See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/347014896

White Man’s Virtual World: A Systematic Content Analysis of Gender and Race
in Massively Multiplayer Online Games

Article in Journal of Virtual Worlds Research · May 2014

DOI: 10.4101/jvwr.v7i2.7096

CITATIONS READS

44 488

5 authors, including:

James D Ivory
Virginia Tech (Virginia Polytechnic Institute and State University)
58 PUBLICATIONS 2,536 CITATIONS

SEE PROFILE

All content following this page was uploaded by James D Ivory on 19 January 2021.

The user has requested enhancement of the downloaded file.

 Volume 7, Number 2
Assembled 2014
May 2014
Editor In-Chief Yesha Sivan,
Tel Aviv-Yaffo Academic College, Israel

Issue Editors Samuel Cruz-Lara,

LORIA (UMR 7503) CNRS - INRIA -
Université de Lorraine, France

Sue Gregory,
University of New England, Australia

Suely Fragoso,
Federal University of Rio Grande do Sul,
Brazil

Urs-Vito Albrecht,
Braunschweig University & Hannover
Medical School, Germany

Christopher Lueg,
University of Tasmania, Australia

Coordinating Editor Tzafnat Shpak

The JVWR is an academic journal. As such, it is dedicated to the open exchange of information. For this reason, JVWR is
freely available to individuals and institutions. Copies of this journal or articles in this journal may be distributed for research
or educational purposes only free of charge and without permission. However, the JVWR does not grant permission for use
of any content in advertisements or advertising supplements or in any manner that would imply an endorsement of any
product or service. All uses beyond research or educational purposes require the written permission of the JVWR. Authors
who publish in the Journal of Virtual Worlds Research will release their articles under the Creative Commons Attribution No
Derivative Works 3.0 United States (cc-by-nd) license. The Journal of Virtual Worlds Research is funded by its sponsors and
contributions from readers.
http://jvwresearch.org Gender and Race in MMO Games 1

Volume 7, Number 2
Assembled 2014
May, 2014

White Man’s Virtual World:

A Systematic Content Analysis of Gender and Race
in Massively Multiplayer Online Games

T. Franklin Waddell
College of Communications,
Pennsylvania State University, University Park, Pennsylvania, USA

James D. Ivory Rommelyn Conde Courtney Long Rachel McDonnell

Department of Communication,
Virginia Tech, Blacksburg, Virginia, USA

Abstract
Based on previous research indicating that character portrayals in video games and other media
can influence users’ perceptions of social reality, systematic content analyses have examined
demographic trends in the way video game characters are portrayed. Although these studies have
extensively documented character portrayals in traditional console and computer video games, there is a
lack of content analyses examining character portrayals in the very popular massively multiplayer online
game (MMO) genre. Such studies are needed because many characters in MMOs are customized avatars
created by users, which may lead to different trends in character demographics. This content analysis
examined representations of gender and race among 417 unique characters appearing 1,356 times in 20
hours of recorded content from four popular commercial MMOs, which was generated by five recruited
users. Characters tended to be disproportionately male and white, with females and racial minorities
appearing much less often. Implications for potential effects on users’ perceptions of social reality are
discussed.

