Acting Like a Tough Guy: Violent-Sexist Video Games, Identification With Game Characters, Masculine Beliefs, & Empathy for Female Violence Victims

Abstract

Empathy-putting oneself in another's shoes-has been described as the "social glue" that holds society together. This study investigates how exposure to sexist video games can decrease empathy for female violence victims. We hypothesized that playing violent-sexist video games would increase endorsement of masculine beliefs, especially among participants who highly identify with dominant and aggressive male game characters. We also hypothesized that the endorsement of masculine beliefs would reduce empathy toward female violence victims. Participants (N = 154) were randomly assigned to play a violent-sexist game, a violent-only game, or a non-violent game. After gameplay, measures of identification with the game character, traditional masculine beliefs, and empathy for female violence victims were assessed. We found that participants' gender and their identification with the violent male video game character moderated the effects of the exposure to sexist-violent video games on masculine beliefs. Our results supported the prediction that playing violent-sexist video games increases masculine beliefs, which occurred for male (but not female) participants who were highly identified with the game character. Masculine beliefs, in turn, negatively predicted empathic feelings for female violence victims. Overall, our study shows who is most affected by the exposure to sexist-violent video games, and why the effects occur. (200 words).

Full text

Running head: VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
1
1"
Acting Like a Tough Guy: Violent-Sexist Video Games, Identification With Game Characters, 2"
Masculine Beliefs, & Empathy for Female Violence Victims 3"
4"
5"
6"
Alessandro Gabbiadini1, Paolo Riva1, Luca Andrighetto2, Chiara Volpato1 and Brad J. Bushman3, 4 7"
1University of Milano Bicocca, Italy 8"
2University of Genova, Italy 9"
3The Ohio State University, USA & 4VU University Amsterdam, The Netherlands 10"
11"
12"
13"
Word c o u n t : 4620 14"
Abstract word count: 201 15"
16"
Author Notes 17"
Correspondence concerning this article should be addressed to 18"
Alessandro Gabbiadini, 19"
University of Milano-Bicocca, Department of Psychology, 20"
Piazza Ateneo Nuovo, 1, 20126 – Milano, Italy. 21"
E-mail: ale.gabbiadini@gmail.com 22"
Phone number : ++39 348.0010477 23"
Fax number : ++39 02.64483716 24"
25"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
2
Abstract 26"
Empathy—putting oneself in another’s shoes—has been described as the “social glue” that 27"
holds society together. This study investigates how exposure to sexist video games can decrease 28"
empathy for female violence victims. We hypothesized that playing violent-sexist video games 29"
would increase endorsement of masculine beliefs, especially among participants who highly 30"
identify with dominant and aggressive male game characters. We also hypothesized that the 31"
endorsement of masculine beliefs would reduce empathy toward female violence victims. 32"
Participants (N=154) were randomly assigned to play a violent-sexist game, a violent-only game, or 33"
a non-violent game. After gameplay, measures of identification with the game character, traditional 34"
masculine beliefs, and empathy for female violence victims were assessed. We found that 35"
participants’ gender and their identification with the violent male video game character moderated 36"
the effects of the exposure to sexist-violent video games on masculine beliefs. Our results supported 37"
the prediction that playing violent-sexist video games increases masculine beliefs, which occurred 38"
for male (but not female) participants who were highly identified with the game character. 39"
Masculine beliefs, in turn, negatively predicted empathic feelings for female violence victims. 40"
Overall, our study shows who is most affected by the exposure to sexist-violent video games, and 41"
why the effects occur. (200 words) 42"
43"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
3
Introduction 44"
“Empathy is about standing in someone else's shoes, feeling with his or her heart, seeing 45"
with his or her eyes. Not only is empathy hard to outsource and automate, but it makes the 46"
world a better place.” 47"
— Daniel H. Pink, author 48"
Empathy is an emotional response that corresponds to the feelings of another person, such as 49"
feeling distress when seeing another person in distress. Empathy does indeed make the world “a 50"
better place” to live and it is one of the best predictors of prosocial behavior [1,2,3]. Numerous 51"
studies have shown that playing violent video games reduces feelings of empathy and makes people 52"
numb to the pain and suffering of others (for a meta-analytic review see [4]). 53"
People feel empathy for other individuals, not for objects. In some video games, such as the very 54"
popular Grand Theft Auto (GTA) games, female characters are treated as sex objects rather than as 55"
individuals worthy of respect. GTA main male characters are always depicted as hyper-masculine, 56"
dominant, and aggressive men. In contrast, the female characters are portrayed as sexual objects—57"
usually prostitutes or pole-dancers—who are peripheral to the game narrative and whose sole 58"
purpose is to entertain the main male characters [5]. For example, after paying a prostitute for sex, 59"
players can kill her and get their money back. Rather than being punished for such behaviors, 60"
players are often rewarded (e.g., through points, extra health to their character, etc.). 61"
Although sex-typed video game characters in the virtual world might affect perceptions of men and 62"
women in the real world, there is a dearth of research on this topic. In one study, male college 63"
students who saw photos of sex-typed male and female video game characters (vs. professional men 64"
and women) were more tolerant of sexual harassment against a female college student by a male 65"
professor [6,7] In another study, female college students who embodied sexualized video game 66"
characters (vs. non-sexualized game characters) were more likely to view themselves as a sexual 67"
object, which in turn increased their acceptance of rape myths [5]. In a third study, both male and 68"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
4
female participants were more aggressive after playing a violent game as male character than as a 69"
female character [8]. Yet, none of these studies assessed the role of individual differences (i.e., 70"
moderators) and/or underlying mechanisms (i.e., mediators) of the obtained effects. The present 71"
research fills these important gaps in the literature by testing a moderator variable (i.e., 72"
identification with the violent-sexist video game character and gender), and a mediator variable 73"
(i.e., masculine beliefs) of the effects of exposure to violent sexist video games on empathy for 74"
female violence victims 75"
Who is Most Affected: The Moderating Role of Identification 76"
