ChapterPDF Available

Violence, Sex, Race, and Age in Popular Video Games: A Content Analysis

Authors:
Chapter

Violence, Sex, Race, and Age in Popular Video Games: A Content Analysis

Abstract

For this study we measured the prevalence of violence in top-selling video games with emphasis on a detailed description of the nature of the violence. Special attention was given to the interaction of video game violence with the demographics of the video game characters. Of primary interest was how the race, sex, and age of the characters related to roles of power, dominance, and aggression. Also reported are the demographics and portrayals of characters in general, notwithstanding the violent content of the game. This study is unique in its analysis of the content of video games and provides important information about this growing medium. Because the content of video games changes rapidly over time, follow-up investigations are necessary. (PsycINFO Database Record (c) 2010 APA, all rights reserved)(chapter)
8
VIOLENCE,
SEX, RACE,
AND AGE IN
POPULAR
VIDEO GAMES:
A
CONTENT
ANALYSIS
KAREN
E.
DILL,
DOUGLAS
A.
GENTILE, WILLIAM
A.
RICHTER,
AND
JODYC. DILL
After
dinner
in a
suburban American home,
a
13-year-old
boy
plays
a
video game alone
in his
bedroom.
On the
screen,
the
boy's hero, Duke
Nukem,
approaches
a
strip club where,
before
entering,
he
guns down
the
local
authorities. Duke
is a
young,
White
man—blond
and
tan, with
huge, rippling muscles.
On
entering
the
club,
he
sees several virtually
naked young women dancing
on
poles, moaning
and
gyrating.
He
shoots
and
kills
one of the
young women.
As her
screams
fade,
Duke
fires
his
witty
retort, "Too bad,
she was
cute."
INTRODUCTION
This
scene
is all too
typical
of
what
is
happening nightly
in
homes across
America.
The
Duke Nukem
3D
video game
is the
11th most popular selling
PC
video game
of all
time
(PC
Data
Top
Games
of
1999, 2000). Forty-nine
percent
of
children have
a
video game player
or
computer
in
their bedrooms
(Song
&
Anderson,
2001).
Griffiths
and
Hunt (1998) reported
that
98.7%
of
the
adolescents they studied were video game players.
Ninety-four
percent
of
eighth
and
ninth
graders reported playing video games, with
59%
report-
ing
playing
at
least
once
a
week
(Gentile,
Lynch, Linder,
&
Walsh.,
2004).
Finally,
Paik (2001) reported
that
87% of
younger children
and 70% of
ado-
lescents play computer games,
and
that
more boys (87%)
than
girls
(79%)
play.
Among eighth-
and
ninth-grade students, boys reported playing
an av-
erage
of 13
hours
per
week, with girls playing
an
average
of 5
hours
per
week
(Gentile
et
al., 2004). Males also reported preferring games
in the
sports
or
violence categories, whereas females
prefer
intellectual-creative
or
action-
US
fantasy
games. Finally, worldwide annual video game sales reached
$20
bil-
lion
by the
year 2000
(Cohen,
2000; Video game sales, 2001). Furthermore,
the
U.S. video game industry
has
become "the
fourth
pillar
of the
entertain-
ment business, alongside movies, music
and
television. Games sales regularly
rival
box-office
receipts,
and the top
game publishers routinely exceed
$1
billion
in
annual revenue" (Reuters, 2005a,
p.
14).
Video Games
as a
Source
of
Information
Mass
media
act as an
agent
of
socialization. Individuals learn cultural
rules
from
the
stories told
in
that
culture (Ryan
&
Wentworth, 1999). Whereas
stories
used
to be
told person-to-person within communities,
the
media
are
now
telling
the
stories (Walsh, 1997); thus, individuals learn rules
from
the
media (Ryan
&
Wentworth, 1999).
The
past
two
decades have seen
an ex-
plosion
of
information sources including
the
Internet, video games,
and an
expanding array
of
television
channels.
The
stories told
by the
media include
messages
about social roles
as
they relate
to
race, sex,
and
age. Many
of
these
stories
are
violent
and
describe,
in
terms
of
race, sex,
and
age,
who is the
powerful
aggressor
and who are the
victims.
