VIOLENT VIDEO GAMES AND DELINQUENT BEHAVIOR 1
This article is a post-print version.
Violent Video Games and Delinquent Behavior in Adolescents:a Risk Factor Perspective
Liese Exelmans, MA
Kathleen Custers, PhD
Jan Van den Bulck, PhD, DSc
Leuven School for Mass Communication Research, University of Leuven, Belgium
Article accepted for publication in Aggressive Behavior.
Please cite as:
Exelmans, L., Custers, K., Van den Bulck, J. (2015). Violent video games and delinquent behavior in adolescents: a risk factor perspective. Aggressive Behavior, 41 (3), 267-279
Correspondence concerning this article should be addressed to:
Liese Exelmans, Leuven School for Mass Communication Research, KU Leuven, Parkstraat 45, 3000 Leuven, Belgium.
E-mail:
Telephone: 003216323231
Abstract
Over the years, criminological research has identified a number of risk factors that contribute to the development of aggressive and delinquent behavior. Although studies have identified media violence in general and violent video gaming in particular as significant predictors of aggressive behavior, exposure to violent video games has been largely omitted from the risk factor literature on delinquent behavior. This cross-sectional study therefore investigates the relationship between violent video game play and adolescents’ delinquent behavior using a risk factor approach. An online survey was completed by 3372 Flemish adolescents, aged 12 to 18 years old. Data were analyzed by means of negative binomial regression modelling. Results indicated a significant contribution of violent videogames in delinquent behavior over and beyond multiple known risk variables(peer delinquency, sensation seeking, prior victimization, alienation). Moreover, the final model that incorporated the gaming genres proved to be significantly better than the model without the gaming genres. Results provided support for a cumulative and multiplicative risk model for delinquent behavior.
Keywords: adolescents, video gaming, delinquency, violence, risk factor
Violent Video Games and Delinquent Behavior in Adolescents:a Risk Factor Perspective
From the 1990s onwards, criminological studies have beeninfluenced increasingly by the risk factor prevention paradigm. Borrowing insights from medical science and public health, the central idea in this paradigm is to identify the risk factors for delinquent behavior in order to design suitable prevention measures (Farrington, 2000; Haines & Case, 2008). Each risk factor contributes to a greater likelihood of engaging in antisocial behavior, and can interact in an additive or multiplicative way (Farrington, 2000; Gentile & Bushman, 2012). Over the years, a significant number of studies have contributed to the identification of these risk factors(see Shader, 2001). Media violence exposure, however,has been largely overlookedin this risk matrix (Boxer, Huesmann, Bushman, O’Brien, & Moceri, 2009; Gentile & Bushman, 2012). According to Boxer et al., (2009), the absence of media violence in the risk factor literature is the result of methodological inconsistencies and the overuse of normative community samples that dominate the research on media violence effects. Still, the omission of media violence as a risk factor seems surprising given recent meta-analyses that have identified media violence exposure as a causal factor for aggressive behavior (see Anderson & Bushman, 2010; Bushman & Huesmann, 2006). The studies that have included exposure to media violence as a risk factor for aggressive and antisocial behavior, have mainlyused a general measure that integrated exposure to violent television, movies and video games(Boxer et al., 2009; Gentile & Bushman, 2012).
The examination of violent video games as a separate risk factor seems worthwhile, given (1) the substantial research effort that has been devoted to the potentially harmful effects of violent video games on adolescents’ aggressive behavior, (2) the observed effect sizes in these studies (Anderson & Bushman, 2010; Anderson & Bushman, 2002; Anderson & Dill, 2000; Bushman & Huesmann, 2006), (3) the call to look at violent video gaming as part of an additive risk model (Anderson, Gentile & Buckley, 2007), and (4) the distinguishing interactive nature of video games (i.e. the user who is an observer and an enactor of aggression simultaneously) that may enhance the acquisition of antisocial beliefs and behavior(Anderson et al., 2008; Glaubke, Miller, Parker, & Espejo, 2001; Vorderer, 2000). While some studies (Boxer et al., 2009; Gentile et al., 2012; Anderson et al., 2007) have paved the way for this approach, there is ample room for expanding the knowledge on video game violence in the risk factor paradigm.
Thegoal of the current study therefore is to examine how the likelihood of delinquent behavior is affected by cumulative risk factors, where we considerviolent video gamingto be a significant risk factor within a larger risk matrix. To analyze the impact of violent games in the context of other sorts of games, a comparison was made with non-violent gaming.