With the Game Character 77"
Previous research has largely ignored moderators of the effects of violent-sexist video 78"
games on players. The notion of identification with a virtual avatar, as an online self-representation, 79"
has been investigated in past research [9,10,11]. In particular, some works have shown that when 80"
experiencing a virtual world, players are likely to establish a connection between themselves and 81"
their game character, and even imagine themselves to be that character [12,13,11]. For example, 82"
participants in one study [13] played either a first-person shooter war game or a racing game and 83"
then completed a measure of automatic attitudes using the Implicit Association Test [14]. The 84"
researchers found that participants who played a war game had stronger associations between 85"
military-related concepts and the self, whereas participants who played a racing game had stronger 86"
associations between racing-related concepts and the self. These results suggest an automatic shift 87"
in players’ implicit self-perceptions. Identification with violent video game characters can also 88"
influence behavior. For example, one study found that the more boys identified with violent game 89"
characters, the more aggressive they were after the game was turned off [11]. 90"
Building on these previous findings, in the present work we tested whether the identification with 91"
the game character would interact with exposure to violent sexist games in predicting a reduction in 92"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
5
empathic feelings toward female violence victims. More specifically, we expected that players who 93"
highly identified with violent sexist game character would display a greater endorsement of 94"
masculine beliefs. Furthermore, we expected that the interactive effects of exposure to violent sexist 95"
games and identification with the game character on masculine beliefs would be stronger for male 96"
(compared to female) players. Drawing from this theoretical framework, we propose that 97"
identification with the game character and participants’ gender could play a key-moderating role in 98"
the effects of violent-sexist games on empathy for female violence victims. 99"
Why the Effect Might Occur: The Mediating Role of 100"
Masculine Beliefs 101"
Previous research has largely ignored mediators of the effects of violent-sexist video games 102"
on players. One scholar notes that the video game culture assumes that the default player is male, 103"
which can lead to the maintenance of masculinity in the virtual world [15]. Another scholar 104"
proposed that adolescent and young adult males often use video game spaces to explore their 105"
masculine identity [16]. Masculinity refers to normative beliefs about how men are expected to 106"
think, feel, and behave [17]. Traditionally, men are considered to be aggressive, dominant, 107"
competitive, strong, powerful, and independent [17,18]. Theories of hegemonic masculinity assert 108"
that modern media convey myths about male dominance and female submission in order to support 109"
a patriarchal social structure [19]. In other words, media stereotypes construct a stylized view of 110 "
masculinity and femininity that influences the thoughts, feelings, and behaviors of those who 111"
consume these media. In the mediated world, men are expected to control their feelings in order to 112"
be less vulnerable and more powerful. Emotions such as fear and empathy are prohibited because 113"
“real men” are not supposed to express these feelings [20,18]. However, not all emotions are 114"
prohibited. Feelings of anger and rage are encouraged in “real men” because they are associated 115"
with high status and power. 116"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
6
The portrayal of men in the media as socially powerful and physically violent reinforces 117"
assumptions about how men and boys should act in society, as well as how they should treat women 118"
and girls [21]. Exposure to sex-typed media characters can have real world consequence. For 119"
example, one study found that television programs that depict women as sex objects increased the 120"
likelihood of sexual harrassment [22]. A meta-analytic review found that masculine beliefs were 121"
positively associated with aggression against women [23]. In general, males are more likely than 122"
females to agree with myths and beliefs supportive of violence against women, show less empathy 123"
for female violence victims, and consider violence against females to be a less serious problem (see 124"
[24]). We propose that the exposure to violent and sexist video games could reinforce masculine 125"
beliefs, hampering the ability for players to feel empathy for female violence victims. Because 126"
masculine norms are strongly reinforced in GTA video games, we propose that GTA gameplay will 127"
increase masculine beliefs. Masculine beliefs, in turn, are expected to be negatively related to 128"
empathy for female violence victims. 129"
Overview 130"
The present study investigates the short-term effects of playing violent-sexist video games 131"
on empathy for female violence victims. We consider identification with the game character as one 132"
possible moderator, expecting larger effects of the game content for participants who strongly 133"
identify with the hyper-masculine, violent, male game character in GTA. Further, we proposed 134"
participants’ gender as second possible moderator, assuming that the effects of the game content 135"
would be stronger for male participants than for female participants. As a possible mediator, we 136"
propose that violent-sexist games reduce empathy for female violence victims by increasing 137"
masculine beliefs. Combining these hypotheses, we propose a conditional process model ([25]; see 138 "
Fig 1). In this model, we predict that violent-sexist video games will increase masculine beliefs, 139"
which in turn will be negatively related to empathic feelings for female violence victims, especially 140"
among male participants who display higher levels of identification with the violent male game 141"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
7
character. To test whether these effects are specific to violent-sexist video games, we also included 142"
violent-only video games in our design, as well as nonviolent (control) video games. 143"
144"
--- Insert Figure 1 about here --- 145"
146"
Fig 1. Conditional process moderated mediation model (Model 11 in PROCESS [24]) 147"
Materials and Method 148"
Participants 149"
Participants were 154 Italian high school student volunteers (43.4% male, 15 to 20 years old, 150"
(M = 16.82, SD = 1.24). The study was reviewed and approved by University of Milano-Bicocca 151"
ethics committee (Prot. N. 0024403/12) before the study began. Both the high school and the 152"
university review board gave their written consent for the study; parents' consent was collected 153"
through an official written communication sent by the office school management before the 154"
