Prosocial, Antisocial, and Other Effects of - APA Home Page

24 Pages·2011·1.43 MB·English

Anderson, C. A., Gentile, D. A., & Dill, K. E. (2012). Prosocial, Antisocial, and Other
Effects of Recreational Video Games. Chapter in D. G. Singer, & J. L. Singer (Eds),
nd
Handbook of Children and the Media, 2 Edition, (pp. 249-272). Thousand Oaks, CA:
Sage. 13
chapter
Prosocial, Antisocial, and Other
Effects of Recreational Video Games
Craig a. anderson
Center for the Study of Violence,
Department of Psychology, Iowa State University
douglas a. gentile
Center for the Study of Violence,
Department of Psychology, Iowa State University
Karen e. dill
Media Psychology, School of Psychology,
Fielding Graduate University
Prosocial, Antisocial, that an electronic device can be operated. The
video game industry’s revenues surpassed the
and Other Effects of
movie industry’s several years ago and sur-
Recreational Video Games passed the music industry’s in 2008. A recent,
nationally representative sample of U.S. teens
Video games are immensely popular around found that 99% of boys and 94% of girls
the world. They are played on computers, played video games (Lenhart et al., 2008).
handheld devices, cell phones, and game con- The amount of time spent playing games
soles. They are played at home, at arcades, at has increased over time (Escobar-Chaves &
school, in automobiles, and virtually anywhere Anderson, 2008; Gentile & Anderson, 2003).
Author Note: Correspondence concerning this manuscript should be addressed to Craig A. Anderson, Iowa State
University, Department of Psychology, W112 Lagomarcino Hall, Ames, IA 50011-3180. Phone: [phone]; Fax:
515–294–6424; E-mail: [email protected]
249
250 • The Popular Media as Educators and Socializers of Growing Children
Many children and adolescents play more The General Aggression Model and its off-
than 20 hours each week; 40 hours of gaming shoot, the General Learning Model, describe
per week is not uncommon among males the basic learning processes and effects involved
(e.g., Bailey, West, & Anderson, 2010). in both short-term and long-term effects of
Much research has examined potential playing various types of games. The Five
positive and negative effects of playing vari- Dimensions of Video Game Effects perspec-
ous types of video games. Most has focused tive describes different aspects of video
on the deleterious effects of violent games games and video game play that influence the
(e.g., Anderson et al., 2010). Other research specific effects likely to occur. The Risk and
has focused on educational games (e.g., Resilience perspective reminds us that the
Murphy et al., 2002). Still other work has effects of video game play—prosocial, anti-
found that total time playing video games social, and other—take place within a com-
is negatively associated with school per- plex set of social and biological factors, each
formance (e.g., Anderson & Dill, 2000; of which contribute to development of the
Anderson, Gentile, & Buckley, 2007), that individual’s thoughts, feelings, and behaviors.
prosocial video games can increase prosocial
behavior (Gentile et al., 2009; Taylor, 2006),
that exercise games are an attractive form General Aggression Model
of physical activity (Rhodes, Warburton, & and General Learning Model
Bredin, 2009; Sell, Lillie, & Taylor, 2008),
The dominant models of social behavior
and that some types of games can improve
in developmental, personality, and social
game-related visual attention skills (e.g.,
psychology are all social-cognitive models.
Green & Bavelier, 2003; Okagaki & Frensch,
Along with numerous colleagues, we devel-
1994). Some of the latest research has studied
oped the General Aggression Model (GAM)
the implications of avatar use on outcomes
to integrate and simplify a wide range of
as diverse as self-concept, social behavior,
more specific social-cognitive models of
and even exercise. Thus, the simple good–bad
aggression (e.g., Anderson & Bushman, 2002;
dichotomy frequently posed by the general
Anderson & Carnagey, 2004; Anderson et al.,
public and the press (“Are video games bad
2007; Anderson & Huesmann, 2003; DeWall
for children?”) is inappropriate.
