Journal of Experimental Psychology: Applied Copyright 2007 by the American Psychological Association

2007, Vol. 13, No. 1, 22–31 1076-898X/07/$12.00 DOI: 10.1037/1076-898X.13.1.22

Virtual Driving and Risk Taking: Do Racing Games Increase Risk-Taking

Cognitions, Affect, and Behaviors?

Peter Fischer Jörg Kubitzki

Ludwig-Maximilians University Allianz Center for Technology

Stephanie Guter and Dieter Frey

Ludwig-Maximilians University

Research has consistently shown that aggressive video console and PC games elicit aggressive cogni-
tions, affect, and behaviors. Despite the increasing popularity of racing (driving) games, nothing is
known about the psychological impact of this genre. This study investigated whether playing racing
games affects cognitions, affect, and behaviors that can promote risk taking in actual road traffic
situations. In Study 1, the authors found that the frequency of playing racing games was positively
associated with competitive driving, obtrusive driving, and car accidents; a negative association with
cautious driving was observed. To determine cause and effect, in Study 2, the authors manipulated
whether participants played 1 of 3 racing games or 1 of 3 neutral games. Participants who played a racing
game subsequently reported a higher accessibility of cognitions and affect positively associated with risk
taking than did participants who played a neutral game. Finally, on a more behavioral level, in Study 3,
the authors found that men who played a racing game subsequently took higher risks in computer-
simulated critical road traffic situations than did men who played a neutral game. Theoretical and
practical implications are discussed.

Keywords: racing games, risk taking, obtrusive driving, risk-related cognitions and affect, risk priming

In the last decade, the playing of video console and PC games most actions in racing games imply a very high risk of having an
has become very popular among young people (Gentile, Lynch, accident or severe crash in a highly realistic virtual road traffic
Linder, & Walsh, 2004; Kubitzki, 2005). Meanwhile, the average environment (see, e.g., GameSpot, n.d., for a review of the game
daily time spent playing these games even exceeds the time spent Burnout Revenge [EA Games]), but to date, no one knows what
watching films and TV (Huston, Wright, Marquis, & Green, 1999). influence these games have on variables related to actual road
After several dramatic and murderous shoot-outs in schools, game traffic behavior.
research focused mainly on the impact of aggressive shooter Media violence research has found that aggressive video games
games on aggression-related cognitions, affect, and behaviors increase aggression-related cognitions, affect, and behaviors
(Anderson & Bushman, 2001; Anderson et al., 2004). In addition (Anderson & Bushman, 2001; Anderson et al., 2004). Even parts
to this important research, we believe that the scope of game of rampage killings in schools are often modeled on the game
research must be extended to include possible detrimental influ- content of ego shooters (Anderson & Dill, 2000; Wikipedia, 2006;
ences of the driving game genre on cognitions, affect, and behav- e.g., it turned out that the killers in the Columbine High School
iors related to risk taking in actual road traffic. In the last few massacre played first-person shooter games excessively and, in
years, racing games have emerged as top sellers in the computer addition, recorded a video before their rampage killing in which
game industry. Within photographically realistic virtual environ- they stated that they were looking forward to using their shotguns
ments (i.e., the visual game environment is extremely similar to just as in the game Doom [id Software], which is one of the earliest
real driving environments), players race through urban and subur- and most famous first-person shooters). What if players of racing
ban traffic. Driving actions in these games often include compet- games similarly model their actual road traffic behavior on their
itive and reckless driving, speeding and crashing into other cars or behavior during these games? For example, a controversy about
pedestrians, or performing risky stunts with the vehicle. In short, racing games developed when on January 26, 2006, following a
horrific accident involving two street racers that resulted in the
death of a taxi driver in Toronto, the racing game Need for Speed
Peter Fischer, Stephanie Guter, and Dieter Frey, Department of Psy- [EA Games] was found in one of the suspects’ cars (Wikipedia,
chology, Ludwig-Maximilians University, Munich, Germany; Jörg Ku- 2006). Moreover, in the context of a focus-group study (Kubitzki,
bitzki, Allianz Center for Technology, Munich, Germany.
2004), in-depth interviews of young male participants in illegal
This research was funded by the Allianz Center for Technology.
Correspondence concerning this article should be addressed to Peter street races revealed that computer and video driving games indeed
Fischer, Department of Psychology, Ludwig-Maximilians University, played an important role in the participants’ driving-related social-
Leopoldstrasse 13, 80802, Munich, Germany. E-mail: pfischer@psy.uni- izations. The present research sought to determine whether a
muenchen.de similar detrimental relationship exists between playing risk-