beginning of the data collection. Parental informed consent and participant written assent rates were 155"
both 100%. 156"
Procedure 157"
Participants were told that investigators were testing cognitive abilities in order to develop a 158"
new video game that would be distributed in the near future. After providing basic demographic 159"
information (age, nationality, gender), participants were randomly assigned to play a violent-sexist 160"
video game (i.e., GTA San Andreas or GTA Vice City; both rated 18+ for violent and sexual 161"
content), a violent-only video game (i.e., Half Life 1 or Half Life 2; both rated 16+ for violent 162"
action, but no sexual content or violence toward women), or a non-violent video game (i.e., Dream 163"
Pinball 3D or Q.U.B.E. 2; both rated 10+ with no violent or sexual content). 164"
Half-Life is a first-person shooter video game where the player has the same visual perspective as 165"
the character. In both games used in this study (Half Life 1, Half Life 2), players fight in a post-166"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
8
apocalyptic future. Although there is a female co-protagonist (Alyx), she is portrayed in a non-167"
sexual manner. In contrast, all female characters in GTA are portrayed in a sexual manner. By 168"
selecting video games that use different representation of women, it is possible to differentiate 169"
between sexual violence exposure (i.e., GTA ) and exposure to violent games not involving sexism 170"
exposure (i.e. , Half Life). 171"
Dream Pinball 3D is a classic pinball simulation game featuring different tables to play on, while 172"
Q.U.B.E. 2 is a first-person puzzler in which the player has to solve an array of physics-based 173"
challenges. 174"
For all the three video game conditions, participants first watched the introductory video of the 175"
game for about one minute, in order to familiarize themselves with the game. Next, they practiced 176"
by playing a preselected scene for 5 minutes. During the practice session, participants were taught 177"
how to control their character and how to interact with the game world. When the practice session 178"
ended, participants played the game alone for 25 minutes 179"
For all the three games, participants were asked to pursue a specific goal or mission. We attempted 180"
to keep the mission as similar as possible for the violent-sexist and the violent-only games. For 181"
violent-sexist games (both GTA San Andreas and GTA Vice City), the mission was to destroy a rival 182"
criminal gang. Both missions started in a private-club, and then move through the streets of the city. 183"
During the gameplay, players were frequently exposed to female prostitutes and lap dancers; 184"
indeed, those are common elements in all the episodes of the GTA saga. For the violent-only games 185"
(both Half Life 1 and Half Life 2), players were asked to complete a mission whose goal was to 186 "
destroy a group of enemies. The game’s action in both Half life 1 and Half life 2, takes place in a 187"
suburban area of the city, and then moves to a few abandoned buildings. For the neutral control 188"
games (both Dream Pinball 3D and Q.U.B.E. 2), the mission was simply to accumulate as many 189"
points as possible. 190"
All games in all conditions were played at an intermediate level to avoid boredom (from the game 191 "
being too easy) and frustration (from the game being too difficult). 192"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
9
After gameplay, participants completed some video game manipulation checks. They were 193"
first asked to report the title of the video game they played, rated how violent, involving, and 194"
exciting they thought the game was (1=not at all to 7=extremely), and rated how sexualized the 195"
female figures in the game were (1=not at all to 7=extremely). Given the popularity of the GTA 196"
series among young people, we also measured how frequently they played the video game they 197"
were randomly assigned to play in this experiment (0=never played before to 7=every day). 198"
One moderator was participant gender. As a possible second moderator, we measured how much 199"
participants identified with their video game character using the 6-item (e.g., “When I am playing, it 200"
feels as if I am my character” and “My character is an extension of myself”; 1=completely disagree 201"
to 7=completely agree) of the Player Identification Scale [26] (Cronbach α=.92). 202"
As a possible mediator, we measured masculine beliefs. We used 12 items (e.g., “Boys should be 203"
encouraged to find a means of demonstrating physical prowess” and “It is OK for a guy to use any 204"
and all means to ‘convince’ a girl to have sex”; 1=completely disagree to 7=completely agree) from 205"
the revised Male Role Norms Inventory (MRNI-R; Cronbach α=.78) [27] Levant, Rankin, 206"
Williams, Hasan, & Smalley, 2010). 207"
The dependent variable was how much empathy participants felt toward female violence victims. 208"
When people become desensitized to violence, they become numb to the pain and suffering of 209"
violence victims [28]. Participants were shown one of two photos (randomly determined) of an 210"
adolescent girl who had been physically beaten by an adolescent boy (see S1 Appendix). 211"
Participants rated whether they felt sympathetic, moved, compassionate, tender, warm, softhearted, 212"
disregarded (reverse-coded) and indifferent (reverse-coded) [see 29] for her (1=not at all to 7=very 213"
much; Cronbach α=.83). 214"
About 2 weeks later, following the completion of the study, all participants were fully debriefed. No 215"
participants expressed suspicion about the true purpose of the study. In particular, none of the 216"
participants reported a link between video games and gender-based violence or sexism. The 217"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
10
experimenter then disclosed the purpose of the study, and discussed the potentially harmful short-218"
terms effects of violent and sexist video games on players. A group discussion followed. 219"
Results 220"
Preliminary Results 221"
Stimulus sampling. 222"
To increase the generalizability of our findings, we used two video games of each type and 223"
two photos of interpersonal violence [30]. Independent-sample t-tests found no significant 224"
differences between the two different violent-sexist games, between the two different violent-only 225"
games, or between the two nonviolent games on identification with the main video game character, 226"
masculine beliefs, or empathy for female violence victims (ps>.12). Thus, the two violent-sexist 227"
games were combined, the two violent-only games were combined, and the two nonviolent games 228"
were combined for subsequent analyses. 229"
There were no significant differences between the two photos of the adolescent boy beating the 230"
adolescent girl on how much empathy participants felt for her (ps>.10). Thus, the data from two 231"
photos were combined for subsequent analyses. 232"
Video game manipulation check items. 233"