A
person's community loses
moral
authority
as
messages come
from
outside
that
community.
The
self
becomes isolated
and the
definitions
of
acceptable
and
deviant behavior
change (Ryan
&
Wentworth, 1999). Therefore,
it is
important
to
analyze
the
content
of
media
and its
role
in
socialization, especially with regard
to
character portrayals.
Portrayal
of
Gender
in
Video Games
There
has
been little examination
of the
role
of
female characters
in
video games.
One
notable exception
is a
study
by
Tracy
Dietz
(1998),
who
analyzed
aggressive
content
and the
portrayal
of
women
in the
top-selling
Sega
and
Nintendo video games.
She
found
that
79% of the
games included
aggression,
with
21%
depicting violence toward women. Twenty-eight per-
cent
of the
games portrayed women
as sex
objects. Only
15% of the
games
portrayed
women
as
heroic characters,
and
even
those
heroic roles were mostly
sexualized
or
trivialized.
In
games with female characters,
the
females were
most likely
to be
depicted
as
victims
or as
damsels
in
distress.
Other
common
depictions included females portrayed
as
visions
of
beauty,
as
evil
or
obstacles
to the
game,
and in
devalued
or
insignificant roles.
It
is
interesting
that
Dietz (1998) found
that
the
most common depic-
tion
of
female
characters
was no
depiction
at
all: Most games simply
did not
have female characters. Similarly, Braun
and
Giroux (1989) found
that
in
arcade video games,
female
screen displays
and
synthetic voice emissions
were
practically
nonexistent.
116
DJLLETAL.
The
Media Education Foundation (Huntemann, 2000) noted
that
video
games
tend
to
send blatant messages about gender
in
Western culture.
A
female
character
is
often
the
damsel
in
distress
or
someone
who
simply
ful-
fills
male desires. Female characters
are
portrayed with distorted body irri'
ages.
These
images
are
hypersexual, with disproportionately large breasts
and
small
waists
and
hips,
and are
often physically impossible, especially given
the
athletic prowess
of the
characters (Huntemann, 2000).
For
men,
the
gender portrayal
is
equally stereotypical
and
blatant, showing
men as
sym-
bols
of
power
and
dominance. Male physical appearance
is
hypermasculine,
often
featuring chest
and arm
muscles
in
massive
and
unrealistic propor-
tions.
This
is the
male hero
fantasy:
the
muscular giant
who
wins battles
and
women
alike (Huntemann, 2000).
In
some
recent
video games, female aggressive heroines
are
emerging.
Tomb
Raider's
heroine, Lara Croft,
is
perhaps
the
most notable,
but is
accom-
panied
by
others such
as
Perfect
Dark's Joanna Dark.
Are
these characters
positive
or
negative role models? Although
the
idea
of
female
heroines
is
positive,
the
highly sexualized
way
they
are
portrayed robs them
of
dignity.
On the
surface, they
may
appear empowered,
but
they
are
actually created
to
appeal
to the
young, male game player (Huntemann, 2000).
Race
and Age in
Violent
Video Games
The
Media
Education
Foundation
(Huntemann,
2000)
reported
that
8
of
the 10
top-selling games feature White characters. Racial minorities
are
depicted
primarily
in
stereotypical
ways.
For
example,
in the
game
Kingpin,
in
which Blacks
are
portrayed
as
street thugs
and
prostitutes,
a
White,
male
character tries
to
overpower
the
Black criminal element.
Vasil
and
Wass (1993) analyzed research
on the
treatment
of age in
several
different
media sources including television
and
print media.
They
concluded that, across these studies,
the
elderly, especially women, were
underrepresented.
When
elderly characters were portrayed, they were
de-
picted negatively, cast
in
minor roles,
and
underdeveloped
as
characters.
The
apparent ages
of
video game characters have never,
to our
knowledge, been
studied
empirically.
If
characters
are
similar
in age to the
player, this
may
help
the
player
identify
with
the
character. Identification with
an
aggressor
has
been shown
to
increase aggressive behavior
in
those exposed
to
aggres-
sive
media (see
Berkowitz,
1993,
for a
discussion).
Content
Analyses:
Violence
A
handful
of
studies have examined
the
percentage
of
popular games
that
are
violent.