22
 RACING GAMES AND RISK TAKING 23

oriented video racing games and variables related to risky driving fact that the latter often contain considerable risky, harmful, and
behavior as exists between playing aggressive video games and aggressive road traffic behavior such as crashing into other players
variables related to aggressive behavior. In short, our studies or driving over pedestrians (e.g., Burnout [EA Games], Grand
investigated the impact of video racing games on cognitions, Theft Auto [Rockstar Games]; for reviews and detailed descrip-
affect, and behaviors related to risk taking. tions of these games, see, e.g., http://www.gamespot.com). Be-
sides eliciting aggressive acts of reckless driving, racing games
Video Game Research allow players to break road traffic rules, and indeed, they often
reinforce this behavior. Given the observed negative impact of
Research on the impact of video games on psychological re- aggressive games on aggressive responding, might racing games
sponses has mostly been conducted in the context of the shooter also have a negative impact on cognitions, affect, and behaviors
genre, which allows players to walk through highly realistic envi- promoting risk taking in actual road traffic situations? In line with
ronments and brutally shoot people and “monsters.” Several stud- the predictions of GAM, we suggest that cognitions, affect, and
ies have revealed that the playing of aggressive games elicits behaviors related to increased risk taking could be activated
aggressive cognitions, affect, and behavioral responses (Anderson (primed) by the playing of racing games. This priming process
& Bushman, 2001; for a recent review on media violence, see also could in turn increase risk taking in real-life driving.
Anderson, Berkowitz, et al., 2003). This research has predomi- So far, little research has addressed this question. In a prelimi-
nantly been performed within the theoretical context of the general nary study, Kubitzki (2005, 2006), who examined 13–17-year-old
aggression model (GAM; e.g., Anderson & Dill, 2000; Anderson boys to quantify racing game activity as well as car-related behav-
et al., 2004; Bushman, 1998), which is based on social– cognitive iors in 657 participants, found that 77% of the respondents played
research, such as the model of social information processing (e.g., racing games regularly; there was a significant positive association
Dodge & Crick, 1990), Bandura’s theory of social learning (e.g., between game-playing intensity and underage driving. Other pre-
Bandura, 1971, 1986), the social– cognitive model of media vio- liminary research conducted for the German Highway Safety Re-
lence (Huesmann, 1986), the excitation transfer model of Zillmann search Institute by Vorderer and Klimmt (2006) reported only
(1983), and the cognitive neoassociationist model of Berkowitz mixed weak and (under specific conditions) short-term effects of
(1984, 1990, 1993). Most important for our line of reasoning,
driving games on road safety. In sum, little research has focused on
Berkowitz’s (1984, 1990, 1993) model suggests that aggressive
this topic, and the research that exists has mostly been often based
ideas promoted by violent media prime other semantically related
on correlative designs or interviews. In addition, the findings of the
thoughts. This priming explanation is, in turn, based on a concept
extant research are equivocal. Thus, it is difficult to answer with
derived from cognitive psychology—namely, spreading activation
certainty the question whether racing games are likely to nega-
within a memory network (e.g., Collins & Loftus, 1975): Thoughts
tively affect road traffic behavior.
(e.g., those triggered by violent games) send out activation along
Below, we present three studies investigating the impact of
associative cognitive pathways and, thereby, activate other asso-
racing games on cognitions, affect, and behaviors positively re-
ciated thoughts. Accordingly, other aggression-related cognitions
lated to risk taking in road traffic. First of all, in an attempt to
and ideas (not identical to those observed in the media) may also
replicate previous correlative findings, we asked young drivers
be evoked by the initial aggressive cognition or idea. Moreover,
within the same associative cognitive network, thoughts are linked questions on their road traffic behaviors (need to impress other
not only to other cognitions but also to behavioral and affective people, competitive driving, cautious driving, accidents) as well as
reactions (e.g., Bower, 1981). In short, with regard to GAM, how often they played current car racing games such as Burnout,
exposure to media violence (e.g., aggressive video and/or PC Midnight Racer (Rockstar Games), or Need for Speed, which are
games) acts as a priming stimulus, which can trigger several among the most popular and frequently played video racing games
associations consisting of aggressive thoughts, expectations, be- worldwide (Study 1). Because correlative studies cannot determine
liefs, and affect related to violence and, thereby, even provide the cause and effect, we directly manipulated whether participants
starting point for aggressive behaviors (Anderson & Bushman, played one of three racing games or one of three neutral (nonrac-
2001, 2002; Anderson et al., 2004; Anderson & Dill, 2000; Bush- ing) games; the dependent variable was the accessibility of cog-
man, 1998). Although Freedman (2002) concluded that research nitions (risk-related terms and ideas) and affect (arousal, excite-
on media violence generally suffers from weak effects, the impact ment) that were supposed to be associated with increased risk-
of violent video games on aggressive behavior has been shown to taking tendencies in road traffic (Study 2). Finally, to more closely
be medium (e.g., Anderson & Dill, 2000 [Study 2], found an effect observe risk taking in actual road traffic behavior, we again had
size [R2] for the impact of violent video game playing on aggres- participants play either a racing game or a neutral game, and we
sive behavior of .14). The present research transfers this line of subsequently measured risk-taking behavior in critical road traffic
reasoning concerning violent media to risk-promoting media con- situations with a standardized and established computer-based
tent. Following social– cognitive models such as GAM, our pro- reaction time measure for risk taking in road traffic (Study 3). In
posal is that racing games can elicit cognitions, affect, and behav- summary, we expected a positive association between the playing
iors that might promote risk taking in real road traffic situations. of racing games and self-reported risk promotion in real road
traffic behavior (Study 1). Furthermore, we expected racing games
(compared with neutral games) to significantly increase cogni-
The Present Research
tions, affect, and behaviors that are supposed to promote risk
In contrast to shooter games, little psychological research has taking in actual road traffic situations (Studies 2 and 3). Because
been carried out on street-racing and driving games, despite the men more frequently play video racing games than women (Ku-
24 FISCHER, KUBITZKI, GUTER, AND FREY

Table 1
Correlations Between Playing of Racing Games and Different Risk-Related Road Traffic Behavior in Study 1

Need to show
Competitive impressive Cautious road Number of tickets
road traffic road traffic traffic Power of Number of reported for norm
Participant group behavior behavior behavior participants’ cars accidents violations

Overall (N ⫽ 198) .49** .43** ⫺.21** .29** .22** .08

Men (n ⫽ 106) .50** .44** ⫺.22** .36** .29** ⫺.05
Women (n ⫽ 92) .33** .28** ⫺.03 .001 ⫺.13 .29**
**
p ⬍ .01.

bitzki, 2005), on an exploratory basis, all analyses statistically (a) competitive road traffic behavior, (b) the need to show impres-
controlled for gender. sive/obtrusive road traffic behavior, and (c) cautious road traffic
behavior. Competitive road traffic behavior was measured with the
Study 1 following items on a scale that ranged from 0 (I don’t agree) to 10
(I definitely agree): (a) “I like to participate in street races”; (b)
Our aim was to collect evidence on whether the playing of “Comparing my driving skills with others is fun”; and (c) “There
racing games is associated with increased risk taking in road should be more street races in German cities” (␣ ⫽ .80). Need to
traffic. Individuals recruited in public places indicated to what show impressive/obtrusive road traffic behavior was measured
extent they played racing games; self-reported risk taking in road with the following items on a scale that ranged from 0 (I don’t
traffic was measured by (a) self-reported competitive behaviors in agree) to 10 (I definitely agree): (a) “I like to be admired when I
road traffic (e.g., street racing), (b) self-reported intentions to overtake other cars”; (b) “Boys should impress their girlfriends by
impress others (e.g., to be admired for risk taking in road traffic), risky car driving”; and (c) “I like to drive very fast when the car is
(c) self-reported attitudes toward cautious behavior in road traffic, full of people” (␣ ⫽ .67). Cautious behavior was measured by the
and (d) self-reported numbers of accidents. The selection of these following items on a scale that ranged from 0 (I don’t agree) to 10
specific driving-related variables was made on the basis of (exist- (I definitely agree): (a) “I mostly respect speed limits”; (b) “I
ing) traffic safety research, which has shown competitiveness, the become anxious when driving too fast”; and (c) “Sometimes I am
desire to impress, sensation seeking, and risk-taking attitudes to be afraid that I will have an accident” (␣ ⫽ .52).
relevant for the prediction of risky driving behavior (e.g., Trimpop Finally, participants indicated on a scale of 0 (never) to 5 (daily)
& Kirkcaldy, 1997). We expected consumption of racing games to to what extent they played the racing games Need for Speed, DTM
be positively associated with competitive and obtrusive behavior Racing (DTM), Colin McRae Rally (Codemasters), and GTR Rac-
in road traffic. Furthermore, consumption of racing games was ing (10tacle Studios) as well as other racing games. We averaged
expected to be negatively associated with cautious behavior and these five items to provide an indicator for the frequency of
positively associated with the number of accidents and norm participants play racing games. After the study, participants were
violations reported. debriefed.