Five items were included to assess whether the video game manipulation was successful. 234"
First, we checked the name of the game reported by each participant. All participants correctly 235"
named the video game they played. We then tested whether the violent video games were rated to 236"
be more violent than the nonviolent games. A one-way between subjects ANOVA found a 237"
significant difference in violence ratings for the video games, F(2,150)=182.16, p<.001, η2=.70. 238"
Post-hoc comparisons using Tukey’s Honestly Significant Difference (HSD) test indicated that the 239"
violent-sexist games (M=5.11, SD=1.23) and the violent-only games (M=4.09, SD=1.25) had 240"
significantly higher violence ratings than the nonviolent games (M=1.20, SD=.53), ds=4.13 and 241"
3.01, respectively. However, the violent-sexist games also had higher violence ratings than the 242"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
11
violence-only games (M=5.11, SD=1.23; d=.82). Then, we checked ratings of familiarity. A one-243"
way between-subjects ANOVA found a significant difference in frequency of play for the video 244"
games, F(2,152)=4.28, p=.015, η2=.053. Post-hoc test revealed significantly that violent-sexist 245"
games had higher frequency of play (M=2.10, SD=1.54; p=.013) than both violent-only (M=1.36, 246"
SD=1.31) and neutral games (M=1.58, SD=1.03). Thus, violence ratings and frequency of play were 247"
included as covariates in all analyses. No statistically significant differences between violent-sexist 248"
(M=4.18, SD=1.56), violent-only (M=3.94, SD=1.42), and neutral games (M=3.82, SD=1.76) were 249"
found for game involvement, F(2,151)=0.70, p>.49. Moreover, no significant differences between 250"
violent-sexist (M=3.71, SD=1.55), violent-only games (M=3.65, SD=1.68), and neutral games 251"
(M=3.18, SD=1.58) were found of game excitement, F(2,151)=1.67, p>.19. Finally, we tested 252"
whether female characters were rated as more sexualized in the violent-sexist video games than in 253"
the violent-only games. As expected, female characters in the GTA games were rated to be more 254"
sexualized (M=5.88, SD=1.43) than female characters in the Half Life games (M=2.16, SD=1.31), 255"
t(101)= 13.67, p<.001, d=2.72. Taken together, these results suggest that the video game 256"
manipulation was successful. 257"
Primary Results 258"
Since playing violent video games that are not necessarily sexist have been already shown to 259"
reduce feelings of empathy, we contrast coded the type of video game as follow: +1=violent-sexist 260"
games, 0=violent-only games, -1=non-violent games. Data analysis revealed that the type of video 261"
game was positively associated with masculine beliefs (r=.203, p=.011). Participant gender (1=male 262"
vs. 0=female) was associated with both masculine beliefs (r=.507, p<.001) and identification with 263"
the game character (r=.241, p=.003). Mean comparison revealed that participant gender impacted 264"
the masculine beliefs such that male participants showed greater endorsement of masculine beliefs 265"
(M=3.23, SD=0.82) than female did (M=2.44, SD=0.52; t(152)=7.25; p<.001; d=1.15). Participant 266"
gender also influenced the identification with the game character such that boys identified more 267"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
12
with the game character (M=4.63, SD=1.20) than girls (M=3.96, SD=1.45; t(152)=3.05; p=.003; 268"
d=0.50). Further, the level of identification with the game character was positively associated with 269"
masculine beliefs (r=.197, p=.014). Finally, masculine beliefs were negatively associated with 270"
empathy for female violence victims (r=-.348, p<.001; see Table 1). 271"
272"
273"
Table 1. Bivariate correlations 274"
1
2
3
4
5
6
7
8
1. Type of video game
1
2. Masculine beliefs
.203*
1
3. Participants' gender
.022
.507**
1
4. Identification with the
game character
.073
.197*
.241**
1
5. Empathy for female
violence victims
.000
-.348**
-.137
-.005
1
6. Violence ratings
.809**
.038
-.016
.082
.094
1
7. Frequency of play
.155
.236**
.335**
.117
-.018
.078
1
8. Participants' age
-.521**
-.085
.100
.046
-.003
-.249**
-.020
1
275"
Note. Type of video game was coded +1=violent-sexist games, 0=violent-only games, and -1=non-violent 276"
games. *p < .05. **p < .01. ***p < .001. 277"
278"
We conducted a series of one-way between-subjects ANOVAs to test the effects of type of 279"
video game played on three dependent variables: (1) identification with the game character, (2) 280"
masculine beliefs, and (3) empathy for female violence victims. The first ANOVA found a 281"
significant effect for type of video game played on identification with the game character, 282"
F(2,152)=6.06, p=.003, η2=.073. Post-hoc comparisons using Tukey’s Honestly Significant 283"
Difference (HSD) test indicated that participants who played a violent game identified with the 284"
character more than did participants who played a neutral game (d=0.62, p=.003) or a violent-sexist 285"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
13
game (d=0.57, p=.04). No differences were found between the neutral and the violent-sexist games 286"
(p=.67). 287"
The second ANOVA found a significant effect for type of video game played on masculine 288"
beliefs F(2,152)=3.33, p=.038, η2=.004. Post-hoc tests indicated that participants who played a 289"
violent-sexist game reported higher masculine beliefs than did participants who played a neutral 290"
game (d=0.50, p=.03). No statistically differences were found between violent-sexist games and 291"
violent-only games (p=.24) or between violent-only and neutral games (p=.57) on masculine 292"
beliefs. 293"
The third ANOVA found no significant effect for type of video game played on empathy 294"
for female violence victims (p=.31). Descriptive statistics by experimental conditions for each 295"
dependent variable are reported in Table 2. 296"
297"
Table 2. Means and standard deviations (in parenthesis) by experimental conditions for each of the 298"
considered dependent variables. 299"
Neutral game
(males n=21; females n=29)
Violent-only game
(males n=25; females n=30)
Violent-sexist game
(males n=22; females n=26)
Identification with
the game character
3.89 (1.57)b
4.76 (1.21)a
4.12 (1.18)b
Masculine beliefs
2.62 (0.72)b
2.77 (0.73)b
3.02 (0.86)a
Empathy for
female violence
victims
5.03 (0.95)a
5.29 (0.92)a
5.02 (1.16)a
Note. Different letters indicate means statistically differences between experimental conditions. 300"
"301"
To test our predictions, we then conducted a conditional process model by using the 302"
PROCESS macro Model 11 for SPSS with 1000 bootstrapping samples [24]. In this model, the type 303"
of video game played was entered as predictor, the identification with the game character as a 304"
moderator, masculine beliefs as the mediator, and empathy toward female violence victims as the 305"
outcome variable. We predicted that participants’ gender would moderate the effects of the 306"
identification with the game character on the relationship between the type of video game and 307"