& Anderson, in press). Briefly, GAM is a 2011
In this chapter, we focus on the effects of
dynamic biosocial cognitive-developmental
playing recreational video games—that is,
model that provides an integrative framework
games not specifically designed for use in
for domain-specific aggression theories. It
educational or therapeutic contexts. (For a
includes situational, personological, and bio-
brief review of all types of video games, see
logical variables. GAM draws heavily on
Barlett, Anderson, & Swing, 2009.) We begin
social-cognitive and social-learning theories
by describing the main theoretical perspec-
that have been developed over the past
tives needed to understand video game effects.
40 years by social, personality, cognitive, and
Next, we review the known effects of recre-
developmental psychologists (e.g., Bandura,
ational video games, focusing on prosocial
1977; Berkowitz, 1989, 1993; Crick & Dodge,
effects (e.g., helping others), antisocial effects
1994; Dodge, 1980, 1986; Geen, 2001;
(e.g., hurting others; stereotyping others), and
Huesmann, 1982, 1988, 1998; Mischel, 1973;
other effects on the individual (e.g., addiction,
Mischel & Shoda, 1995). These perspectives
cognitive skills, exercise, attention control).
paved the way for understanding the learn-
ing and developmental processes involved
in shaping aggressive behavior and how
Theoretical Overview
such processes contribute to the develop-
ment and change of personality. Figure 13.1
Three complementary theoretical perspec- presents an overview of the model. (For
tives are particularly useful when contem- more detailed views and descriptions, see the
plating the effects of playing video games. works cited earlier.)
Prosocial, Antisocial, and Other Effects of Recreational Video Games • 251
Figure 13.1 The General Aggression Model, Overview
Personality Biological Environmental
Development: Modifiers Modifiers
Distal Factors
and Processes
Personality
Social
Encounters: Person Situation
Proximal
Factors and
Processes
Cognition
Present
Internal
State
Affect Arousal
Appraisal
Decision
Behavior
Source: Adapted from Violent Evil and the Aggression Model (pp. 168–192), by C. A. Anderson and N. L. Carnagey,
2004, In A. Miller (Ed.), The social psychology of good and evil. New York: Guilford.
To understand aggression, or any social interprets his or her environment and the
behavior, we must understand how such behav- people therein—on expectations regarding
ior in general depends on cognitive, affective, the likelihood of various outcomes, on knowl-
and arousal factors within the individual. edge and beliefs about how people typically
Note that in Figure 13.1 no specific reference respond in certain situations, and on efficacy
is made to aggressive behavior. One implica- beliefs about the ability to respond to the
tion is that the model can be applied to other ongoing events. By understanding these percep-
types of social behavior, such as helping other tions and cognitions, researchers have a basis
people. This generalization is the essence of for understanding both within-person cross-
what has become known as the General Learn- situational stability across time (because people
ing Model (for more details, see Barlett & show stability in how they perceive their
Anderson, in press; Buckley & Anderson, 2006; social world over time), between person vari-
Gentile et al., 2009). ability within the same situations (because dif-
In any specific social encounter (lower ferent people perceive situations differently),
portion of Figure 13.1), social behavior and between person similarity within the
depends on how an individual perceives and same or similar situations (because situations
252 • The Popular Media as Educators and Socializers of Growing Children
frequently impose realistic demands that of as personality, illustrated by the dashed line
limit the number of options regarding how linking the proximal portion of Figure 13.1 to
people can construe the situation). Further- the distal portion. In this way, the present event
more, such social-cognitive models also influences the future.
account for variability in aggression across The past influences the present by affect-
time, people, and contexts because different ing what the person brings with them to the
knowledge structures develop and change and situation (e.g., knowledge structures) and by
different situational contexts prime different influencing the kinds of situations the person
knowledge structures. The three main keys to is likely to encounter. Figure 13.2 illustrates
these models involve discovering (1) what one way in which repeated exposure to a spe-
the person brings with him or her to the situ- cific type of environmental modifier—media
ations (knowledge structures, such as attitudes, violence—creates an aggressive personal-
beliefs, scripts, perceptual biases), (2) what ity, which in turn increases the likelihood of
types of knowledge structures are primed aggressive behavior in a specific situation. A
or activated by the features of the current situ- similar figure could be produced to illustrate
ation, and (3) how various life experiences how repeated exposure to prosocial media
combine with biological predispositions to can increase an altruistic personality and pro-
create and change personality, conceived of social behavior.