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 2

1. Introduction
Increasingly, video game users have seen their entertainment options expand beyond traditional
console-based games to include more and more persistent massively multiplayer online games (MMOs).
These online games occupy a substantial amount of leisure time for millions of game players. Recent
research indicates that the average MMO user spends more than 25 hours a week using the popular
online games, with MMO use often displacing game players’ exposure to other traditional media
(Williams, Yee, & Caplan, 2008).
There is evidence in the existing literature that sexism and racism are noteworthy concerns
among some video game player communities (Fox & Tang, in press; Kolko, Nakamura, & Rodman,
2012; Salter & Blodgett, 2012). While the cultural causes of such intolerance may be complex, a
possible socializing influence on video game users’ stereotypical perceptions and behavior regarding
gender and race may be in the content of the games they play. Given that MMOs are a substantial part of
some users’ media diets, MMO play may have a socializing influence on some heavy players’
perceptions of social reality. For example, previous longitudinal investigations have found that MMO
players’ perceptions of real-life crime may be shaped by the extent of their exposure to violence
encountered in an MMO (Williams, 2006). Video game use has been shown to foster the development of
gender stereotypes (Behm-Morawitz & Mastro, 2009) and to threaten the racial identity of non-white
players (Lee & Park, 2011).
These results are not surprising given a large body of content analyses finding disproportionate
portrayals of gender and race in console-based video games, with male and white characters
overrepresented at the expense of female characters and characters representing most racial minorities
(e.g., Beasley & Standley, 2002; Dietz, 1998; Williams, Martins, Consalvo, & Ivory, 2009). However,
while the content of traditional video game use has been extensively investigated, the content of MMOs
has been relatively underexplored. This is a shortcoming in the existing research on video game content,
not only because MMOs are very popular, but because MMOs provide users with an unprecedented role
in determining the game characters’ demographic makeup through the creation of user-customized
avatars.
Although similar to many traditional console-based games in terms of theme and plot, MMOs
are substantially different from console-based games because they are populated at any given time by
thousands of avatars whose appearance is determined by the game users. While most console-based
games involve one or a few user-controlled characters interacting with a large group of computer-
controlled characters created by the game producers, a mix of computer-controlled agent characters and
customized user-controlled avatars inhabit MMOs. These MMOs allow users to control the way they
represent themselves in the games’ virtual environments by personalizing their avatars’ appearance
along a multitude of potential characteristics. For example, MMO users can vary the appearance of their
avatars by choosing gender, skin tone, hairstyle, and height (Ducheneaut, Wen, Yee & Wadley, 2009).
The opportunity to manipulate one’s identity in MMOs has important implications for research on the
representation of gender and race in online video games, as Williams, Martins, Consalvo and Ivory
(2009) argue that user avatars represent an increasingly important source of character portrayals for
online games. However, the portrayal of characters in MMOs has been more or less unexamined in
systematic content analyses.
The present study addressed this gap in existing research through a systematic content analysis of
gender and race among 417 unique characters appearing 1,356 times in 20 hours of content from four
popular commercial MMOs. This analysis is intended to provide baseline data about portrayals of

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 3

gender and race in MMOs to tell us more about what content users encounter in these popular games and
what possible social effects of these games’ content might be worth investigating in the future. After
reviewing previous literature related to the representation of gender and race in traditional and online
video games, this paper describes the research questions investigated by this study, details the sample
and coding protocol used to complete the content analysis, reports the study’s findings, and discusses the
implications and limitations of the study.

2. Literature Review
2.1 Representations of Gender and Race in Traditional Console-Based Video Games
Media can be a powerful force in shaping perceptions of social reality, including media users’
perceptions of gender roles and the social role of ethnic groups (Behm-Morawitz & Mastro, 2009;
Comstock & Cobbey, 1979; Mastro & Behm-Morawitz, 2005; Mastro, Behm-Morawitz, & Ortiz, 2007;
Rivadeneyra, Ward, & Gordon, 2007). Just as representations of gender and race are often unrealistic in
media such as television (Harwood & Anderson, 2002) and magazines (Soley and Reid, 1988), the
distribution of gender and race in video games is frequently unrepresentative. For example, a
comprehensive content analysis of video game characters across 150 popular commercial video games
found that male and white characters were overrepresented at the expense of women and most other
ethnic groups (Williams et al., 2009). Incongruities were further heightened when differentiating
between primary and secondary characters, as white males were even more likely to be featured as the
primary protagonist controlled by the user. Underrepresentation of women and non-white ethnic groups
has also been observed by earlier content analyses of gender and race in video games which have found
only 13% of characters were female (Beasley & Standley, 2002; Dietz, 1998) and more than 87% of
leading characters were white (Heintz-Knowles, Henderson, Glaubke, Miller, Parker, & Espejo, 2001).
These trends in character representation have been attributed to the corresponding demographics
of game designers, as Williams (2006) argues the absence of non-white, female characters in games can
be explained by the absence of females and minority designers in the gaming industry. Given this
assertion and the consistently disproportionate representations of gender and race in previous studies of
video game content, it is possible that the demographic representations of characters in console-based
games may remain disproportional until there is a substantial change in the demographic makeup of
people involved in game design and production. However, it is possible that these trends in character
portrayals may change in the case of MMOs because of the users’ role in determining the characteristics
of many of the characters encountered in those games.

2.2 A New Frontier: Character Demographics in MMOs and other Virtual Worlds
MMOs feature regular interaction between user-created avatar characters online. Although the
extent of customization options for users’ avatar characters varies between MMO games, the ability to
choose the skin tone and gender of one’s avatar is a common feature across MMO games (Ducheneaut
et al., 2009). Given that the prevalence of customized user avatars has given MMO users an
unprecedented role in determining the demographics of video game characters, previous findings about
representations of gender and race in traditional console-based games may not be generalizable to the
content of MMO games. While prior research has yet to systematically examine portrayals of character
gender and race in MMO games, some studies have examined relationships between MMO users’
characteristics and the way those users represent themselves with MMO avatar characters (Bessière,
Seay, & Kiesler, 2007; Duchenaut et al., 2009; Huh & Williams, 2008; Hussain & Griffiths, 2008).