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
"
"
14
masculinity beliefs. (see Figure 1). Participant age, video game violence rating, and frequency of 308"
video game play were also included as covariates. 309"
Considering the moderated path from the type of video game to masculine beliefs, analyses 310"
revealed that the main effects of type of video game played, identification with the game character, 311"
participant gender, participant age, and frequency of gameplay were not significant (bs<.56, 312"
ts(140)<1.47, ps>.14), whereas the main effect of violence rating was significant (b=-.10, SE=.047, 313"
t(140)=-2.07, p=.039). The two-way interaction between type of video game played and participant 314"
gender was significant (b=-1.05, SE=.45, t(140)=-2.29, p=.023), whereas the interactions between 315"
type of video game played and identification with the game character (b=.02, SE=.065, t(140)= .32, 316"
p=.74) and between participant gender and identification with the game character (b=-.03, SE=.083, 317"
t(140)=.36, p=.71) were both nonsignificant. Crucially, all these effects were qualified by our 318"
predicted 3-way interaction between type of video game played, participant gender, and 319"
identification with the game character on masculine beliefs (b=.26, SE=.10, t(140)=2.55, p=.011; 320"
see Table 3). 321"

Running head: VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
15
Table 3. Regressions of type of video game (neutral, only-violent, sexist-violent) on empathy for female violence victims when masculine beliefs is the mediator 322#
and participants’ gender and identification with the game character are the moderators 323#
b
SE
p
Masculine beliefs
Type of video game
.229
.273
.40
Participants’ gender
.561
.38
.14
Identification with the game character
.063
.049
.20
Age
-.013
.052
.80
Violence rating
-.099
.047
.04
Frequency of play
.026
.043
.54
Type of video game X Participants’ gender
-1.05
.457
.02*
Type of video game X Identification with the game character
.021
.065
.74
Participants’ gender X Identification with the game character
.030
.083
.71
Type of video game X Participants’ gender X Identification with the game character
.265
.103
.01*
Empathy for female violence victims
Age
-.039
.079
.61
Violence ratings
.099
.074
.18
Frequency of play
.052
.061
.38
Masculine beliefs
-.465***
.109
.0001***
Note. b = unstandardized beta weight; SE = standard error; *p < .05. **p < .01. ***p < .001. 324#

Running head: VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
16
For male participants, simple slope analyses showed a significant positive relationship between 325$
identification with the game character and masculine beliefs for males who played with a violent-326$
sexist game (b=.32, SE=.12, t(62)=2.65, p=.009), but not for males who played violent-only game 327$
(b=.095, SE=.078, t(62)=1.21, p=.23) or for males who played a nonviolent game (b=-.13, SE=.97, 328$
t(62)=-1.33, p=.18; see Fig 2). For female participants, there was no significant relationship 329$
between identification with the game character and masculine beliefs in any of the three video game 330$
conditions (bs<.082, ts(76)<1.81, ps>.075). 331$
332$
--- Insert Figure 2 about here --- 333$
334$
Fig 2. Interactive effect of video game content and identification with the game character on 335$
masculine beliefs for male participants. The difference between violent-sexist game players 336$
and violent-only and nonviolent game players is significant for values of identification with 337$
video game characters greater than 4.2931. 338$
339$
Then, considering the path from masculine beliefs (i.e., the mediator variable) to empathy 340$
for female violence victims (i.e., the outcome variable), analyses revealed that masculine beliefs 341$
negatively impacted the dependent variable (b=-.45, SE=.11, t(146)=-4.17, p<.001). 342$
As first support of our moderated mediational hypothesis, we found a significant indirect 343$
effect of the type of video game on empathy for female violence victims via increased masculine 344$
beliefs for males [-0.1846 (95% CI=-0.3984 to -0.0410)] but not for females [-0.0406 (95% CI=-345$
0.1318 to 0.0097)] who played with a violent-sexist game. No effect was found for males [-0.0486 346$
(95% CI=-0.1733 to 0.0363)] or females [-0.0294 (95% CI=-0.0799 to 0.0041)] in the violent-only 347$
game condition. Similarly, no indirect effect of type of video game on empathy was found for males 348$
[-0.0874 (95% CI=-0.0134 to 0.3363)] or females [-0.0181 (95% CI=-0.0783 to 0.286)] in the 349$
nonviolent game condition. Further, conditional indirect effects of type of video game on empathy 350$

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
$
$
17
for female violence victims were significant only for males in the violent-sexist game condition at 351$
values of identification with the game character greater than 4.2583 [-0.1920 (95% CI=-0.4668 to -352$
0.0492] (see Fig 2). These results suggest that violent-sexist games decreased empathy for female 353$
violence victims for boys who strongly identified with the violent game character, and did so by 354$
increasing masculine beliefs. 355$
Discussion 356$
The present research is in line with previous studies showing that violent video games can 357$
desensitize individuals to real-life violence [4,31], including violence against women [6]. More 358$
important, it moves beyond the question of whether violent games are harmful per se to address the 359$
important questions of who is most likely to be harmed by violent-sexist video games, and through 360$
what mechanism does the harm occur. 361$
A possible answer to the “who” question is players that identify with the violent-sexist game 362$
character. Results support the prediction that playing violent-sexist video games increases 363$
masculine beliefs and decreases empathy for female violence victims, especially for boys and 364$
young men who highly identified with the male game character. Previous research has shown that 365$
video games are especially likely to increase aggression among players who identify with violent 366$
game characters [11], and that a reduced empathy is one of the major predictor for aggression 367$
against women [32]. Exposure to media violence is one of the many factors that can influence 368$
empathy levels [33]. Violent video games, in particular, might reduce empathy levels because 369$
players are linked to a violent character. If the video game is a first person shooter, players have the 370 $
same visual perspective as the killer. If the video game is third person, players control the actions of 371$
the violent character from a more distant visual perspective. Because they are forced to adopt the 372$
visual perspective of the perpetrator, it is difficult for players to put themselves in the shoes of the 373$
victim. In general, we argue that it is important to take individual differences into account when 374$
considering violent video game effects (see also [34]). 375$