Method Results and Discussion

Participants and design. One hundred and ninety-eight men
The raw correlations and associated statistics are depicted in
and 92 women participated in this study. Participants ranged in age
Table 1. Table 2 presents an overview of the means and standard
from 16 to 45 years (overall: M ⫽ 23.43, SD ⫽ 4.65; men: M ⫽
deviations for the measures used. Because regression procedures
23.59, SD ⫽ 5.29; women: M ⫽ 23.24, SD ⫽ 3.80).1
Materials and procedure. Participants were recruited in public
places (e.g., pedestrian zones, fast food restaurants, gas stations). 1
In the context of preliminary survey research on the racing game topic,
They were asked whether they would be willing to participate in a which involved 13–17-year-old participants, Jörg Kubitzki consulted the
study dealing with road traffic behavior. First, participants an- research department of the German government’s youth ministry (which
swered general questions about their driver’s license (“How many studies drug consumption among young Germans, etc.). On the basis of the
months ago did you gain your driver’s license?” [open question]), information provided by the ministry, no parental consent is necessary to
power of their car (“How much horsepower has your current car?” collect data from adolescents under 18 years of age provided (a) the survey
[open question]), numbers of accidents (“How many accidents research has no economic aims, (b) the survey does not delude participants
have you had since you gained your driver’s license?” [open (e.g., by specific manipulations), (c) the young participants are able to
understand the aim and content of the research, (d) no information is
question]), and numbers of tickets received for norm violations in
collected about the parents. In addition, from a technical perspective,
road traffic (“How many tickets have you received since you adolescents have an active as well as a passive interest in car driving
started driving a car?” [open question]). Afterward, we constructed because, in many countries, they are allowed to drive at the age of 16. In
items (derived from previous findings on determinants of risky sum, in Germany, collecting data from 16- and 17-year-old participants
driving behavior; e.g., Trimpop & Kirkcaldy, 1997) that aimed to without a consent form signed by their parents posed no ethical problem in
measure the most important constructs for our research question: Study 1.
 RACING GAMES AND RISK TAKING 25

Table 2
Means and Standard Deviations (in Parentheses) for Playing of Racing Games and Risk-Related Road Traffic Behavior Measured in
Study 1

Need to show
Racing Competitive impressive Cautious Number of
game road traffic road traffic road traffic Power of Number of tickets for
Participant group consumption behavior behavior behavior participants’ carsa reported accidents norm violations

Overall (N ⫽ 198) 0.41 (0.78) 3.02 (2.22) 2.51 (1.71) 4.64 (1.93) 101.08 (53.84) 1.28 (2.50) 2.12 (4.69)
Men (n ⫽ 106) 0.63 (0.84) 3.56 (2.37) 2.82 (2.00) 4.14 (1.71) 116.44 (64.27) 1.59 (3.14) 1.58 (3.40)
Women (n ⫽ 92) 0.15 (0.62) 2.39 (1.84) 2.18 (1.22) 5.23 (2.00) 82.10 (27.65) 0.92 (1.36) 2.70 (5.75)
a
Values represent horsepower.

would allow us to statistically control for other potentially irrele- are positively related to risk taking (the counterpart to aggression-
vant (but confounded) factors (e.g., age), we ran hierarchical related cognitions in aggression research), arousal/excitement
regression analyses to determine the effect of frequency of playing (which has been shown in previous work to be a frequent affective
video racing games on driving-relevant variables. Therefore, we consequence of positive risk priming; e.g., Guter, 2006), and
combined the four major dependent variables— competitive road positive and negative affect (to statistically control for a possible
traffic behavior, need to show impressive/obtrusive road traffic general impact of affect). On the basis of the effect of aggressive
behavior, cautious road traffic behavior (reversed), and number of media on aggressive thoughts and affect (which are regarded as the
accidents reported (all rs ⫽ .19 –.56, all ps ⬍ .01)—to form a scale main determinants of aggressive behavior), we expected that par-
of risky road traffic behavior (␣ ⫽ .62). We used hierarchical ticipants who played a racing game would have a higher accessi-
regression analysis to determine potential main effects and inter- bility of cognitions and affects (arousal/excitement) that are sup-
actions. In the first step, we included the two main effects of posed to promote risk taking than would participants who played
frequency of playing racing games, ␤ ⫽ .42, t(195) ⫽ 6.50, p ⬍ a neutral game.
.01, and gender, ␤ ⫽ ⫺.18, t(195) ⫽ ⫺2.72, p ⬍ .01. In the second
step, we included the Frequency of Playing Racing Games ⫻ Method
Gender interaction, which reached significance, ␤ ⫽ ⫺.49,
t(194) ⫽ ⫺2.64, p ⬍ .01 (this interaction remained significant Participants and design. Eighty-three students at Ludwig-
[p ⬍ .05] when the age of participants was statistically controlled Maximilians University, Munich, Germany, volunteered to partic-
for in Step 1 of the hierarchical regression analysis), and indicates ipate for experimental credit. The sample consisted of 36 women
that the positive association between frequency of playing racing and 47 men, ranging in age from 19 to 42 years (overall M ⫽
games and risky road traffic behavior is stronger for men than for 23.86, SD ⫽ 4.40; men: M ⫽ 23.96, SD ⫽ 4.47; women: M ⫽
women. The effect size of the overall model (Cohen’s f2) was .25. 23.72, SD ⫽ 4.36). The experiment had a 2 (type of game: racing
Study 1 provided initial evidence that the consumption of racing game vs. neutral game) ⫻ 2 (gender) between-subjects design.
games is associated with increased obtrusive and competitive road Materials. Six different current computer games were used.
traffic behavior, reduced cautious road traffic behavior, and an The selection of the computer games was made on the basis of
increased number of reported accidents. These associations were attractiveness and sales rankings in different computer game mag-
more pronounced for male participants than for female partici- azines. Three were typical racing games (Burnout, Midnight
pants. A limitation of Study 1 is that because of time restrictions Racer, and Need for Speed), and three were neutral games that
and restrictions imposed by the maximum feasible questionnaire contained no racing content (Tak [THQ], Crash Bandicoot [Vi-
length (the study was conducted in public places), we used only vendi Universal Games], and Fifa 2005 [EA Games]). All games
self-constructed items. As a consequence, only two (competitive were played on a Sony PlayStation 2 platform and a 72-in.
behavior: ␣ ⫽ .80; impressive/obtrusive behavior: ␣ ⫽ .67) of the (182.88-cm) TV screen. All racing games involved participants
three scales provided a good reliability coefficient. The reliability racing through cities against other (computer) gamers. To win,
coefficient of cautious behavior (␣ ⫽ .52) was rather low, and participants had to massively violate traffic rules (e.g., drive on the
thus, future research should use standardized scales with more sidewalk, crash into other cars, drive at high speed). In contrast,
items to obtain higher reliabilities for this construct. Moreover, neutral computer games involved participants in soccer games
because Study 1 had a correlative design, we could not determine (Fifa 2005) or jump and run actions (Tak, Crash Bandicoot). The
cause and effect. Perhaps a priori competitive or obtrusive road assignment of participants to games was random, each game was
traffic behavior might lead to preferences for increased racing played by an approximately equal number of participants, and
game consumption. Hence, to make a stronger causal point, we gender was approximately proportional across experimental con-
used an experimental research design in Study 2. ditions.
As dependent variables (derived from aggression research; e.g.,
Anderson & Bushman, 2002; Anderson, Carnagey, & Eubanks,
Study 2
2003), we measured accessibility of risk-related cognitions, risk-
Participants played either one of three racing games or one of related affect (excitement/arousal), and general positive and neg-
three neutral computer games. We then measured cognitions that ative affects (to statistically control possible affect-related effects).
26 FISCHER, KUBITZKI, GUTER, AND FREY