A
content
analysis
of
Canadian arcade games revealed that
71%
were violent (Braun
and
Giroux,
1989).
Provenzo (1991) found
that
about
85% of the
most popular games were
violent.
Funk (1993) asked ado-
VIDEO
GAME
CONTENT
ANALYSIS
117
lescents
to
report their
favorite
video games
and
found
that
50%
included
a
violent game
on
their list. Buchman
and
Funk (1996) reported
the
same
numbers
in
their study
of
fourth through eighth graders.
Similarly,
Dietz
(1998)
found
that
79% of the
popular Sega
and
Nintendo titles
she
studied were aggressive
in
nature. Twenty-one percent
included violence
specifically
directed toward women. Nearly half
the
games
included violence directed specifically
at
other characters, with
the
major-
ity
of
these characters being human
or
human-like
and
including graphic
violence.
Video Game
Violence
Linked
to
Aggression
Anderson
and
Bushman (2001) conducted
a
meta-analysis
on 35
dif-
ferent
studies
of
violent video games
to see if
these would reveal similar pat-
terns
in
their findings. They identified
the
following consistent
pattern:
Ex-
posure
to
violent video games increases physiological arousal, aggressive
thoughts, aggressive emotions,
and
aggressive actions
and
decreases positive,
prosocial actions. Thus, there
is
reason
to be
concerned
about
the
prevalence
of
violence
in
video games (for reviews
of the
literature,
see
Dill
&.
Dill,
1998;
Gentile
&
Anderson, 2003).
Rationale
for the
Present
Investigation
For
this study
we
measured
the
prevalence
of
violence
in
top-selling
video games with emphasis
on a
detailed description
of the
nature
of the
violence. Special
attention
was
given
to the
interaction
of
video game vio-
lence with
the
demographics
of the
video game characters.
Of
primary inter-
est
was how the
race, sex,
and age of the
characters related
to
roles
of
power,
dominance,
and
aggression. Also reported
are the
demographics
and
por-
trayals
of
characters
in
general, notwithstanding
the
violent
content
of the
game.
This
study
is
unique
in its
analysis
of the
content
of
video games
and
provides important information about this growing medium. Because
the
content
of
video games changes
rapidly
over time,
follow-up
investigations
are
necessary.
METHOD
The
Games
The
games
chosen
for
this analysis came
from
a
published list
of the 20
top-selling
PC
video games
of
1999
(PC
data
top
games
of
all
time,
1999;
PC
data
top
games, 2000).
PC
games were
chosen
because they
are
available
to a
wide
audience
and
because past
content
analyses
of
video games have stud-
118
DILLETAL.
led
other
game formats such
as
dedicated game systems
and
arcade games.
Following
is the
list
of the
games
and
their publishers:
RollerCoaster
Tycoon
(Hasbro Interactive), SimCity
3000
(Electronic Arts),
Who
Wants
To Be a
Millionaire?
(Disney),
Age
of
Empires
H: The Age
of
Kings
(Microsoft), StarCroft
(Havas Interactive),
Half-Life
(Havas Interactive), Command
&
Conquer:
Tiberian
Sun
(Westwood),
Microsoft
Flight
Simulator (Microsoft),
Frogger
(Hasbro Interactive),
Baldur's
Gate (Interplay),
Cabela's
Big
Game Hunter
2
(Activision),
Wheel
of
Fortune
(Hasbro Interactive),
Tom
Clancy's Rainbow
Six
Gold
Edition
(Red Storm), StarCra/t
Expansion:
Brood
War
(Havas Inter-
active),
Need
for
Speed
3: Hot
Pursuit
(Electronic Arts),
Monopoly
(Hasbro
Interactive),
Deer
Hunter
III
(GT
Interactive), Star Wars
Episode
I:
Phantom
Menace
(LucasArts),
Microsoft
Combat
Flight
Simulator (Microsoft),
and Tom
Clancy's
Rainbow Six:
Rogue
Spear
(Red Storm).
Each game
was
played
by a
professional male gamer
in his
early 20s.
His
game play
was
recorded
on
videocassettes, which were subsequently viewed
by
the
coders. Game play
was
standardized
in the
following ways: Wherever
applicable,
the
same game settings were used.