here as the person’s set of operative knowl- Repeated exposure to media violence, such
edge structures. as playing violent video games, increases the
GAM focuses heavily on how the devel- accessibility of a host of aggression-related
opment and use of knowledge structures influ- knowledge structures while simultaneously
ences both early (e.g., basic visual perception) decreasing feelings of empathy for victims of
and downstream (e.g., judgments, decisions, violence and decreasing negative emotional
behaviors) psychological processes (e.g., reactions to violent thoughts, images, and
Bargh, 1996; Collins & Loftus, 1975; Fiske scripts. In essence, such exposure increases
& Taylor, 1991; Higgins, 1996; Wegner & the aggressive personality. When an ambigu-
Bargh, 1998). Of particular interest from cog- ous provocation of some kind occurs (e.g.,
nitive psychology are findings showing that getting bumped in a cafeteria or bar), the
through repeated practice and exposure, com- increased accessibility of aggressive schemata
plex judgments and choices become automa- increases the likelihood that a hostile attribu-
tized, requiring little or no mental effort tion for the event will be made. This increases
or conscious awareness (Bargh & Pietromonaco, the likelihood that aggressive response options
1982). With practice, one can learn to auto- will be generated and that they will be
matically scan the environment for threat, selected for action. In short, the person
to perceive threat even in ambiguous situa- becomes more likely to retaliate. That retali-
tions, and to respond to threat with aggres- ation itself produces an outcome that influ-
sive action. Alternatively, one can learn to ences one’s expectations and beliefs about
automatically look for people in need of the future. The dashed line indicates that such
help, to perceive such needs quicker or more learning has an impact (small) on further per-
frequently than others, and to respond to sonality development. The victim’s response,
this perception by offering help. Indeed, it very likely an aggressive one, sets the stage
is possible that both of these sets of knowl- for the next episodic cycle of this social inter-
edge structures could be well learned (autom- action event.
atized) within the same person. Again, keep in mind that the same type of
The learning processes involved include dynamic process occurs with repeated expo-
all of the well-studied processes of classical sure to prosocial media. Of course, in this
conditioning, instrumental conditioning, case more prosocial thoughts, feelings, and
imitation, and higher order forms of learning. shifts in personality are instigated by the
These learning processes, which result from media exposure, resulting in more prosocial
discrete social encounters, affect what we think behavior under the right circumstances.
Prosocial, Antisocial, and Other Effects of Recreational Video Games • 253
Figure 13.2 Media violence effect on personality, and personality effect on aggression
Repeated Exposure to
Media Violence
Increased
Aggressive Decreased empathy,
aggressive
personality anxiety to violent
cognitions
thoughts
Initial Generate Select
Make a hostile
ambiguous aggressive aggressive
interpretation
provocaion options options
Victim
responds with
Retaliate
aggression
with
Retaliation is
aggression
sometimes
rewarded, but also
provokes victim
Aggression escalation cycle
demonstrated a relationship between amount
Five Dimensions of Video
of sedentary gaming and obesity (Berkey et al.,
Game Effects
2000; Laurson et al., 2008; Vandewater, Shim,
Gentile and his colleagues (Gentile, in & Caplovitz, 2004). Again, it is likely that sed-
2011
press; Gentile & Stone, 2005; Khoo & Gentile, entary games displace other more physically
2007) have proposed five dimensions along active activities, and children may also snack
which video games can have effects—the more while gaming than they otherwise would.
amount, content, context, structure, and mechan- Amount of gaming has also been implicated in
ics. This approach explains how research find- repetitive stress disorders (Brasington, 1990),
ings that initially appear contradictory are and in video game addiction (see later section).