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
$
$
18
A possible answer to the “why” question (i.e., through which mechanism) is masculine beliefs. 376$
Results showed that masculine beliefs were negatively related to empathic feelings for female 377$
violence victims. To our knowledge, the present research is the first to elucidate the underlying 378$
mechanism that links violent video games playing to desensitization of violence against women. 379$
Speaking to the specificity of our effects, it is noteworthy that type of video game played 380$
directly affected identification with the game character and masculine beliefs, whereas it did not 381$
directly affect empathy for female violence victims. However, we found an indirect conditional 382$
effect of violent-sexist games (vs. only-violent vs. neutral) on empathy, which consistently emerged 383$
through the mediation of masculine beliefs and the moderation of identification with the game 384$
character. Accordingly, the effects were statistically significant only for highly identified male 385 $
participants who played the GTA games, which are both violent and sexist. We found no significant 386$
effects for violent-only or nonviolent video games on masculine beliefs or empathic feelings. As in 387$
previous research (e.g., [8]), gender differences in identification with the game character emerged 388$
(i.e., males identified more with the male game character than females did). In addition, males had 389$
higher masculine beliefs than females did. This finding can be interpreted in light of the theoretical 390$
framework of identification with a virtual character. Video game identification has been considered 391$
as an altered experience of the self, in which players may come to perceive themselves to actually 392$
be their game character, thus assuming the character’s point of view [13]. This process has been 393$
described also as “an imaginative process” involving cognitive, emotional, and motivational 394$
dimensions ([35], p. 250). The player shares the character’s perspective (cognitive), feelings 395$
(emotional), and goals (motivational) [35,36,37]. Furthermore, identification with a virtual 396$
character has been found to be greater in video games with an articulate plot, in which the assigned 397$
role fosters a sense of ''vicarious self-perception'' [38]. This is the case with GTA, in which players 398 $
assume the role of a man who is aggressive, misogynistic, cruel, and greedy. Since that, it is not 399$
surprising that we found that male players who identified with the main video game character ended 400$

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
$
$
19
up adopting his point of view more easily than female players, as indicated by an increase in 401$
masculine beliefs and a decrease in feelings of empathy for female violence victims. 402$
This investigation of virtual representations of males and females in video games is 403$
extremely relevant, because video games have distinct features compared to other forms of media 404$
[5] and different effects on males and females. Unlike images in traditional media, game characters 405$
are designed to respond to a user’s actions [39], which can promote a powerful experience that goes 406$
beyond passive media consumption. Often these interactions mirror communication in the physical 407$
world, and users often react to virtual situations in natural and social ways [39,5]. 408$
Limitation and Future research 409$
Our study, like all studies, has limitations. Few main limitations stand out. First, we used a 410$
self-report measure of empathy. Future research should consider other measures of empathy that are 411$
less subject to demand characteristics, such as physiological measures (e.g., heart rate, skin 412$
conductance). 413$
Second, we examined only two moderators (i.e., participant gender and identification with the game 414$
character) and only one mediator (i.e., masculine beliefs). Future research should examine other 415$
possible moderators (e.g., trait aggressiveness, social dominance orientation) and mediators (e.g., 416$
dehumanization, women objectification; [see 22]). 417$
Third, our study was based on a short exposure to violent-sexist video games (i.e., about 25 418$
minutes). Although it is impressive that we were able to obtain significant effects after such a brief 419$
exposure, we do not know what the consequences would be for longer exposures. If the effects 420$
occur after only 25 minutes of play in a laboratory experiment, they are probably magnified after 421$
longer periods of play outside the lab. Indeed, individuals usually play video games for much 422$
longer periods of time (from 8 hours to 13 hours per week; [40, 41]). Previous experimental 423$
research has shown that the effects of violent video games can accumulate and get larger over time, 424$
at least over a three-day period [41]. 425$

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
$
$
20
Fourth, we did not measure sexist thoughts after a certain amount of time, so we cannot ascertain if 426$
and how long the observed effects last. However, previous experimental research has shown that 427$
the effects of violent video games can last at least 24 hours after gameplay if players ruminate about 428$
the content of the game [43]. 429$
Fifth, GTA is a well-known game, and the simple act of playing it could have primed masculine and 430$
sexist thoughts regardless of the actual gameplay. Indeed, media audiences often make their 431$
assessment of characters and narratives using existing schemes rather than actual on-screen 432$
action/content [44]. Thus, future studies should test whether similar effects are obtained for less 433$
well-known sexist-violent videogames than GTA. Furthermore, future studies should test whether 434$
the empathy reduction linked with violent-sexist video games is specific to female violence victims 435 $
or whether it extends to male violence victims. Future research should also examine video games 436$
with female characters that are not depicted in a sexualized manner. 437$
Sixth, we did not actually test whether feelings of empathy mediate aggression against women. As a 438 $
first step, this study focused on empathy for female violence victims. Future studies should explore 439$
whether violent-sexist video games also increase aggression against women. 440$
Conclusion 441$
One of the best predictors of aggression against girls and women is lack of empathy [32]. 442$
The present research shows that violent-sexist video games such as GTA reduce empathy for 443$
female violence victims, at least in the short-term. This reduction in empathy partly occurs because 444$
video games such as GTA increase masculine beliefs, such as beliefs that “real men” are tough, 445$
dominant, and aggressive. Our effects were especially pronounced among male participants who 446$
strongly identified with the misogynistic game characters. Daniel Pink was correct in noting that 447$
empathy makes the world a better place. Unfortunately, it appears that GTA might make the world 448$
a worse place for females. 449$
450$