Accessibility of risk-promoting cognitions was measured by a because of recent guidelines on the university experimental
homonymous decision task. This type of measure belongs to the course-credit system.
class of implicit (aggression) measures and is based on the as- After 20 min of game play, the experimenter came back and told
sumption that violent media primes (increases the accessibility of) participants that they should now answer several questions on the
aggression-related thoughts, affects, and scripts in implicit seman- game. These questions covered the homonymous decision task
tic memory (see, e.g., Anderson, Carnagey, & Eubanks, 2003). described above (risk-promoting cognitions), arousal, and positive
Participants were given a list of 10 words: Each word had two and negative affect. At the end of the experiment, participants were
possible meanings (homonym), either a positively risk-related thoroughly debriefed about the real aim of the study. Special care
meaning (e.g., kick: stimulation from risky behavior such as fast was taken to ensure that no participant left the experimental lab
driving) or a nonrisk-related meaning (e.g., to kick a ball in with negative or risk-promoting thoughts and/or feelings.
soccer). Participants were asked to define each word. We used the
absolute number of risk-related homonym decisions as a depen-
Results and Discussion
dent variable (for a complete list of the homonyms used, see Table
3). Because no homonym measure existed for the accessibility of Check for interfering effects. Participants’ age, the time that
risk-promoting cognitions in German, we had to construct this had elapsed since they had gained their driver’s license, and the
measure (and, thus, no validity or reliability coefficients are avail- number of kilometers normally driven in 1 yr had no significant
able for it at present). As a second dependent variable, we mea- effect on dependent variables, nor did they interact with experi-
sured arousal by the adjectives (a) aroused, (b) excited, (c) bored mental conditions. With regard to the dependent variables (risk-
(reverse coded), and (d) experiencing a “kick” on a scale of 0 (not promoting cognitions, arousal/excitement, positive and negative
at all) to 10 (extremely; ␣ ⫽ .71). According to the Collins affect), Tukey tests revealed no significant differences among the
Cobuild Advanced Learner’s Dictionary (2003), “if something three (different) racing games or among the three (different) neu-
gives you a ‘kick,’ it makes you feel very excited or very happy for tral video games.
a short period of time” (pp. 603– 604). The arousal scale was Accessibility of risk-related cognitions (homonyms). Means
derived from research on positive risk priming by Guter (2006). and standard deviations are depicted in Table 4. Analysis of
Positive affect and negative affect were measured by four items variance (ANOVA) data are displayed in Table 5. A 2 (game) ⫻
each, derived from the Positive and Negative Affect Scales (PA- 2 (gender) ANOVA revealed a significant main effect of game,
NAS; Watson, Clark, & Tellegen, 1988). indicating that participants who played a racing game (M ⫽ 4.84,
Procedure. After participants arrived in the experimental lab, SD ⫽ 1.84) exhibited a higher accessibility of risk-promoting
they learned that the aim of our study was to test video games. cognitions than did participants who played a neutral game (M ⫽
They would first play a video game for 20 min and subsequently 4.03, SD ⫽ 1.55).
answer questions concerning the game. The experimenter ex- Arousal/excitement. Means and standard deviations are de-
plained how to play the game and observed and/or assisted for the picted in Table 4. ANOVA data are displayed in Table 6. A 2
first 5 min. Afterward, the experimenter left the lab so as to reduce (game) ⫻ 2 (gender) ANOVA revealed a significant main effect of
social-desirability and/or demand reactions from the participants. game, indicating that participants who played a racing game (M ⫽
Note that each participant played exactly one (racing or neutral) 5.29, SD ⫽ 1.88) reported a higher level of arousal/excitement
game before answering the dependent measures. This procedure than did participants who played a neutral computer game (M ⫽
was necessary because of time restrictions: Experimental experi- 4.28, SD ⫽ 2.00).
ence has shown that it takes a minimum of 20 min of playing one Positive and negative affect (PANAS). A 2 (game) ⫻ 2 (gen-
single game before a participant is really “in” the game. So if der) ⫻ 2 (type of affect: positive vs. negative) ANOVA with
participants had played all three racing or all three neutral games, repeated measures on the last factor revealed a significant main
respectively, the experimental procedure (including dependent effect of type of affect, F(1, 78) ⫽ 153.46, p ⬍ .01, Cohen’s f ⫽
measures) would have exceeded 1 hr, which was not possible 1.39, indicating that participants experienced more positive (M ⫽

Table 3
Homonyms Used in Studies 2 and 3

Homonym (in German) Risk-related meaning Neutral meaning

WAGEN To dare A car

SCHNEIDEN To diss somebody on the street To cut (e.g., with scissors)
KICK To get high/excited To kick the ball (e.g., in a soccer game)
RENNEN A race To sprint, to run
SPEED To be fast (e.g., when driving a car) A drug
TRAUEN To dare to do something To get married
FEIGE To be gutless, cowardly Fig, the fruit
RASEN To drive with high speed Grass, lawn
FETZEN Driving fast A rag
SCHLAGEN To beat someone in a race To hit someone in a fight

Note. Words used are homonyms in German but not in English.