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 4

Some investigations of avatar customization choices in MMOs have focused on the concept of
gender bending, or the frequency with which users choose to employ avatars with genders opposite to
their offline biological sex. Estimations of gender bending’s occurrence fluctuate across different studies
of online games, ranging from 15% of users in EverQuest (Huh & Williams, 2008) to more than half of
all male players in World of Warcraft (Hussain & Griffiths, 2008; Yee, Ducheneaut, Yao, & Nelson,
2011). Subsequent research additionally suggests that the identities users choose vary according to the
demographics and related goals of the game environment (Ducheneaut et al., 2009; Nowak & Rauh,
2008). For example, Ducheneaut and colleagues (2009) suggest that users of World of Warcraft choose
opposite gendered avatars for the purpose of standing out within a predominantly male population, while
users of Second Life may prefer an idealized, same-gender virtual self to mediate interactions in a more
gender-balanced population. Phenomena such as gender-bending by MMO users further underscore the
possible need for a thorough examination of the gender makeup of virtual world characters (Huh &
Williams, 2008; Hussain & Griffiths, 2008). Therefore, we ask:
RQ1: How frequently are male and female characters represented across massively multiplayer
online games?
RQ2: Does the representation of male and female characters vary between massively multiplayer
online games?
In addition, previous studies have yet to examine the distribution of character race among
characters in MMOs. Though previous research on user demographics in the MMO EverQuest II found
that 87% of MMO players were white (Williams et al., 2008), it remains unclear whether users’ replicate
their offline race or engage in virtual identity exploration through the selection of avatar skin color. As a
result, we ask:
RQ3: How frequently are different ethnic groups represented across massively multiplayer online
games?
RQ4: Does the representation of ethnic groups vary between massively multiplayer online games?

3. Method
3.1 Sample
Characters from a total of four commercially popular massively multiplayer online games were
analyzed. These games were chosen based on sales figures from the NPD Group for 2010 (NPD Group
Data, 2010), which identified World of Warcraft, Guild Wars, Dungeons and Dragons Online, and
RuneScape as the most popular MMO games for that year. Given that the four games selected accounted
for 61% of the North American MMO market (Bayer, 2010), these four MMOs represent the majority of
the MMO content encountered by MMO users during that year.
Five undergraduate students recruited from a student gaming organization at a major university
were each randomly assigned to play two of the four MMOs used in this study. These students had
extensive prior experience with MMO games, which is consistent with the type of game player recruited
to generate game content in prior content analysis research (Beasley & Standley, 2002; Downs & Smith,
2010). For each game, 2.5 hours of game content were sampled from each of two separate players,
resulting in five total hours of content for each game and a total sample of 20 hours of content across the
four separate MMOs. Although previous studies have typically recorded segments of 20 to 30 minutes
for each video game included in their samples (e.g., Beasley & Standley, 2002; Downs & Smith, 2010;

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 5

Williams et al., 2009), this study included five hours of content per MMO game given the longer time
commitments characteristic to MMO gameplay (Yee, 2006). Two users were assigned to play each
MMO to ensure that the content included in the sample for each game represented the game experience
of multiple users rather than the idiosyncratic experience of only one user. Prior research has typically
only employed one player to generate game content (Beasley & Standley, 2002; Downs & Smith, 2010;
Williams et al., 2009). The sex, physical appearance, name, and server of the users’ character avatars
were also randomly generated so that any avatar or server-dependent game content (e.g., instances
where users’ characters start the game in a region populated by characters with similar appearance)
would be determined randomly and not by a user decision. The users’ game sessions were digitally
recorded on a high-end desktop computer for later coding.

3.2 Unit of Analysis

Data for variables under study were coded for each unique game character that a user’s character
avatar encountered during a 2.5-hour MMO game session. A character was included in the sample if a
user’s character avatar interacted with that character or viewed the character’s basic information using
the game interface. Specifically, the protocol guiding whether a character was coded was
operationalized as to whether that character’s avatar and name icon appeared as “selected” in the game’s
user interface, a sign that the user’s character was somehow engaged with that character in the game
environment. The sample included 417 such characters encountered in the 20 hours of MMO content
sampled.