For
example,,
if
there
was a
choice
of
difficulty
level,
the
highest level
was
chosen.
For
each game,
the
opening segments
and
control screen were recorded
so the
coders could view
the
story line
and the
general game controls.
The
time each game
was re-
corded
was set at a
minimum
of 10
minutes
and a
maximum
of 30
minutes.
The
minimum time
was
used
on
games with straightforward
and
relatively
unvarying
story lines;
the
maximum time
was
used
for
games with
different
levels
or
story lines, with
an
attempt made
to
demonstrate
all the
major
play
modes. Because these
are
samples
of
game play,
not all
characters
and
set-
tings were
necessarily
recorded,
although
an
attempt
was
made
to
sample
each
of the
main actions
of the
game
in
question.
Three
trained
coders
(a
male
and a
female psychologist
and a
male com-
munications professor) rated each
of the 20
video games. After watching
the
complete video recording
of
each
game,
each
coder independently rated
the
game according
to the
categories described below. Next,
the
coders com-
pared ratings.
Any
discrepancies were resolved
by
viewing disputed footage
(e.g.,
to
determine
the
race
of a
character whose race
was
disputed). After
discrepancies were resolved,
a
single list
of
characteristics resulted. Interrater
reliability
was
acceptably high.
The
average reliability coefficient
for
the
scaled
items
was r =
0.81.
The
average reliability coefficient
for the
categorical items
was
kappa
=
0.98.
Demographics
of
Video Game
Characters
Sex,
Race,
and Age
Coders
judged
each
character's
gender, race (Black,
White,
Latino,
Asian, Indian,
or
other),
and age
(child, adolescent, adult,
or
senior).
VIDEO
CAME
CONTENT
ANALYSIS
119
Role
Coders categorized
the
character's role
in the
game
as
being
a
main
character (hero
or
primary
player),
target (object
of
violence
in
aggressive
games),
or
secondary
character
(other).
Life-Form
Type
Raters coded
the
life-form
type
of the
video game characters according
to the
following categories: human, humanoid/alien, animal,
cartoon,
and
robot/mechanical.
Life-form
type
is an
important variable, because where
modeling
behavior
is
concerned, media
viewers
tend
not
only
to
imitate
the
aggressive
behavior they witnessed
but to
take into account
the
personal
characteristics
of the
perpetrator
and
victim. Phillips (1986)
found
increases
in
homicide rates
after
highly
publicized
prizefights.
It is
interesting
that
the
races
of the
murderer
and
victim matched those
of the
winner
and
loser
in
the
prizefight.
Violence
that
is
perceived
as
more realistic
is
more likely
to
incite violence
in the
viewer (see Geen, 2001,
for a
review).
Humans, ani-
mals,
and
robots
are all
concrete
characters, whereas cartoons
and
aliens
are
less
concrete.
Characterisation
For
each video game character, raters selected whether
the
character
fit
any
of the
following important roles: sexualized, comrade, obstacle, needs
rescue,
positive role,
and
negative role.
The six
categories were considered
independent. Characters were rated
as
sexualized
if
they were depicted
as
scantily clad,
as sex
objects,
or
with sexualized
features
such
as
large breasts,
or
if the
dialogue surrounding
that
character
was
sexual
in
nature. Comrades
were
secondary characters
who
aided
.or
otherwise befriended
the
main char-
acter.
The
obstacle category
followed
from
Dietz's (1998) description. Dietz
(1998) also
found
that
female
video game characters were often depicted
as
damsels
in
distress.
This
study expanded
the
term
damsel
in
distress
to
include
characters
of
both
sexes with
the
term
needs
rescue.
Finally,
a
rating
of
whether characters were depicted particularly posi-
tively
or
negatively during game
play
was
included.
These
terms were con-
fined
to
anything outside
the
main action
of the
game
(a
soldier
in a war
game
would
not be
coded
as a
negative role,
but
characters
who
demon-
strated
vices aside
from
the
main action
of the
game were rated
as
negative).
Aggressive
Content
of the
Video
Games
Realism
Several
items tapping
into
this realism were constructed.
All
items
de-
scribed
below were rated
on
five-point Likert scales with
0
(not
at
all)
and 4
(extremely)
being
the
endpoints.