 RACING GAMES AND RISK TAKING 27

Table 4 Table 6
Means and Standard Deviations (in Parentheses) for Risk- Arousal/Excitement Analysis of Variance for Study 2
Related Cognitions (Homonyms) and Arousal/Excitement as a
Function of Type of Game Played and Gender in Study 2 Source df F Cohen’s f
*
Game 1 4.91 0.25
Dependent variable
Gender 1 0.26 0.05
Participant group
Game ⫻ Gender 1 0.95 0.03
and experimental Risk-related cognitions Arousal/
Within-group error 78
condition (homonyms) excitement
Note. MSE ⫽ 3.85.
Men
Racing games 4.78 (1.76) 5.43 (1.85)
*
p ⬍ .05.
Neutral games 4.15 (1.23) 4.33 (1.79)
Women
Racing games 4.94 (2.01) 5.07 (1.97)
Neutral games 3.89 (1.85) 4.24 (2.26)
Method
Participants and design. Sixty-eight participants (29 men and
Note. N ⫽ 83.
39 women), ranging in age from 19 to 35 years (overall M ⫽
23.62, SD ⫽ 4.14; men: M ⫽ 24.76, SD ⫽ 3.87; women: M ⫽
5.98, SD ⫽ 2.00) than negative affect (M ⫽ 1.69, SD ⫽ 1.58). No 22.77, SD ⫽ 4.17), participated in this study. The experiment had
further significant effects or interactions were observed. a 2 (type of game: racing vs. neutral) ⫻ 2 (gender) between-
Study 2 revealed that a higher accessibility of risk-promoting subjects design.
cognitions and stronger arousal/excitement was subsequently ob- Materials. Six different games were used. The three racing
served among racing gamers than among neutral gamers. No games were the same as in Study 2; the neutral games were two of
significant differential effect on positive and negative affect by those used in Study 2 (Fifa 2005 and Tak) and an aggressive
racing games was observed. Men and women reacted similarly to shooter game (Medal of Honor [EA Games]), included to assess
racing versus neutral games. Although cognitions and affect are possible effects of aggressive media on risk taking. Gaming con-
crucial mediators of related behaviors elicited by modern media soles were identical to those in Study 2. As in Study 2, the
(e.g., Anderson & Bushman, 2002; Anderson, Carnagey, & Eu- assignment of participants to games was random, each game was
banks, 2003), a shortcoming of Study 2 is that we did not measure played by an approximately equal number of participants, and
behavior directly related to the readiness to take risks in actual gender was approximately proportional across experimental con-
road traffic situations. Hence, Study 3 focused on this matter. ditions.
The main dependent variable was a standardized and widely
accepted test called the Vienna Risk-Taking Test, which is a
Study 3
module from the Vienna Test System (Schuhfried, 2006) and
Participants played either one of three racing games or one of measures— on the basis of reaction times—individual willingness
three neutral games. Then they worked on the Vienna Test System to take risks in road traffic. This test is mainly applied in traffic
for risk-taking (Schuhfried, 2006), a standardized and widely psychology and requires a computer system with a monitor to be
accepted measure of readiness to take risks in actual road traffic conducted. The theoretical background for the test is provided by
situations. Finally, as in Study 2, participants’ accessibility of Wilde’s (1994) theory of risk homeostasis. Participants sit in front
risk-promoting cognitions was measured (homonymous task). We of a computer monitor and learn that they will be confronted with
expected participants who played a racing game to more readily 15 different videotaped risky situations in road traffic (driver’s
take risks in typical road traffic situations than would participants perspective), such as planned overtaking maneuvers and arrival at
who played a neutral game. Furthermore, we expected to replicate railroad crossings that have begun to close. First, the specific
the findings of Study 2; that is, after playing racing games, par- traffic situation was described verbally. Then participants saw the
ticipants would have higher accessibility to positively risk-related critical situation two times. The first time, participants were in-
(risk-promoting) cognitions than would participants who played a structed only to watch the situation. The second time, they decided
neutral game. when they would abandon their maneuver by pressing a key. The
time that elapsed between the start of the sequence and the deci-
sion to abandon it was used as the dependent variable as an
Table 5 indicator of risk taking (the longer the reaction time, the higher the
Risk-Related Cognitions (Homonyms) Analysis of Variance for risk taking). The whole test procedure lasted about 10 min. Ac-
Study 2 cording to its publisher, the reliability (␣) of the basic test is .92.
The construct validity of the test has been shown in three inde-
Source df F Cohen’s f
pendent studies (e.g., Scheiblechner, 1985). A study by Sommer,
Game 1 4.73 *
0.25 Arendasy, Schuhfried, and Litzenberger (2005) revealed that the
Gender 1 0.14 0.01 test correctly identified 89% of accident-free drivers and drivers
Game ⫻ Gender 1 0.30 0.06 who had multiple accidents (criteria validity: R2 ⫽ .636). Unlike
Within-group error 78
the cars and traffic environments in racing games, these situations
Note. MSE ⫽ 2.99. were “real” videotaped situations, not produced by computer
*
p ⬍ .05. graphics. Moreover, the task that the participants had to perform
28 FISCHER, KUBITZKI, GUTER, AND FREY