3.3 Coded Variables

Gender. Each character was coded as “male,” “female,” or “not applicable.” Non-human
anthropomorphic characters without an apparent gender (e.g., animals; characters with a non-human-like
morphology) were coded as “not applicable.” This treatment of quasi-human characters is consistent
with the protocol of previous video game content analyses (e.g., Downs & Smith, 2010; Williams et al.,
2009).
Race. Characters were assigned an apparent race based on the heuristic cue of avatar skin color
and subsequently coded as “White,” “Black,” “Hispanic,” “Asian/Pacific Islander,” or “Native
American.” Nonhuman anthropomorphic characters without features consistent with human racial
characteristics (e.g., skin tone variation not typically present among humans) were coded as “not
applicable.” This protocol is consistent with the treatment of race utilized by previous content analyses
(e.g., Downs & Smith, 2010; Williams et al., 2009).
Number of appearances. The number of times each unique character appeared in a game
session was recorded so that analyses could account for the unequal prevalence of different characters’
appearances in the recorded game content. This variable was operationalized as the number of separate
times the same character (or an identical version of that character in the case of some computer-
controlled characters) appeared during users’ 2.5-hour game session.
Character type. Characters were identified as either a customized avatar character controlled by
another game user or a computer-controlled agent character created by the game designer. This
distinction was made using information provided in the games, such as the text and icons accompanying
characters in the game display.

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 6

3.4 Coding and Reliability

A total of four graduate student coders were trained in the coding protocols and completed
coding for all variables for all characters in the 20 hours of recorded MMO content. Each 2.5-hour game
session was randomly assigned to one of the coders, with two sessions randomly assigned to all coders
(without the coders’ knowledge as to which sessions were assigned to multiple coders) to allow
assessment of intercoder reliability.
Scott’s π was used to assess intercoder reliability for the categorical measures of character
gender (π = .76) and race (π = .81). Pearson’s r was used to determine intercoder reliability for the
continuous measure of total number of character appearances (r = .81). Intercoder reliability for all
variables was acceptable.

4. Results
4.1 Character Gender
The first research question asked how frequently male and female characters were represented in
the overall sample. Analyses were first carried out for unique characters, then repeated for the total
number of appearances by all characters. As Figure 1 shows, male characters represented 62.59% (n =
261, 95% CI = 57.85%-67.09%) of the unique characters in the sample, female characters represented
13.67% (n = 57, 95% CI = 10.70%-17.29%) of the unique characters in the sample, and 23.74% (n = 99,
95% CI = 19.91%-28.05%) of the characters in the sample did not have an apparent gender. This means
that of the 318 characters who had an apparent gender, 82.08% (n = 261, 95% CI = 77.49%-85.90%)
were male and 17.93% (n = 57, 95% CI = 14.10%-22.52%) were female.

Figure 1: Gender representation for unique characters and total character appearances,
all games (n = 417 characters, 1,356 appearances).
Gender prevalence for the total appearances among all characters followed a similar pattern.
Male characters accounted for 57.38% (n = 778, 95% CI = 31.74%-36.79%) of total character
appearances, female characters accounted for 8.41% (n = 114, 95% CI = 7.05%-10.00%) of total
character appearances, and 34.22% (n = 464, 95% CI = 31.74%-36.79%) of total character appearances

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 7

in the sample were by characters with no apparent gender. This means that of the 892 appearances by
characters who had an apparent gender, 87.22% (n = 778, 95% CI = 84.87%-89.25%) were appearances
by male characters and 12.78% (n = 114, 95% CI = 10.75%-15.13%) were appearances by female
characters.
The second research question asked if gender representation was consistent across the four games
that constituted the sample. Analyses were first carried out for unique characters, then repeated for the
total number of appearances by all characters. A chi-square analysis comparing prevalence of unique
characters’ gender across the MMO games found significant differences in gender prevalence between
games (χ² = 14.34, p = .026). As Table 1 shows, all games featured a majority of male characters,
though the prevalence of characters with no apparent gender and female characters differed somewhat
across games. A chi-square analysis comparing total character appearances across games also found
significant differences in gender prevalence between games (χ² = 205.33, p < .001). All of the games in
the sample featured more appearances by male characters than by female characters, but the total
proportion of male character appearances was particularly high for Dungeon and Dragons Online, the
total proportion of female character appearances was relatively high (though still less than half) for
Guild Wars, and the proportion of appearances by characters with no apparent gender was particularly
high for World of Warcraft.

Table 1: Gender representation for unique characters

and total character appearances across games (n = 417 characters, 1,356 appearances).