The
first
set of
questions
was as
follows:
120
DILLETAL.
"How aggressive
is
this video game?"
"When
violence happens,
how
violent
are
the
visual graphics?"
and
"When
violence happens,
how
violent
are the
auditory
elements?"
Highly
concrete aggressive
stimuli
elicit more aggression
than
do
less
concrete stimuli (Turner
&
Goldsmith, 1976).
The
next
set of
questions
distinguished between
the
terms
realistic
and
reality-based
for
several dimen-
sions
of the
games.
Realistic
referred
to the
quality
of the
computer graphics.
Something
was
reality-based
if "it
exists
or
could exist
in the
real world." Both
the
realistic
and
reality-based nature
of the
settings, main characters, targets,
and
weapons were assessed.
The
next questions expanded
the
idea
of
realism
for the
weapons.
In-
cluded
was the
question "How accessible
are the
weapons?"
Weapon
accessi-
bility
was
defined
as the
perceived likelihood that
a
person could attain
the
weapons
in
real
life.
With
a
checklist, weapons were placed into
the
follow-
ing
categories: real, science fiction,
fantasy,
and
hands.
More
than
one
cat-
egory
could
be
chosen
for a
given game.
Real
weapons were weapons
that
are
currently
available
in the
real world.
Hands
denoted aggression through hand-
to-hand combat.
The
science
fiction
and
fantasy
categories were used
for
categorizing
nonreal weaponry.
The
difference
between
the two is
that
fan-
tasy
weapons
do not
exist
and
cannot theoretically
be
created (e.g., magic
wands)
whereas science fiction weapons
do not
exist,
but
could theoretically
be
created
in the
future
on the
basis
of
scientific
theory (e.g., spaceships that
fire
lasers).
Next,
the
main
view
of the
game
was
coded
by
checking
one of
four
categories:
from
person's
eyes
(e.g., seeing hands shoot), shows main character's
entire body
(or
vehicle)
from
behind, shows characters (vehicle)
as
seen
from
directly
above,
or
shows characters (vehicles) interacting
from
a
distance.
The
first
category
is the
most realistic,
and the
categories decline
in
realism
from
there.
These
categories
are
also theoretically related
to the
degree that
the
player would
be
likely
to
identify
with
the
aggressor.
It was
theorized that
identification
should
be
greatest when seeing
the
action through
the
character's eyes
and
least when seeing
the
game characters interact
from
a
distance. Finally, whether
any of the
targets
of
aggression were objects
or
implicit
targets
was
rated.
Objects
were defined
as
nonliving entities,
and
implicit
targets
were
defined
as
targets that were supposed
to be
life
forms
but
could
not be
directly viewed
by the
player (e.g., ships
that
would presumably
be
piloted
by a
living being).
Categories
According
to a
modified version
of
Buchman
and
Funk's (1996) video
game
categories,
all
games were
classified
as
life-form
violence, general
en-
tertainment, sports violence,
or
nonviolent
sports.
VIDEO
GAME
CONTENT
ANALYSIS
121
RESULTS
Game
Characteristics
Of the
top-20 selling games
from
1999,
12
(60%) have violence
as a
major
theme.
The
games were classified
as
life-form
violence (50%), general
entertainment (35%), sports violence (10%),
and
nonviolent sports (5%).
The
amount
of
aggression
was
rated
on a 0-4
scale, with
4
representing
ex-
tremely
aggressive.
Sixty-five
percent
of
games have nonzero aggression scores,
and 12
(60%) have scores greater
than
the
midpoint
of the
scale, hereafter
referred
to as
high
scores. Among
the 12
games with violence
as a
major
theme,
the
mean aggression score
is 3.6 (SD =
0.47).
Half
of the top 20
games
are
presented
from
the
player's eyes
(first-
person viewpoint),
45% are
presented showing
the
characters interacting
from
a
distance,
and 5%
show characters
as
seen
from
directly above.
Within
violent games, half
are
presented
from
a
first-person
viewpoint,
and
half show
the
characters interacting
from
a
distance.
Three
of the 20
games (15%)
include profanity (all
three
are
violent games
and
constitute
25% of the
vio-
lent games).