actually congruent. Games and the ways people Most of the research on video game effects
interact with them are multidimensional, and has focused on the content dimension, with
each dimension is likely to be associated with the bulk of that research focused on violent
specific types of effects. content. It should be no surprise that people
The amount of time that people spend on learn the content of whatever games they
recreational games can have effects on them, play. If they play educational games, they
regardless of specific game features. Studies learn the educational content and can apply it
have demonstrated that amount of time play- to their schoolwork (Murphy et al., 2002); if
ing games predicts poorer school performance they play games designed to teach health con-
(see later section). Theoretically, this effect is tent, they learn those concepts and apply them
likely due to displacement of other academi- to their lives (Beale, Kato, Marin-Bowling,
cally beneficial activities. Other studies have Guthrie, & Cole, 2007; Lieberman, 2001); if
254 • The Popular Media as Educators and Socializers of Growing Children
they play violent games, they learn the vio- Bowlby, 2008) or to improve dynamic bal-
lent content and may apply it to their lives ance control after brain surgery (Betker, Szturm,
(see sections that follows). Moussavi, & Nett, 2006). The intersection of
The context of game play may produce structure and mechanics is the continuous
differential effects, but this is the dimension feedback loop that is often referred to as hand-
with the least research at this time. Context eye coordination.
can be defined within the game or outside of One benefit of noting the dimensions on
the game. One type of within-game context which games can have effects is that it allows
can be seen in violent games that allow for us to recognize that the dichotomous question
either team-based or free-for-all modes of of whether games are good or bad is too sim-
play. Both may be equally violent, but play- plistic. Games have multiple effects at multi-
ing in an everyone-for-oneself mode might ple levels of analysis, some of which may be
lead to greater aggressive thoughts, lower beneficial and some of which may be harmful,
empathy, and greater desensitization. If the even within the same game.
in-game context requires players to cooperate
to achieve goals, this might also teach team- Research Designs and
work and social coordination skills. Further-
Scientific Causality
more, the social context outside of the game
may matter. Playing a violent game in a room Scientists use three main types of research
with other friends (virtual or real) might designs to test theoretical models. Each type
increase the aggression effect because players has characteristic strengths and weaknesses.
are giving each other social support for No single study is thought of as conclusive,
aggression. It might actually reduce the though some are stronger than others. Because
aggression effect, however, if one’s motiva- no single empirical study can be wholly con-
tions are prosocial (to help your friends). To clusive, researchers create and test theories
our knowledge, no studies have yet tested using multiple studies and multiple study
these hypotheses. designs in order to triangulate on the clearest
How the game is structured and displayed answer to the research question. Different
on the screen can also have effects. This designs are used to test various plausible alter-
screen structure provides information that is native hypotheses. By ruling out such alterna-
learned, similar to how we learn to perceive tives, the main causal hypothesis gains strength.
other visual information (Gibson, 1979). If a plausible alternative explanation remains
Perceptual skills can be improved through viable, then additional empirical tests or a modi-
practice, as has been shown in several studies fication of the theory is required. In sum, it
(Green & Bavelier, 2003, 2006a, 2006b, 2007; is the total picture of combined studies that
Greenfield, DeWinstanley, Kilpatrick, & Kaye, answers the question of whether or not a particu-
1994). Other studies have demonstrated an lar variable (e.g., violent video game expo-
improved ability to get three-dimensional infor- sure) is a causal factor for a particular outcome
mation from flat screens (Greenfield, Brannon, (e.g., aggressive behavior).
& Lohr, 1994) or improved mental rotation The three major types of studies are exper-
skills (Cherney, 2008) after playing games that imental, cross-sectional correlational, and longi-
require those skills. tudinal studies. Experimental studies randomly
Finally, the mechanics dimension refers to assign participants to different groups: for
what can be learned from practice with differ- example, playing either a violent or non-violent
ent types of game controllers. Depending on video game. A major result of random assign-
the type of controller, several different skills ment is that individual differences among the
could be improved, including fine motor skills participants should be equally distributed
(e.g., by using a thumb controller), gross motor between the groups, creating groups that
skills (e.g., by swinging the Wii remote like are equivalent (more or less) at the outset—
a golf club), or balance skills (e.g., using the even on individual differences that have
Wii balance board). These effects can be used not been measured. The experimenter measures
intentionally, such as in physical therapy the outcome variable after the experimental
(Deutsch, Borbely, Filler, Huhn, & Guarrera- manipulation takes place. If the groups differ
Prosocial, Antisocial, and Other Effects of Recreational Video Games • 255
on the outcome variable, the researcher can end of the study. This design, when used prop-
conclude with confidence that the manipu- erly, allows some causal conclusions and use
lated variable caused the obtained difference of real world outcome variables. The major
in outcome. This is the great strength of limitations of longitudinal studies are that
experimental studies. The major weakness is they are difficult and expensive.