Unique Characters Total Appearances

Game Male Female None Male Female None
D & D Online 61.5% 19.2% 19.2% 73.0% 5.5% 21.5%
Guild Wars 66.3% 16.8% 16.8% 66.7% 17.4% 17.4%
RuneScape 61.3% 13.3% 25.3% 63.2% 8.7% 28.1%
World of Warcraft 61.3% 7.3% 31.4% 35.8% 7.4% 56.9%
Total (All Games) 62.6% 13.7% 23.7% 8.4% 57.4% 34.2%

4.2 Character Race

The third research question asked how frequently different ethnic groups were represented in the
overall sample. Analyses were first carried out for unique characters, then repeated for the total number
of appearances by all characters. As Figure 2 shows, white characters made up 43.41% (n = 181, 95% CI
= 38.73%-48.20%) of the unique characters in the sample, while black characters accounted for 3.84%
(n = 16, 95% CI = 2.38%-6.14%) of the sample and no human racial group was apparent for the other
52.76% (n = 220, 95% CI = 47.96%-57.50%) of the characters. Hispanics, Asian/Pacific Islanders, and
Native Americans were not represented by any of the characters in the sample. Although slightly more
than half of the characters were not identified with any apparent human racial group, this means that of
the 197 characters who did have an apparent racial group, 91.88% (n = 181, 95% CI = 87.22%-94.94%)
were white and 8.12% (n = 16, 95% CI = 5.06%-12.78%) were black.

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 8

Figure 2: Race representation for unique characters and total character appearances,
all games (n = 417 characters, 1,356 appearances).

Trends were similar for analyses of racial prevalence among the total appearances of all
characters. White characters accounted for 29.57% (n = 401, 95% CI = 27.20%-32.06%) of total
character appearances, black characters accounted for 7.89% (n = 107, 95% CI = 6.57%-9.45%) of total
character appearances, and 62.54% (n = 848, 95% CI = 59.93%-65.07%) of total character appearances
in the sample were by characters who were associated with no apparent racial group.
The fourth research question asked if racial representation was consistent across the four games
that constituted the sample. Analyses were first carried out for unique characters, then repeated for the
total number of appearances by all characters. A chi-square analysis found significantly different racial
distribution across games (χ²= 35.00 p < .001). As Table 2 shows, this was mainly due to some
differences between games as to whether White characters were represented more often than characters
with no apparent human racial association. For example, Dungeons and Dragons Online and RuneScape
featured a majority of white characters while Guild Wars and World of Warcraft featured a majority of
characters without an apparent human racial association. For all games, though, black characters
represented a small minority of characters, and World of Warcraft featured no black characters at all in
this particular sample.

Table 2: Race representation for unique characters

and total character appearances across games (n = 417 characters, 1,356 appearances).

Unique Characters Total Appearances

Game White Black None White Black None
D & D Online 52.9% 3.9% 43.3% 30.1% 17.6% 52.4%
Guild Wars 40.6% 7.9% 51.5% 43.5% 5.31% 51.2%
RuneScape 56.0% 5.3% 38.7% 44.4% 5.9% 49.7%
World of Warcraft 31.4% 0.0% 68.6% 18.0% 0.0% 82.0%
Total (All Games) 43.4% 3.8% 52.8% 29.6% 7.9% 62.5%

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 9

Repeating the chi-square analysis for all character appearances also identified significantly
different racial distribution across games (χ²= 216.80 p < .001). As with the results for unique
characters, this was mainly due to differences between games in terms of the relative prevalence of
appearances by White characters and by characters with no apparent human racial association, though
appearances by Black characters did constitute a larger minority of the appearances in Dungeons and
Dragons Online than in other games.

4.3 User-Controlled Avatar Characters and Computer-Controlled Agent Characters

Given that this research was motivated in part by interest in whether the presence of customized
user-controlled avatar characters would influence gender and racial representation among MMO
characters, supplementary analyses explored how many characters in the sample were user-controlled
avatar characters and how many of them were computer-controlled agent characters. User-controlled
characters actually made up only 4.07% (n = 17, 95% CI = 2.56%-6.43%) of the unique characters in the
sample, while computer-controlled characters comprised 95.92% (n = 400, 95% CI = 93.57%-97.44%)
of the unique characters. User-controlled characters accounted for only 2.29% (n = 31, 95% CI = 1.62%-
3.23%) of total character appearances, with computer-controlled characters making 97.71% (n = 1,325,
95% CI = 96.77%-98.39%) of the character appearances in the games.
Although user-controlled characters were not common in the sample, their gender and racial
makeup was slightly more equitable than the overall sample. As Figure 3 shows, 76.47% (n = 13, 95%
CI = 52.74%-90.45%) of user-controlled characters were male and 23.53% (n = 4, 95% CI = 9.56%-
47.26%) of user-controlled characters were female. None of the user-controlled characters had no
apparent gender. Male user-controlled characters accounted for 87.10% (n = 27, 95% CI = 71.15%-
94.87%) of appearances by user-controlled characters in the sample, while female user-controlled
characters accounted for the other 12.90% (n = 4, 95% CI = 51.34%-28.85%) appearances by user-
controlled characters.