Games
are
very
realistic,
both
in
terms
of the
portrayals
of
characters
and
settings
and in
terms
of how
realistic
the
graphics are.
When
considering
how
reality-based
the
settings are, raters
gave
scores
at or
above
the
scale
midpoint
for
100%
of the
games
(M = 3.6 on the 0-4
scale,
SD =
0.67).
When
considering
how
reality-based
the
main characters are, raters again
gave high scores
to
100%
of the
games
(M =
3.6,
SD =
0.54). Raters gave
high scores
to 85% of
games when considering
how
realistic
the
graphics
depicting
the
game setting are,
and to 57% of
games when considering
the
realism
of the
graphics depicting
the
game characters.
The
results
are
similar
for
violent video games.
Violent Game Characteristics
A
number
of
game characteristics
of the 12
violent games were rated
(again
on a 0 to 4
scale).
Two
thirds
of
violent games received scores
at or
above
the
scale midpoint
for how
violent
the
graphics
are (M =
2.2,
SD =
0.77),
and 58%
received high scores
for how
violent
the
auditory elements
are
(M =
1.9,
SD =
0.63). Although most games (83%) received high scores
for
how
reality-based
the
weapons are, only
50%
received high scores
for
how
realistic
the
weapons graphics are.
Two
thirds
of
games similarly
re-
ceived
high scores
for how
reality-based
the
targets
of
violence are, although
only
50%
received high scores
for how
realistic
the
target graphics are. Half
of
violent games include objects
as
targets,
and 42%
include implicit targets
(e.g.,
airplanes piloted
by
unseen living beings). Among violent games,
92%
J
22
DILL
ET AL.
include real weapons,
33%
include
science
fiction weapons,
and 8%
include
fantasy
weapons.
Eighty-three
percent
of
violent games received high scores
for how in-
strumental
the
aggression was. Although only
45% of
violent games received
high scores regarding their extent
of
having
a
theme
of
retributional aggres-
sion,
82% had
nonzero scores, indicating
at
least some amount
of
aggression
as
justified
retribution.
All of the
violent games received high scores
for re-
warding
aggressive behaviors. Twenty-five
percent
of
violent games com-
bine aggression
and
humor
at
least some
of the
time.
Although
only
one
game
received
a
high score
for
showing disrespect
for
life
(StarCroft),
42% of
violent games received nonzero scores. Over half (58%)
of the
games include
at
least some elements
of
gratuitous violence.
Of
those
games
that
include
gratuitous
violence,
22%
reward
the
gratuitous violence
and 33%
punish
the
gratuitous
violence.
Seventeen
percent
of
violent games
have
some aspects
of
sexualized
violence.
Characteristics
of
Game Characters
Across
the top 20
games,
the
characters
are
predominantly
White
male
adults.
Across
the
violent games,
the
characters
are
also predominantly
White
male
adults.
Main Characters
Across
the top 20
games,
the
main characters (28%)
are
predominantly
male. Only
10% of
main characters
are
female,
whereas
70% are
male
and
20%
could
be
either. Over
two
thirds
of the
main characters (68%)
are
White,
with
11%
Black
and 11%
Latino.
Of the
specifically
male main characters,
77%
are
White,
8% are
Black,
and 15% are
Latino.
Of the two
specifically
female
main characters,
one is
Black
and one is
Israeli. Ninety-five percent
of
main characters
are
portrayed
as
adults,
and 90% are
humans.
Secondary Characters
Across
the top 20
games,
a
majority
of the
secondary characters (48%)
are
male (55%). Thirty-one percent
of
secondary characters
are
female
and
14%
can be
either male
or
female
or are
mixed groups
of
males
and
females.
Over
two
thirds
of the
secondary characters (72%)
are
White,
with
10%
Black,
and 0%
Latino,
although
9%
include mixed-race groups.
Of the
spe-
cifically
male secondary characters,
78% are
White,
11% are
Black,
0% are
Latino,
6% are
American Indian,
and 5% are
other races.
Of the
specifically
female
secondary characters,
87% are
White
and 13% are
other races. Sixty-
nine
percent
of
secondary characters
are
portrayed
as
adults,
and 73% are
humans.
VIDEO
GAME
CONTENT
ANALYSIS
123
The
category
of
secondary characters tends
to
show<