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
$
$
21
Acknowledgments 451$
The authors declare that they have no competing interests or any business relationships with 452$
the manufacturers of the video games and products cited in the text. 453$
This work is supported by PRIN (2012)-20123X2PXT_003 grant to the first and the third authors. 454$
We want to thank “G. Maironi da Ponte” high school institute, especially Prof. Mazzotti and Mr. 455$
Maffeis, for supporting us in the data collection process. 456$
457$

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
$
$
22
References 458$
1. Batson CD. Prosocial behavior and altruism. In: Lindzey G, Gilbert D, Fiske ST, editors. 459$
The Handbook of Social Psychology, Oxford University Press; 1998. pp. 282–316. 460$
2. Eisenberg N, Miller PA. The relation of empathy to prosocial and related behaviors. Psychol 461$
Bull. 1987;101: 91-119. doi: 10.1037/0033-2909.101.1.91 462$
3. Hoffman ML. Empathy and prosocial behavior. In: Lewis M, Haviland-Jones JM, Barrett 463$
LF, editors. Handbook of Emotions, Third Edition. Guilford Publications; 2008. pp. 440–464$
455. 465$
4. Anderson CA, Shibuya A, Ihori N, Swing EL, Bushman BJ, Sakamoto A, Rothstein HR, 466$
Saleem M, Barlett CP. Violent video game effects on aggression, empathy, and prosocial 467$
behavior in Eastern and Western countries: A meta-analytic review. Psychol Bull. 2010;136: 468 $
151-173. doi: 10.1037/a0018251 469$
5. Fox J, Ralston RA, Cooper CK, Jones KA. Sexualized avatars lead to women's self-470$
objectification and acceptance of rape myths. Psychol Women Quart. 2014. In press; doi: 471$
10.1177/0361684314553578. 472$
6. Dill K, Brown B, Collins M. Effects of exposure to sex-stereotyped video game characters 473$
on tolerance of sexual harassment. J Exp Soc Psychol. 2008;44: 1402-1408. 474$
doi:10.1016/j.jesp.2008.06.002 475$
7. Dill K, Thill K. Video game characters and the socialization of gender roles: Young 476$
people’s perceptions mirror sexist media depictions. Sex Roles. 2007;57: 851-864. doi: 477$
10.1007/s11199-007-9278-1 478$
8. Yang GS, Huesmann LR, Bushman BJ. Effects of playing a violent video game as male 479$
versus female avatar on subsequent aggression in male and female players. Aggressive 480$
Behav. 2014;40: 537-541. doi: 10.1002/ab.21551 481$
9. Bailenson JN, Beall AC. Transformed social interaction: Exploring the digital plasticity of 482$
avatars. In: Schroeder R., Axelsson A, editors. Avatars at work and play: Collaboration and 483$
interaction in shared virtual environments. London: Springer-Verlag; 2006. pp. 1-16. 484$
10. Dunn RA, Guadagno RE. My avatar and me: Gender and personality predictors of avatar-485$
self discrepancy. Comput Hum Behav. 2012;28: 97-106. doi: 10.1016/j.chb.2011.08.015 486$
11. Konijn EA, Nije Bijvank M, Bushman BJ. (2007). I wish I were a warrior: The role of 487$
wishful identification in effects of violent video games on aggression in adolescent 488$
boys. Dev Psychol. 2007;43: 1038-1044. doi: 10.1037/0012-1649.43.4.1038 489$
12. Gabbiadini A, Mari S, Volpato C, Monaci MG. Identification processes in online groups: 490$
Identity motives in the virtual realm of MMORPGs. J Media Psychol. 2014;26: 141-152. 491$
doi: 10.1027/1864-1105/a00119 492$

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
$
$
23
13. Klimmt C, Hefner D, Vorderer, P., Roth, C., & Blake, C. (2010). Identification with video 493$
game characters as automatic shift of self-perceptions. Media Psychol. 2010;13: 323-338. 494$
doi: 10.1080/15213269.2010.524911 495$
14. Greenwald AG, McGhee DE, Schwartz JLK. Measuring individ-ual differences in implicit 496$
social cognition: The Implicit Association Test. J Pers Soc Psychol. 1998;74: 1464–1480. 497$
15. Gray KL. (2012). Deviant bodies, stigmatized identities, and racist acts: Examining the 498$
experiences of African–American gamers in Xbox Live. New Rev Hypermed Multimed. 499 $
2012;18: 261–276. 500$
16. Jansz J. The emotional appeal of violent video games for adolescent males. Commun Theor. 501$
2005;15: 219–241. 502$
17. Mahalik JR, Locke BD, Ludlow LH, Diemer MA, Scott RPJ., Gottfried M, et al. 503$
Development of the Conformity to Masculine Norms Inventory. Psychol Men Masc. 504$
2003;4: 3-25. doi: 10.1037/1524-9220.4.1.3 505$
18. Volpato C. Psicosociologia del maschilismo. Roma-Bari: Laterza; 2013. [Psychology of 506$
masculinity]. 507$
19. Connell R. Gender and Power: Society, the Person, and Sexual Politics. Stanford, Calif: 508$
Stanford University Press; 1987. 509$
20. Levant R.F., Majors R.G., Kelley M. Masculinity ideology among young African American 510$
and European American women and men in different regions of the United States. Cult Div 511 $
Mental Health. 1998;4: 227–236. doi:10.1037/1099-9809.4.3.227 512$
21. Götz M, Lemish D, Aidman A, Moon H. Media and the make-believe worlds of children: 513$
When Harry Potter meets Pokémon in Disneyland. Mahwah, NJ: Lawrence Erlbaum; 2005. 514$
22. Galdi S, Maass A, Cadinu M. Objectifying media: their effect on gender role norms and 515$
sexual harassment of women. Psychol Women Quart. 2014;38: 398-413. doi: 516$
0361684313515185. 517$
23. Murnen SK, Wright C, Kaluzny G. If boys will be boys, then girls will be victims? A meta-518$
analytic review of the research that relates masculine ideology to sexual aggression. Sex 519$
Roles. 2002;17: 359-375. doi: 10.1023/A:1020488928736 520$
24. Flood M, Pease B. Factors influencing attitudes to violence against women. Trauma, 521$
Violence Abuse. 2009; 10(2):125-142. doi: 10.1177/1524838009334131 522$
25. Hayes AF. Introduction to mediation, moderation, and conditional process analysis. New 523$
York: The Guilford Press; 2013. 524$