that one frequently cannot use strong real- The strongest case for establishing that a
world measures of the conceptual outcome hypothesized effect is causal arises when the
variable. For example, it would be unethical same conceptual results are obtained regard-
to actually allow study participants to physi- less of the research design—and when plau-
hit cally beat each other; therefore, more ethical sible alternative explanations have been tested
measures must be used. In such cases, research- and ruled out. In the video game domain, such
ers use ethically appropriate measures that variables might include sex, personality trait
faithfully represent the conceptual outcome hostility, parental education level, and paren-
variable. Ideally, though not necessarily, one tal monitoring of media, among others. Experi-
would use measures that also predict or are mental studies by their very nature control for
predicted by more extreme real-world mea- such individual difference variables, even if
sures. For example, the competitive reaction the variables haven’t been measured. So do
time task frequently used to measure physical longitudinal studies, at least to some extent.
aggression in the lab has repeatedly shown Cross-sectional studies have no inherent con-
such high external validity. trol for such potential confounds, but if such
Cross-sectional correlational studies allow variables are measured, then they can be sta-
researchers to get beyond the outcome vari- tistically controlled.
able limitation of experimental studies. In a
cross-sectional study, for example, research- Scientific Causality,
ers might survey children about the video
Risk, and Resilience
games they play and about several real-world
types of aggressive behavior, such as how It is important to note what scientifically-
many physical fights they have had. The established causality means and does not
major weakness of cross-sectional studies is mean, as this often creates confusion for the
that claims of causality are more risky, espe- public. Modern scientific causality is not the
cially from a single study. For example, it “necessary and sufficient” causality of old.
might be that playing violent games causes Instead, it is probabilistic. For example, “smok-
aggressive behavior, or that aggressive chil- ing causes lung cancer” means that repeated
dren are more likely to play violent games, or smoking of tobacco products increases the
that some third variable causes both (such as likelihood that one will contract lung cancer.
poor impulse control). Cross-sectional stud- It does not mean that all smokers get lung
ies are strong where experimental studies are cancer or that only smokers get it. Similarly,
weak and vice versa. According to the trian- “media violence causes aggression” means that
gulation notion, if both types of studies show exposure to violent media increases the like-
similar results, researchers can be reasonably lihood of later aggressive behavior. It does
comfortable in assuming that they have dis- not mean that all violent game players will
covered a real causal effect on important, real behave aggressively, or that only violent
outcomes. And of course, many plausible game players behave aggressively. The prob-
alternative explanations can be tested using abilistic nature of modern causality results
cross-sectional designs. from the fact that human health and behavior
Longitudinal studies document changes is multicausal. This approach to understand-
over a longer period of time. For example, ing the multicausality of behavior is sometimes
one might measure children’s video game known as the risk and resilience approach
habits and their aggressive behaviors at two (Gentile & Sesma, 2003).