Figure 3: Gender representation for unique user-controlled characters

and total user-controlled character appearances, all games (n = 17 characters, 31 appearances).

As shown in Figure 4, white user-controlled characters comprised 76.47% (n = 13, 95% CI =

52.74%-90.45%) of the total unique user-controlled characters, while another 11.77% (n = 2, 95% CI =
3.29%-34.34%) of the unique characters were black and a final 11.77% (n = 2, 95% CI = 3.29%-

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 10

34.34%) had no apparent human racial association. White user-controlled characters accounted for
77.42% (n = 24, 95% CI = 60.19%-88.61%) of total appearances by user-controlled characters, with
black characters accounting for 16.13% (n = 5, 95% CI = 7.09%-32.63%) of user-controlled character
appearances and user-controlled characters with no apparent human racial association making 6.45% (n
= 2, 95% CI = 7.09%-32.63%) of user-controlled character appearances.

Figure 4: Race representation for unique user-controlled characters

and total user-controlled character appearances, all games (n = 17 characters, 31 appearances).

5. Discussion
Though previous studies of gender and race in traditional console-based video games have found
games’ portrayals to be unrepresentative in favor of male characters and white characters, character
portrayals in the increasingly popular genre of massively multiplayer online games have previously been
relatively unexamined. An understanding of the content of these games is increasingly important, as
users of MMOs report playing online games more frequently than their use of television and other
popular media (Williams et al., 2008). Given that previous research has found that the extended use of
online video games can lead to the cultivation of offline perceptions of social reality, (Williams, 2006),
an accompanying examination of the popular but relatively under examined content of MMOs warrants
comprehensive analysis. This study attempted to expand our understanding of how video game
characters are represented in this novel game genre by identifying patterns of gender and race
representation in MMOs to determine whether these patterns mirror those found by other research or
whether MMOs exhibit new trends in the way they represent gender and race among their characters.
Specifically, the study examined the distribution of character gender and race across 20 hours of video
game content generated by five MMO players.
Generally, this study found that, as with research on console-based video games, characters in
MMOs tended to be disproportionately male and white. Most characters in the sampled content were
male, particularly among those characters who had an apparent gender. While many characters in the
sample did not have an apparent association with a human racial group, those who did were
overwhelmingly white, and black was the only other racial group represented at all. Although MMOs
differ from other video game genres in many ways, it appears that gender and race representation may
not be one of them. It should also be noted that these inequities cannot be explained away as simply a

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 11

reflection of the MMO user demographics. While the gender distribution of characters and character
appearances in this sample is perhaps comparable to, though in cases slightly more male than, the MMO
gender distributions of MMO users reported in recent surveys (e.g., Williams, Consalvo, Caplan, & Yee,
2009; Williams et al., 2008), the distribution of race among characters in our sample badly fails to
represent the ethnic diversity of MMO users.
While we had speculated that the presence of user-created and user-controlled avatar characters
might influence the gender and race makeup of MMO characters, user-controlled characters were
relatively rare in the sample. This suggests that users of MMOs encounter computer-controlled agent
characters much more frequently than user-controlled avatar characters, meaning that the decisions of
game designers and producers still play a very strong role in the character portrayals that MMO users
are exposed to while using MMO games. This finding is consistent with user behavior previously
observed in the popular MMO World of Warcraft which found many users prefer to be “alone together”
in virtual environments (Ducheneaut, Yee, Nickell, & Moore, 2006). Therefore, while users may enjoy
the presence of an audience of user-controlled avatar characters while completing their virtual conquests
(Ducheneaut et al., 2006), the most frequently-viewed character representations in MMOs may still be
determined by games’ designers and producers rather than by the users.
It also appears that the presence of user-controlled characters may not have a large effect on the
gender and racial makeup of MMOs. While the small group of user-controlled characters we
encountered was slightly less disproportionately male and white, male and white characters still
constituted a large majority of the user-controlled avatar characters in the sample. Although players of
MMOs are provided the opportunity to customize the gender and skin color of their own avatar
characters, users of MMOs in our sample seem to be—for the most part—replicating patterns of gender
and race representation common to computer-controlled characters in MMOs, as well as in other video
games, with their avatar character choices. Although the gender makeup of the small group of user-
controlled characters is generally comparable to that found in recent surveys of MMO users (e.g.,
Williams et al., 2009; Williams et al., 2008), it certainly does not represent the span of races represented
by MMO users. That said, a larger group of user-controlled character avatars should be sampled to
provide a more full understanding of how MMO users represent their characters’ gender and race.
Non-human characters appeared frequently in the current sample and were more likely to appear
than female or non-white characters. Prior research (Williams, 2006) has attributed the infrequent
appearance of female characters to the predominance of male game designers, although it less clear why
non-human characters are more likely to populate MMO games than non-white ethnic groups. While the
design motivations for the frequent appearance of non-human characters are beyond the scope of the
current study, the results nonetheless suggest that both creators and users of MMOs may further
perpetuate disproportionately male and white character representations that have been observed in video
game content for more than a decade. Although the precise implications of these unrepresentative
portrayals for users’ perceptions of social reality are unclear, previous research on the effects of
disproportionate representations of gender and race (Behm-Morawitz & Mastro, 2009; Comstock &
Cobbey, 1979; Mastro & Behm-Morawitz, 2005; Mastro et al., 2007; Rivadeneyra, Ward, & Gordon,
2007) suggest concern about the way that the large and dedicated MMO audience may view gender and
race in society.
While the present study focused specifically on the representation of gender and race in MMOs,
there is a need for more research on the content of MMOs and related environments. Subsequent studies
could examine other popular virtual worlds, as applications like Second Life and other social
environments may contain demographic representations inconsistent with those of the MMO games