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
$
$
24
26. Van Looy J, Courtois C, De Vocht M, De Marez L. Player identification in online games: 525$
validation of a scale for measuring identification in MMOGs. Media Psychology. 526$
2012;15:197–221. 527$
27. Levant RF, Rankin TJ, Williams C, Hasan NT, Smalley KB. Evaluation of the factor 528$
structure and construct validity of the Male Role Norms Inventory-Revised (MRNI-R). 529$
Psychol Men Masc. 2010;11: 25-37. doi: 10.1037/a0017637 530$
28. Bushman BJ. Anderson CA. Comfortably numb: Desensitizing effects of violent media on 531$
helping others. Psychol Sc. 2009;20: 273-277. doi: 10.1111/j.1467-9280.2009.02287.x 532$
29. Batson CD, Fultz J, Schoenrade PA. Distress and empathy: Two qualitatively distinct 533$
vicarious emotions with different motivational consequences. J Pers. 1987;55: 19–39. doi: 534$
10.1111/j.1467-6494.1987.tb00426.x. 535$
30. Wells GL, Windschitl PD. Stimulus sampling and social psychological experimentation. 536$
Pers Soc Psychol Bull. 1999;25: 1115–1125. doi:10.1177/01461672992512005 537$
31. Carnagey NL, Anderson CA, Bushman BJ. The effect of video game violence on 538$
physiological desensitization to real-life violence. J Exp Soc Psychol. 2007;43: 489–496. 539$
doi:10.1016/j.jesp.2006.05.003 540$
32. Sanday PR. The socio-cultural context of rape: A cross-cultural Study. J Soc Issues. 541$
1981;37: 5-27. doi: 10.1111/j.1540-4560.1981.tb01068.x 542$
33. Krahé B., Möller I. Longitudinal effects of media violence on aggression and empathy 543$
among German adolescents. J App Dev Psychol. 2010;31: 401-409. doi: 544$
10.1016/j.appdev.2010.07.003 545$
34. Gabbiadini A, Riva P, Andrighetto L, Volpato C, Bushman BJ. Interactive effect of moral 546$
disengagement and violent video games on self-control, cheating and aggression. Soc 547$
Psychol Personal Sci. 2013;5: 451-458. doi: 10.1177/1948550613509286 548$
35. Cohen J. Defining Identification: A Theoretical Look at the Identification of Audiences with 549$
Media Characters. Mass Comm & Soc. 2001;4: 245–264. 550$
doi:10.1207/S15327825MCS0403_01 551$
36. Lin JH. Identification matters: A moderated mediation model of media interactivity, 552$
character identification, and video game violence on aggression. J Commun. 2013;63: 682-553$
702. doi:10.1111/jcom.12044 554$
37. Peng W. The mediational role of identification in the relationship between experience mode 555$
and self-efficacy: Enactive role-playing versus passive observation. Cyberpsychol Behav. 556$
2008;11: 649-652. doi: 10.1089/cpb.2007.0229 557$
38. Goldstein NJ, Cialdini RB. The spyglass self: A model of vicarious self-perception. J Pers 558$
Soc Psychol. 2007;92: 402-417. 559$

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
$
$
25
39. Blascovich J, Loomis J, Beall A, Swinth K, Hoyt C, Bailenson JN. Immersive virtual 560$
environment technology as a methodological tool for social psychology. Psychol Inq. 561$
2002;13: 103-124. 562$
40. ESRB. How much do you know about video games? ESRB, 2014; Available: 563$
http://www.esrb.org/about/video-game-industry-statistics.jsp 564$
41. Statista. Daily time spent playing video games per capita in the United States in 2008, 2013 565$
and 2018 (in minutes). Statista. 2015: Available: 566$
http://www.statista.com/statistics/186960/time-spent-with-videogames-in-the-us-since-567$
2002/ 568$
42. Hasan Y, Bègue L, Scharkow M, Bushman BJ. The more you play, the more aggressive you 569$
become: A long-term experimental study of cumulative violent video game effects on 570$
hostile expectations and aggressive behavior. J Exp Soc Psychol. 2012;49: 224-227. DOI: 571$
10.1016/j.jesp.2012.10.016 572$
43. Bushman, BJ. Gibson B. Violent video games cause an increase in aggression long after the 573$
game has been turned off. Soc Psychol Pers Sc. 2011; 2: 29-32. doi: 574$
10.1177/1948550610379506 575$
44. Raney AA. Expanding disposition theory: Reconsidering character liking, moral 576$
evaluations, and enjoyment. Comm Theory, 2004;14: 348-369. 577$
578$
579$
580$
581$
582$

VIOLENT SEXIST VIDEO GAMES, MASCULINE BELIEFS, & EMPATHY
$
$
26
Supporting Information files 583$
S1 Fig 1 caption. Figure 1 (for illustrative purposes only) used to measure empathy toward a 584$
female victim of violence. Participants indicated how much pain they thought the girl was feeling, 585$
and how much they thought she was suffering (1=not at all to 7=very much; Cronbach α=.79). 586$
S1 Fig 2 caption. Figure 2 (for illustrative purposes only) used to measure empathy toward a 587$
female victim of violence. 588$
589$
590$
591$