was entirely different: Playing racing games involved using a Table 8

typical video console joypad, whereas responding to the risk- Means and Standard Deviations (in Parentheses) for Risk-
taking test involved pressing a button on the computer keyboard Related Cognitions (Homonyms) and Risk-Taking Behavior
So the environment as well as the task facing the participants and (Reaction Times) as a Function of Type of Game Played and
the response action required in the risk-taking test were not so Gender in Study 3
similar to the gaming environment as to make transfer inevitable
and, thus, produce some kind of demand effect. Dependent variable
After the Vienna Risk-Taking Test, participants worked on the Participant group Risk-taking behavior
same material that was used in Study 2 to measure risk-related and experimental (reaction times in Risk-related cognitions
cognitions. Finally, to check for interfering effects, we asked condition seconds) (homonyms)
general questions concerning driver’s license, accident history,
Men
sensation seeking (Zuckerman, 1994), and enjoyment of the games
Racing games 7.54 (1.34) 4.71 (1.77)
(see Table 7). After the experiment, participants were debriefed Neutral games 6.41 (1.20) 2.87 (1.55)
and dismissed. Women
Racing games 5.87 (1.21) 4.58 (1.68)
Neutral games 6.49 (1.26) 4.35 (1.53)
Results and Discussion
Note. N ⫽ 68.
Check for interfering effects. Age was significantly related to
risk-taking behavior (r ⫽ .32, p ⬍ .01) but did not interact with
experimental conditions (i.e., the Game ⫻ Gender interaction was displayed in Table 10. A 2 (game) ⫻ 2 (gender) ANOVA revealed
still significant, p ⬍ .01, when age was statistically controlled for a significant main effect of game, indicating that participants who
as a covariate). The time that had elapsed since participants gained played a racing game (M ⫽ 4.64, SD ⫽ 1.69) exhibited a higher
their driver’s license, the number of accidents reported, sensation accessibility of risk-related cognitions than did participants who
seeking, and enjoyment of the games had no significant effect on played a neutral game (M ⫽ 3.71, SD ⫽ 1.69). This main effect
risk-taking behavior, nor did they interact with experimental con- was qualified by a significant Game ⫻ Gender interaction, indi-
ditions. Tukey tests revealed that the aggressive video game did cating that the impact of racing games on risk-related cognitions
not significantly differ from the other neutral games with regard to was more pronounced for men (M ⫽ 2.87, SD ⫽ 1.55 [neutral
readiness for risk taking and accessibility of risk-related cogni- game] vs. M ⫽ 4.71, SD ⫽ 1.77 [racing game]) than for women
tions. A similar picture was obtained when the same analyses were (M ⫽ 4.35, SD ⫽ 1.53 [neutral game] vs. M ⫽ 4.58, SD ⫽ 1.67
separately computed for women and men. Although the sample [racing game]). The main effect of game was still significant when
size was rather small (making it difficult to detect a potential age was statistically controlled as a covariate ( p ⬍ .05). The
difference between the aggressive shooter game and the other Game ⫻ Gender interaction was not significant when age was
control games), on the basis of the present analyses, it seems statistically controlled for.
unlikely that the violent shooter game had a specific aggression- Study 3 replicated and extended the findings of the previous
mediated effect on risk-related variables and, thus, can be com- studies on a more behavioral level: Racing games increased readi-
bined with the sample of control nonracing games. ness to take risks in critical road traffic situations as well as the
Risk-taking behavior. Means and standard deviations are de- accessibility of cognitions positively related to risk taking. This
picted in Table 8. ANOVA data are displayed in Table 9. A priori effect was especially pronounced for men. Women were less prone
contrasts revealed that men took more risks in critical road traffic to be affected by racing games. It is interesting to note that this
situations when they had previously played a racing game (M ⫽ interaction of racing game exposure and gender was not found in
7.54, SD ⫽ 1.34) rather than a neutral game (M ⫽ 6.41, SD ⫽ Study 2. There are a number of conceivable reasons for this: For
1.20), t(64) ⫽ ⫺2.43, p ⬍ .05. In contrast, women’s readiness for example, the realistic driving situations (i.e., the dependent vari-
risk taking was not differentially affected by racing (M ⫽ 5.87, able of risk taking) experienced prior to completing the risky-
SD ⫽ 1.21) versus neutral games (M ⫽ 6.49, SD ⫽ 1.26), t(64) ⫽ cognitions task might have made women more risk averse than
1.54, ns. Moreover, men who played a racing game took more men, resulting in a lower availability of risk-promoting cognitions
risks than did women who played a racing game, t(64) ⫽ 3.79, for women than men. Thus, it would be a fruitful endeavor for
p ⬍ .01. future research to statistically control for possible order effects of
Accessibility of risk-related cognitions (homonyms). Means different risk-related measures as a function of gender and expo-
and standard deviations are depicted in Table 8. ANOVA data are sure to racing games.

Table 7
Wordings of General Questions in Study 3

Topic Question

Driver’s license “How many months have elapsed since you gained your driver’s license?”
Accident history “How many accidents have you had since you gained your driver’s license?”
Enjoyment of the games “How much did you enjoy playing the video game in the current study?”
(0 ⫽ not at all; 10 ⫽ extremely)
 RACING GAMES AND RISK TAKING 29

General Discussion Table 10

Risk-Related Cognitions (Homonyms) Analysis of Variance for
Much evidence has shown that aggressive games increase play- Study 3
ers’ aggression-related thoughts and behaviors. Although it is a
topic of high societal relevance, much less attention has been paid Source df F Cohen’s f
to the detrimental effects of racing games on risk-taking variables.
Consistent with social– cognitive and priming explanations of me- Game 1 6.75* 0.32
Gender 1 2.85 0.21
dia effects, the present research suggests that the playing of racing Game ⫻ Gender 1 4.10* 0.25
games (a) increases the accessibility of thoughts that are positively Within-group Error 64
related to risk taking, (b) leads to enhanced arousal and excite-
ment, and (c) increases risk-taking behavior in critical road traffic Note. MSE ⫽ 2.65.
*
p ⬍ .05.
situations. After finding strong positive correlations between the
playing of racing games and self-reported risk-taking behavior in
road traffic (Study 1), we directly manipulated whether partici- have more traffic violations and accidents (e.g., Laapotti, Keski-
pants played racing or neutral games to determine cause and effect. nen, Hatakka, & Katila, 2001), and have higher sensation-seeking
After playing racing games (compared with control games), par- scores (e.g., Jonah, 1997) than do young women. These significant
ticipants exhibited a higher accessibility of risk-promoting cogni- gender differences, which are evident in accident statistics and past
tions, were more aroused/excited (Study 2), and, most important, research, might be explained by certain effects of socialization,
were more ready to take risks in critical (computer-displayed) road with car driving and motor sports being considered a masculine
traffic situations (Study 3). Whereas racing games significantly matter. Furthermore, because men, in general, are more favorably
affected risk-promoting cognitions and affect of both male and inclined to risk taking (e.g., Elliott, Shope, Raghunathan, &
female participants (although these patterns were more pro- Waller, 2006; Taubman-Ben-Ari & Findler, 2003) and more fre-
nounced for men than for women), significantly increased risk- quently play video racing games than do women (Kubitzki, 2005),
taking behavior in critical road traffic situations was found only for risk-related cognitions, affect, and behaviors might be more
male participants. strongly internalized and embedded in the male cognitive system.
This, in turn, should enhance the likelihood of priming of these
Theoretical and Practical Implications constructs by risk-related media stimuli (similar effects have been
On a theoretical level, our results support social– cognitive ex- shown for aggressive media: participants with a trait disposition of
planations of media priming effects. Similar to the priming effects aggressiveness respond more aggressively to aggressive media
and modeling impact of aggressive media stimuli, we have shown content than do participants with low dispositional aggressiveness;
that “positively framed” risky media content (as provided by see, e.g., Anderson, Carnagey, & Eubanks, 2003). Finally, from a
racing games) activates cognitions and feelings of arousal and more methodological perspective, the gender difference might
excitement that are positively related to increased risk taking, derive from performance differences in the racing games used. For
which in turn (at least for male participants) can lead to actual example, in Study 3, the finding that women in the racing game
increased risk taking in simulated critical road traffic situations. condition scored lower on readiness for risk taking than did those
Hence, social– cognitive models addressing the influence of media in the neutral game condition might be attributable to a stronger
on psychological variables, such as GAM (Anderson & Bushman, feeling among women of having failed at the racing games (e.g.,
2002), should be extended to include possible effects of risk- having more accidents or poorer score performances because of
related media content. less experience with computer games), which could have made
An interesting but not totally unexpected finding was that, some of them more risk averse. To test this account of the observed
overall, female participants’ risk-related responding was less gender differences, subsequent research should try to additionally
strongly affected by racing games than was that of male partici- assess participants’ performance (e.g., number of accidents, score
pants. There are several possible reasons for this finding. For level, rankings) in racing games played. In short, future research
example, numerous studies have shown that young men are more should always statistically control for possible gender effects and,
attracted by the video game medium (e.g., Krahé & Möller, 2004), in addition, empirically examine possible reasons for the different
are more affected by motor sports in any respect (Kubitzki, 2004), impact of racing games on risk-related responding of women and
men.
On a practical level, our results pose the question whether
Table 9 playing racing games leads to accidents in real-life road traffic. We
Risk-Taking Behavior Analysis of Variance for Study 3 demonstrated that playing racing games affects risk-taking cogni-
tions, affect, and behaviors. Research on media violence has
Source df F Cohen’s f shown that cognitions, affect, and behavioral intentions are the
Game 1 0.70 0.11 most important mediators for determining effects of media content
Gender 1 6.71* 0.32 on real-life behaviors (e.g., Anderson, Carnagey, & Eubanks,
Game ⫻ Gender 1 8.11** 0.36 2003, Anderson et al., 2004). From traffic safety research, it is
Within-group error 64 known that driving-related attitudes and intentions strongly deter-
Note. The Game ⫻ Gender interaction was still significant when age of
mine driving behavior (Kubitzki, 2005; for a detailed theoretical
participants was statistically controlled for ( p ⬍ .01). MSE ⫽ 1.56. discussion supporting this argument, see also Åberg, 2001, and de
*
p ⬍ .05. ** p ⬍ .01. Pelsmacker and Janssens, in press, who reported a correlation of
30 FISCHER, KUBITZKI, GUTER, AND FREY