points in time separated by 6 months. One Years of study have documented scores of
can then test whether children who play vio- variables that individually increase the likeli-
lent games at the beginning of the study change hood of concurrent or future aggression. These
to become relatively more aggressive by the include variables at many different levels of
256 • The Popular Media as Educators and Socializers of Growing Children
analysis, including poverty, having been bul- media violence (Anderson et al., 2003; Gentile,
lied, taking drugs, genetic risk for aggression, 2003). Unfortunately, many parents who take
poor parenting, and media violence (Surgeon great pains to keep children from witnessing
General, 2001). Each of these individually real violence in the home and neighborhood
is a risk factor for aggression—that is, each often do little to keep them from viewing
increases the risk of current or future aggres- large quantities of violence on television, in
sive behavior. But no single factor alone is movies, and in video games. This lack of
sufficient to elicit more extreme forms of parental concern about media violence is per-
aggression. This does not mean that we can plexing given the research on its harmful
ignore any of them—each is important, and effects, the large number of national reviews
steps could be taken to minimize them. of the research that have publicized these
In addition, there are factors that help effects, and the strong critique of media vio-
lower the risk of aggression: These are pro- lence by pediatricians. We do not mean to
tective factors. These also include variables suggest that there is no controversy over how
at many different levels of analysis (e.g., hav- to interpret the results of studies. What appear
ing prosocial peers, having highly involved to us to be overreactions happen in both
parents, being female, and certain genetic/ directions, from some people claiming that
biological factors). To predict and understand violent games are training a generation of
which people will behave aggressively, there- killers to others claiming there is no effect at
fore, we need to understand the risk and pro- all. For example, attorney (now disbarred)
tective factors that each individual has. Each Jack Thompson stated on ABC’s World News
additional risk factor increases the risk, Now program (March 23, 2000): “In every
whereas each protective factor decreases the school shooting, we find that kids who pull
risk. Not every risk and protective factor has the trigger are video gamers.” On the other
the same effect size; some are more important extreme, psychologist Christopher Ferguson
than others (Anderson et al., 2003; Surgeon claimed that the existing research is largely
General, 2001). Furthermore, some risk fac- “pseudoscience” (2009) and that researchers
tors may interact, increasing their effects are attempting to create a “moral panic”
more together than they would individually, (2008; 2010). Neither of these claims seems
although much more research is needed in reasonable to us. The most recent compre-
this area. Ultimately, it is important to under- hensive review of the overall media violence
stand that when scientific psychological literature (including television, movies, and
research documents a causal effect of violent video games) documents the “. . . unequivocal
games on aggression, for example, this means evidence that media violence increases the
that violent games are one risk factor increas- likelihood of aggressive and violent behavior
ing the likelihood of aggressive behaviors, in both immediate and long-term contexts”
not that they are the only cause of aggressive (Anderson et al., 2003, p. 81). Interestingly, a
behavior. In order for a child to behave seri- 2004 survey of pediatricians found that over
ously violently, he or she would need to have 98% believed that the media affect childhood
multiple risk factors and few protective fac- aggression (Gentile et al., 2004). But some-
tors (Gentile & Sesma, 2003). how this message has not been convincingly
delivered to, or always understood by, aver-
age Americans (Dill, 2009a).
Violent Video
Game Effects A Recent Meta-Analysis
In this section, we review the available
Much research over several decades has research on the effects of playing violent
documented how witnessing violence and video games. Six outcome variables have
aggression leads to a range of negative out- received sufficient research attention to war-
comes for children. Negative outcomes result rant inclusion in a recent comprehensive
both from witnessing real violence (e.g., meta-analytic review: aggressive behavior,
Osofsky, 1995) as well as from viewing aggressive cognition, aggressive affect,
Prosocial, Antisocial, and Other Effects of Recreational Video Games • 257
physiological arousal, prosocial behavior, Main Findings
desensitization/low empathy (Anderson et al.,
The main findings can be succinctly sum-
2010). This meta-analysis is considerably larger
marized: Playing violent video games causes
than any prior meta-analysis of the violent
an increase in the likelihood of physically
video game effects literature (e.g., Anderson,
aggressive behavior, aggressive thinking,
2004; Anderson & Bushman, 2001) for two
aggressive affect, physiological arousal, and
primary reasons: (1) It included a large num-
desensitization/low empathy. It also decreases
ber of previously unavailable studies from
Japan; and (2) there has been an explosive helpful or prosocial behavior. With the excep-
growth in research on this topic in recent tion of physiological arousal (for which there
years. Specifically, 381 effect sizes were are no cross-sectional or longitudinal stud-
obtained from 136 research reports involving ies), all of the outcome variables showed the
over 130,000 participants. In the following same effects in experimental, cross-sectional,
sections, all reported meta-analytic results are and longitudinal studies. The main effects
from this new meta-analysis, and all are occurred for both males and females, for par-
based on the sample of studies that met all ticipants from low-violence collectivistic type
best practices criteria (the “Best Raw” sam- Eastern countries (e.g., Japan), and from high-
ple in Anderson et al., 2010). This sample violence individualistic type Western countries
included 221 effect sizes involving over (e.g., United States, Europe). Table 13.1 dis-
61,000 participants.1 plays the major findings of the meta-analyses.