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 12

investigated here. Other content variables that have been widely investigated in console-based games,
such as prevalence of sexualized portrayals, should also be examined in MMOs and other virtual worlds.
Future research should also examine other indicators of race aside from character appearance, including
the environment of the MMO or the dialogue used between characters. Finally, future investigations
could supplement the present study with a content analysis of game content generated in even longer
MMO game sessions with more developed MMO characters. Since more dedicated users of MMOs
interact more frequently with other user-controlled avatar characters (Ducheneaut et al., 2006), future
investigations could expand upon the foundation set by the current study to examine whether the content
trends observed here continue to present themselves after scores of hours of game play.
Until then, however, this study provides useful baseline data describing the way gender and race
tend to be portrayed in several popular MMOs. Unfortunately, though, it seems that while MMOs are
something new compared to the video games that came before them, their representation of gender and
race may not be new at all.

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
http://jvwresearch.org Gender and Race in MMO Games 13

References

Bayer, R. (2010, May 13). DDO jumps to third-most-popular spot in MMO survey. Massively. Retrieved
from http://massively.joystiq.com/2010/05/13/ddo-jumps-to-third-most-popular-spot-in-mmo-
survey/.
Beasley, B. & Standley, T. C. (2002). Shirts vs. skins: Clothing as an indicator of gender role
stereotyping in video games. Mass Communication & Society, 5(3), 279-293.
doi:10.1207/S15327825MCS0503_3.
Behm-Morawitz, E., & Mastro, D. (2009). The effects of the sexualization of female video game
characters on gender stereotyping and female self-concept. Sex Roles, 61, 808-823.
doi:10.1007/s11199-009-9683-8.
Bessière, K., Seay, A. F., & Kiesler, S. (2007). The ideal elf: Identity exploration in World of Warcraft.
Cyberpsychology, Behavior, and Social Networking, 10(4), 530-535. doi:10.1089/cpb.2007.9994.
Comstock, G., & Cobbey, R. (1979). Television and the children of ethnic minorities. Journal of
Communication, 29(1), 104-15. doi:10.1111/j.1460-2466.1979.tb0168 8.x.
Dietz, T. L., (1998). An examination of violence and gender role portrayals in video games: Implications
for gender socialization and aggressive behavior. Sex Roles, 38(5/6), 425-442.
doi:10.1023/A:1018709905920.
Downs, E. & Smith, S. L. (2010). Keeping abreast of hypersexuality: A video game character content
analysis. Sex Roles, 62, 721-733. doi:10.1007/s11199-009-9637-1.
Ducheneaut, N., Wen, M.-H., Yee, N., & Wadley, G. (2009). Body and mind: A study of avatar
personalization in three virtual worlds. Proceedings of the 27th annual International Conference
on Human Factors in Computing Systems, Boston, MA. doi:10.1145/1518701.1518877.
Ducheneaut, N., Yee, N., Nickell, E., & Moore, R. J. (2006). “Alone together?” Exploring the social
dynamics of massively multiplayer online games. Proceedings of the 24th annual International
Conference on Human Factors in Computing Systems, Montreal, CA.
doi:10.1145/1124772.1124834.
Fox, J., & Tang, W. Y. (in press). Sexism in online video games: The role of conformity to masculine
norms and social dominance orientation. Computers in Human Behavior.
doi:10.1016/j.chb.2013.07.014
Glaubke, C. R., Miller, P., Parker, M. A., & Espejo, E. (2001). Fair play? Violence, gender and race in
videogames. Oakland, CA: Children Now.
Harwood, J., & Thompson, K. M. (2002). The presence and portrayal of social groups on prime-time
television. Communication Reports, 15, 81-97. doi:10.1080/08934210 209367756.
Huh, S. & Williams, D. (2009). Dude looks like a lady: Gender swapping in an online game. In W. S.
Bainbridge (Ed.), Online worlds: Convergence of the real and the virtual (161-174). New York:
Springer. doi:10.1007/978-1-84882-825-4_13.
Hussain, Z. & Griffiths, M. D. (2008). Gender swapping and socializing in cyberspace: An exploratory
study. Cyberpsychology, Behavior, and Social Networking, 11(1), 47-53.
doi:10.1089/cpb.2007.0020.
Kolko, B., Nakamura, L., & Rodman, G. (2012). Race in cyberspace. New York: Routledge.
Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2
 http://jvwresearch.org Gender and Race in MMO Games 14