.47 between driving attitudes [intentions] and speed violation arousal and excitement, arousal was regarded as a crucial affective
behavior). Putting these arguments and empirical evidence to- difference between racing and nonracing games (and, therefore, it
gether, we conclude that playing racing games could provoke was used as a dependent variable rather than statistically controlled
unsafe driving (e.g., we found that male players of racing games for in a pretest). However, GAM and previous violent video game
reacted about 1 s later to critical road traffic situations than did research do not necessarily suggest that arousal will be higher in
male players of neutral games). To conclude, practitioners in the high-risk games (violent or racing) than in the games used for
field of road traffic safety should bear in mind the possibility that comparison. The level of arousal strongly depends on the charac-
racing games indeed make road traffic less safe, not least because teristics of the games used for comparison; one can choose very
game players are mostly young adults, acknowledged as the high- arousing but nonrisky games. Therefore, several studies of violent
est accident-rate group. video games have preselected violent and nonviolent games that
were equally arousing so that the potentially observed differences
Limitations and Future Research in aggressive responding could be clearly attributed to violent
game content rather than to differences in arousal. Future research
A limitation of our studies is that we only used games with road on the impact of racing games on risk-related road traffic behavior
traffic (street-racing) settings. Many racing games (e.g., Formula 1 should use similar pretests and research designs.
[SCEE]) only let players drive around dedicated racing circuits Finally, on the basis of the present studies, we cannot determine
without traffic and, thus, do not allow players to break traffic rules. whether the effects reported are only short-term or are also long-
Hence (although it is not empirically clear whether Formula 1 term effects. Theoretically, Kubitzki (2005) argued that racing
driving simulations are qualitatively different from the kinds of games reward illegal, reckless, competitive, and aggressive driv-
racing games used in the present research), we do not know ing, and thus— given that the average initial age for playing racing
whether similar effects of racing games on risk-related responding games is 10 —this systematic reward could have a long-lasting
could be observed for the whole driving genre. influence on the development of attitudes and habits in driving.
Another point to discuss is whether our dependent variables are Moreover, the positive correlation between frequency of playing
valid indicators of real-world driving behavior. To answer with racing games and risk taking in real-life driving situations further
certainty whether racing games increase risk taking in road traffic, suggests that there might be a long-term connection. However,
studies demonstrating a game effect on players’ accident and future research should indeed investigate whether it is possible to
traffic violation rates would be needed. Also, simulator studies (in extrapolate from short-term effects of racing games to possible
which actual driving behavior, or more realistic simulated driving long-term effects by using a longitudinal study design.
behavior, serves as a dependent variable) are desirable (although
these measures might be even more prone to be biased by social
desirability than the measure we used in the present Study 3). References
Another shortcoming of Studies 2 and 3 is that participants only
Åberg, L. (2001). Attitudes. In P.-E. Barjonet (Ed.), Traffic psychology
completed a single game before answering the dependent mea-
today (pp. 119 –136). Norwell, MA: Kluwer Academic.
sures. Although this procedure was necessary because of time Anderson, C. A., Berkowitz, L., Donnerstein, E., Huesmann, L. R., John-
restrictions, the heterogeneity of the tasks (games) could possibly son, J., Linz, D., et al. (2003). The influence of media violence on youth.
have resulted in effects being obscured in the present study. Future Psychological Science in the Public Interest, 4, 81–110.
research should plan longer playing sessions (e.g., 2 hr) to permit Anderson, C. A., & Bushman, B. J. (2001). Effects of violent video games
participants to play a wider range of a specific type of game, thus on aggressive behavior, aggressive cognition, aggressive affect, physi-
preventing effects of possible heterogeneity among the stimulus ological arousal, and prosocial behavior: A meta-analytic review of the
material. scientific literature. Psychological Science, 12, 353–359.
We have predicted and interpreted our findings solely in context Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual
of the GAM. However, one could also construct an alternative Review of Psychology, 53, 27–51.
Anderson, C. A., Carnagey, N. L., & Eubanks, J. (2003). Exposure to
framework for the present findings that is based on the classic
violent media: The effects of songs with violent lyrics on aggressive
frustration–aggression hypothesis. That is, individuals who play
thoughts and feelings. Journal of Personality and Social Psychology, 84,
street-racing games experience a variety of frustrating events (e.g., 960 –971.
being overtaken, speeded out, or rammed by other cars), which, in Anderson, C. A., Carnagey, N. L., Flanagan, M., Benjamin, A. J., Eubanks,
turn, might lead to a higher risk acceptance by the player so as to J., & Valentine, J. C. (2004). Violent video games: Specific effects of
respond aggressively to other road users. Furthermore, an in- violent content on aggressive thoughts and behavior. Advances in Ex-
creased level of desensitization for risky road traffic situations (on perimental Social Psychology, 36, 199 –249.
a more perceptual level) might also be an explanation for our Anderson, C. A., & Dill, K. E. (2000). Video games and aggressive
present results. Because we cannot rule out these alternative ac- thoughts, feelings, and behavior in the laboratory and in life. Journal of
counts on the basis of the present data, it might be a fruitful Personality and Social Psychology, 78, 772–790.
Bandura, A. (1971). Social learning theory of aggression. In J. G. Knutson
endeavor for future research to investigate whether frustration–
(Ed.), Control of aggression: Implications from basic research (pp.
aggression theory or desensitization accounts can also be applied
201–250). Chicago: Aldine-Atherton.
to explain the impact of racing games on risk-related behavior. Bandura, A. (1986). Social foundations of thought and action: A social
A further limitation of the present research concerns the role of cognitive theory. Englewood Cliffs, NJ: Prentice Hall.
arousal/excitement as risk-related affect. Derived from previous Berkowitz, L. (1984). Some effects of thoughts on anti- and prosocial
work (Guter, 2006), which found that risk priming (e.g., by pic- influence of media events: A cognitive neoassociationist analysis. Psy-
tures, movies, and games) systematically increases feelings of chological Bulletin, 95, 410 – 427.
 RACING GAMES AND RISK TAKING 31

