJ: Prentice Hall.
Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.),
Advances in experimental social psychology (pp. 1–62).
New York, NY: Academic Press.
Berger, J. (2011). Arousal increases social transmission of infor­
mation. Psychological Science, 22, 891–893.
Dijksterhuis, A., Spears, R., Postmes, T., Stapel, D., Koomen,
W., van Knippenberg, A., & Scheepers, D. (1998). Seeing
one thing and doing another: Contrast effects in automatic
behavior. Journal of Personality and Social Psychology, 75,
862–871.
Fischer, P., Kastenmüller, A., & Greitemeyer, T. (2010). Media
violence and the self: The impact of personalized gaming
characters in aggressive video games on aggressive behav­
ior. Journal of Experimental Social Psychology, 46, 192–195.
Galinsky, A. D., Wang, C. S., & Ku, G. (2008). Perspective-
takers behave more stereotypically. Journal of Personality
and Social Psychology, 95, 404–419.
Gentile, D. A., Anderson, C. A., Yukawa, S., Ihori, N., Saleem,
M., Ming, L. K., . . . Sakamoto, A. (2009). The effects of pro­
social video games on prosocial behaviors: International
evidence from correlational, longitudinal, and experimen­
tal studies. Personality and Social Psychology Bulletin, 35,
752–763.
Greitemeyer, T., & Osswald, S. (2010). Effects of prosocial
video games on prosocial behavior. Journal of Personality
and Social Psychology, 98, 211–221.
Nelson, L. D., & Norton, M. I. (2005). From student to super­
hero: Situational primes shape future helping. Journal of
Experimental Social Psychology, 41, 423–430.
Rosenberg, R. S., Baughman, S. L., & Bailenson, J. N. (2013).
Virtual superheroes: Using superpowers in virtual real­
ity to encourage prosocial behavior. PLoS ONE, 8(1),
e55003. Retrieved from http://www.plosone.org/article/
info:doi/10.1371/journal.pone.0055003
Weinstein, N., Przybylski, A. K., & Ryan, R. M. (2009). Can
nature make us more caring? Effects of immersion in nature
on intrinsic aspirations and generosity. Personality and
Social Psychology Bulletin, 35, 1315–1329.