Lee, J.-E. R., & Park, S. G. (in press). Whose Second Life is this? How avatar-based racial cues shape
ethno-racial minorities’ perception of virtual worlds. Cyberpsychology, Behavior, and Social
Networking. doi:10.1089/cyber.2010.05 01.
Mastro, D., & Behm-Morawitz, E. (2005) Latino representation on primetime television. Journalism and
Mass Communication Quarterly, 82, 110-130. doi:850334381.
Mastro, D., Behm-Morawitz, E., & Ortiz, M. (2007). The cultivation of social perceptions of Latinos: A
mental models approach. Media Psychology, 9(2), 347-365. doi:10.1080/15213260701286106.
Nowak, K. L., & Rauh, C. (2006). The influence of the avatar on online perceptions of
anthropomorphism, androgyny, credibility, homophily, and attraction. Journal of Computer-
Mediated Communication, 11, 153-178. doi:10.1111/j.1083-6101.2006.tb00308.x.
Rivadeneyra, R., Ward, L. M., & Gordon, M. (2007). Distorted reflections: Media exposure and Latino
adolescents’ conceptions of self. Media Psychology, 9(2), 261-290.
doi:10.1080/15213260701285926.
Salter, A., & Blodgett, B. (2012). Hypermasculinity and dickwolves: The contentious role of women in
the new gaming public. Journal of Broadcasting and Electronic Media, 56, 401-416.
doi:10.1080/08838151.2012.705199
Soley, L. C., & Reid, L. N. (1988). Taking it off: Are models in magazine ads wearing less? Journalism
Quarterly, 65, 960-966.
Williams, D. (2006). A brief social history of video games. In Vorderer, P., & Bryant, J. (Eds.), Playing
computer games: Motives, responses, and consequences (197-212). Mahwah, NJ: Lawrence
Erlbaum Association.
Williams, D. (2006). Virtual cultivation: Online worlds, offline perceptions. Journal of Communication,
56, 69-87. doi:10.1111/j.1460-2466.2006.00004.x.
Williams, D., Consalvo, M., Caplan, S., & Yee, N. (2009). Looking for gender (LFG): Gender roles and
behaviors among online gamers. Journal of Communication, 59, 700-725. doi:10.1111/j.1460-
2466.2009.01453.x.
Williams, D., Martins, N., Consalvo, M., & Ivory, J. D. (2009). The virtual census: Representations of
gender, race, and age in video games. New Media & Society, 11(5), 815-834.
doi:10.1177/1461444809105354.
Williams, D., Yee, N., & Caplan, S. E. (2008). Who plays, how much, and why? Debunking the
stereotypical gamer profile. Journal of Computer-Mediated Communication, 13, 993-1018.
doi:10.1111/j.1083-6101.2008.00428.x.
Yee, N. (2006). The demographics, motivations, and derived experiences of users of massively
multiplayer multi-user online graphical environments. Presence, 15(3), 309-329. doi:
10.1162/pres.15.3.309.
Yee, N., Ducheneaut, N., Yao, M., & Nelson, L. (2011). Do men heal more in drag? Conflicting identity
cues between user and avatar. Proceedings of the 29th annual International Conference on Human
Factors in Computing Systems, Vancouver, CA. doi:10.1145/1978942.1979054.

Assembled 2014 / May 2014 Journal of Virtual Worlds Research Vol. 7, No. 2

View publication stats