Berkowitz, L. (1990). On the formation and regulation of anger and Vol. 03/04). Berlin, Germany: Verkehrstechnisches Institut der Deut-
aggression. American Psychologist, 45, 494 –503. schen Versicherer.
Berkowitz, L. (1993). Aggression: Its causes, consequences, and control. Kubitzki, J. (2005). Zur Problematik von Video-Rennspielen [The problem
Philadelphia: Temple University Press. of video racing games]. Zeitschrift fuer Verkehrssicherheit, 3, 135–138.
Bower, G. H. (1981). Mood and memory. American Psychologist, 36, Kubitzki, J. (2006, July). Video driving games and traffic safety—A matter
129 –148. of concern? Results from a questionnaire survey in 13–17-year-old male
Bushman, B. J. (1998). Priming effects of violent media on the accessi- adolescents. Paper presented at the 26th International Congress of Ap-
bility of aggressive constructs in memory. Personality and Social Psy- plied Psychology, Athens, Greece.
chology Bulletin, 24, 537–545. Laapotti, S., Keskinen, E., Hatakka, M., & Katila, A. (2001). Novice
Collins, A., & Loftus, E. (1975). A spreading activation theory of semantic drivers’ accidents and violations—A failure on higher or lower hierar-
memory. Psychological Review, 82, 407– 428. chical levels of driving behaviour. Accident Analysis & Prevention, 33,
Collins Cobuild advanced learner’s dictionary (4th ed.). (2003). London: 759 –769.
Collins. Scheiblechner, H. (1985). Psychometric models for speed-test construc-
de Pelsmacker, P., & Janssens, W. (in press). The effect of norms, attitudes tion: The linear exponential model. In S. E. Embretson (Ed.), Test
and habits on speeding behavior: Scale development and model building design: Developments in psychology and psychometrics (pp. 219 –244).
and estimation. Accident Analysis & Prevention. New York: Academic Press.
Dodge, K. A., & Crick, N. R. (1990). Social information-processing bases Schuhfried, G. (2006). Wiener Testsystem [Vienna test system]: Vienna
of aggressive behavior in children. Personality and Social Psychology Risk-Taking Test. Vienna, Austria: Author.
Bulletin, 16, 8 –22. Sommer, M., Arendasy, M., Schuhfried, G., & Litzenberger, M. (2005).
Elliott, M. R., Shope, J. T., Raghunathan, T. E., & Waller, P. F. (2006). Diagnostische Unterscheidbarkeit unfallfreier und mehrfach unfall-
Gender differences among young drivers in the association between belasteter Kraftfahrer mit Hilfe nichtlinearer Auswertungsmethoden [Di-
high-risk driving and substance use/environmental influences. Journal agnostic discernibility of accident-free and nonaccident-free drivers with
of Studies on Alcohol, 67, 252–260. nonlinear analysis methods]. Zeitschrift für Verkehrssicherheit, 51, 82–
Freedman, J. L. (2002). Media violence and its effect on aggression: 86.
Assessing the scientific evidence. Toronto, Ontario, Canada: University Taubman-Ben-Ari, O., & Findler, L. (2003). Reckless driving and gender:
of Toronto Press. An examination of a terror management theory explanation. Death
GameSpot. (n.d.). Burnout Revenge for PlayStation 2 [Review]. Retrieved Studies, 27, 603– 618.
August, 2006, from http://www.gamespot.com/ps2/driving/burnoutre- Trimpop, R., & Kirkcaldy, B. (1997). Personality predictors of driving
venge/review.html?q⫽burnout accidents. Personality and Individual Differences, 23, 147–152.
Gentile, D. A., Lynch, P. J., Linder, J. R., & Walsh, D. A. (2004). The Vorderer, P., & Klimmt, C. (2006). Zum Einfluss von Computerspielen mit
effects of violent video game habits on adolescent hostility, aggressive Fahrzeugbezug auf das Fahrverhalten junger Fahrer: Abschlussbericht
behaviors, and school performance. Journal of Adolescence, 27, 5–22. an die Bundesanstalt für Straßenwesen [The influence of vehicle-related
Guter, S. (2006). Risikoverhalten und Massenmedien: Der Einfluss von computer games on the driving behavior of young drivers: Final report
Risikodarstellungen in den Massenmedien auf risikobezogene Kogni- to the Federal Highway Research Institute (BASt)]. Hannover, Ger-
tionen, Emotionen und Verhaltensweisen [Risky behavior and mass many: Hochschule für Musik und Theater.
media: The impact of risky media content on risk-related cognitions, Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and vali-
emotions, and behaviors]. Taunusstein, Germany: Driesen. dation of brief measures of positive and negative affect: The PANAS
Huesmann, L. R. (1986). Psychological processes promoting the relation scales. Journal of Personality and Social Psychology, 54, 1063–1070.
between exposure to media violence and aggressive behavior by the Wikipedia. (2006). Video game controversy. Retrieved August, 2006, from
viewer. Journal of Social Issues, 42, 125–139. http://en.wikipedia.org/wiki/Video_game_controversy
Huston, A. C., Wright, J. C., Marquis, J., & Green, S. B. (1999). How Wilde, G. J. S. (1994). Target risk: Dealing with the danger of death,
young children spend their time: Television and other activities. Devel- disease and damage in everyday decisions. Toronto, Ontario, Canada:
opmental Psychology, 35, 912–925. PDE.
Jonah, B. A. (1997). Sensation seeking and risky driving: A review and Zillmann, D. (1983). Cognition– excitation interdependencies in aggressive
synthesis of the literature. Accident Analysis & Prevention, 29, 651– 665. behavior. Aggressive Behavior, 14, 51– 64.
Krahé, B., & Möller, I. (2004). Playing violent electronic games, hostile Zuckerman, M. (1994). Behavioural expressions and biosocial bases of
attributional style, and aggression-related norms in German adolescents. sensation seeking. Cambridge, England: Cambridge University Press.
Journal of Adolescence, 27, 53– 69.
Kubitzki, J. (2004). Von der Playstation auf die Autobahn: Auto-Tuning, Received June 26, 2006
illegale Auto-Rennen und reckless driving [From Playstation to German Revision received October 18, 2006
Freeway: Car-tuning, illegal car-racings, and reckless driving] (Report Accepted October 18, 2006 䡲