Average effect size of violent video game play. (Results from the “Best Raw” data,
Table 13.1
Anderson et al., 2010.)
Design Total N K Ave. Effect (r+) Z
Physical Aggression
Experimental 2513 27 .210 10.512**
Longitudinal 4526 12 .203 13.787**
Cross-Sectional 14,642 40 .262 32.291**
Aggressive Cognition
Experimental 2887 24 .217 11.695**
Longitudinal 3408 8 .115 6.728**
Cross-Sectional 9976 27 .183 18.445**
Aggressive Affect
Experimental 1454 21 .294 11.289**
Longitudinal 2602 5 .075 3.836**
(Continued)
1Note that the other two samples—the Full Sample (N = 130,296) and the Best Partials Sample (N = 53,034)—yielded
essentially the same results. That is, in each sample, the average violent video game effect sizes were significant.
258 • The Popular Media as Educators and Socializers of Growing Children
Table 13.1 (Continued)
Design Total N K Ave. Effect (r+) Z
Cross-Sectional 5135 11 .101 7.227**
Prosocial (helping) Behavior
Experimental 633 4 -.182 -4.599**
Longitudinal 2778 5 -.114 -6.022**
Cross-Sectional 3495 7 -.093 -5.506**
Empathy/Desensitization
Experimental 249 1 -.138 -2.175*
Longitudinal 2421 4 -.184 -9.147**
Cross-Sectional 3910 10 -.203 -12.845**
Notes: Total N is the total number of participants in all of the summarized studies. K is the number of different studies.
The Average effect (r+) is the weighted average effect size, expressed as an r-value. Z is the Z-test of whether the effect
is significantly different from zero.
* p < .05. ** p < .001.
Additional Findings some children are more vulnerable to the
effects than others. Markey and Markey (2010),
From a triangulation perspective, the results
for example, provide data suggesting that chil-
of the meta-analysis seem clear. There is
dren with certain personality features (e.g.,
strong evidence that playing violent video
high neuroticism, low agreeableness, and low
games increases aggression and affects a host
conscientiousness) may be the most vulner-
of relevant outcome variables in theoretically
able to violent game effects. Although other
expected ways. The violent game effect
studies have not found a profile that strongly
occurs in both the immediate situation (exper-
suggests increased vulnerability for some
imental studies) and across time (cross-
groups (e.g., boys, highly aggressive individu-
sectional and longitudinal studies). The
als, etc.), this is a promising area of research.
triangulation notion applies not only to using
multiple research designs, but to using mul-
tiple methods and participant populations
Prosocial Video
within each design. The findings were quite
Game Effects
consistent in these ways as well, showing that
the main findings held across culture, age,
and sex. The range of measures used in these There has been little research on the effects of
studies was also impressive. For example, prosocial, nonviolent video games. There are
physical aggression measures included few nonviolent games that have lots of pro-
standard laboratory measures, self-reports of social content—that is, games in which the
fights at school, peer reports, teacher reports, main character’s primary task is to help other
and reports of truly violent behavior. game characters in nonviolent ways. Although
Overall, the basic question of whether vio- some might argue that games in which the
lent games are a potential risk factor for player’s character is a hero killing “bad guys”
increased aggressive thoughts, feelings, and and saving other “good” characters are proso-
behaviors appears to have been answered. cial, the large research literature on TV and
Research is beginning to move into potentially film violence effects clearly demonstrates that
more interesting questions, such as whether such heroic violence increases later aggression,