Video Gaming and Aggressive Behavior
The consumption of video games is at its peak during adolescence. American youth (8-18 years old) play approximately 1.15 hours of videogames per day. Average time spent playing videogames has continuously increased over the past years: from 24 minutes per day in 1999, to 49 minutes in 2004 and 73 minutes in 2009 (Rideout, Foehr, & Roberts, 2010). In Europe, earlier studies found that Belgian adolescents played an average of 1.20 hours of video games per day(Vandercammen & Vandenbrande, 2011), whereas their Dutch counterparts played about 1.61 hours per day (Lemmens, Valkenburg, & Peter, 2011). Content analyses have shown that violence is a key element in video games: Almost 90% of the games contain some violent content, and 40% of the games include serious violence against other characters (Glaubke et al., 2001). These statisticshave fuelled a lot of discussionabout the potential harmful effects of violent video games on their players (Anderson & Dill, 2000; Gentile, Lynch, Linder, & Walsh, 2004; Savage, 2004)
The majority of the studies on violent videogame effects has focused on the question of whether game play increases aggressive behavior and aggression-related thoughts(Anderson & Bushman, 2002; Anderson et al., 2010; Gentile et al., 2004). For example, in a meta-analysis of 136 studies, Anderson et al. (2010) found that exposure to video game violence was associated with higher levels of (1) aggressive behavior, (2) aggressive cognitions, (3) aggressive affect, (4) desensitization and (5) decreases in prosocial behavior.
The General Aggression Model (GAM), which providesa comprehensivesynthesis of prior theoretical models, is commonly referred to for explaining these relationships(Anderson & Bushman, 2001; Anderson & Groves, 2013; Carnagey & Anderson, 2003). The GAM is a social cognitive model that describes the short and long term effects of exposure to media violence on aggressive behavior (Anderson & Groves, 2013; Dewall & Anderson, 2005). For the short term effects or single episode cycle, the GAM postulates a three-stage process wherebysituations (i.e. provocation, unpleasant environment) and person factors (i.e. traits, values, attitudes) indirectly influence the likelihood of aggressive behavior through thoughts, feelings or physiological arousal. For example, in a first stage, trait aggressiveness and exposure to violent media content caninteractively influence the accessibility of aggressive thoughts and emotions, and arousal in a second stage (Anderson & Bushman, 2001). In a final stage, an appraisal and decision process takes place that results in either thoughtful (reappraisal) or impulsive (immediate appraisal) aggressive behavior. The resulting behavior then becomes part of the input in the next episode cycle(Anderson & Groves, 2013; Bushman & Anderson, 2002; Carnagey & Anderson, 2003; DeWall, Anderson, & Bushman, 2011; Dewall & Anderson, 2005). For the long term effects, the GAM specifies that chronic exposure to aggression-related stimuli such as media violence results in the development of a more aggressive personality. By changing an individual’s attitudes and beliefs and their beliefs aboutother’s behavior, an aggressive personality develops. In other words, individuals who repeatedly experience situations that provoke aggressive thoughts gradually develop and reinforce aggression-related knowledge structures. The repeated activation of these knowledge structures results in a higher accessibility and thus higher likelihood of them being used in social situations. In particular, repeated exposure to media violence may result in decreased attention to violence, a more positive attitude towards violence and increased antisocial behaviors such as aggression and delinquency (Anderson & Bushman, 2001; Anderson & Bushman, 2002; Anderson & Dill, 2000; Bushman & Anderson, 2002).
There is a large body of research in support of the propositions advanced by the GAM. Studies support the idea of both immediate and long term effects (Anderson & Bushman, 2001; 2002; Anderson & Dill, 2000; Anderson et al., 2010; Markey & Scherer, 2009), in experimental, correlational and longitudinal studies (Anderson & Dill, 2000; Bushman & Anderson, 2002; Holtz & Appel, 2011; Morris & Johnson, 2010) and across demographics such as gender, age and social status (Anderson & Bushman, 2001; Barlett, Anderson, & Swing, 2008; Delisi, Vaughn, Gentile, Anderson, & Shook, 2013). It has been argued that there is sufficient evidence to support the idea that the relationship is causal (Anderson et al., 2010; Bushman & Anderson, 2001; Huesmann, 2007).
However, with other authors arguing that their studies yielded null effects(Ferguson et al., 2008; Ferguson, Olson, Kutner, & Warner, 2010; Funk et al., 2002; Savage & Yancey, 2008), the relationship between violent video gamesand aggression has fostered some debate. Criticism is raised on (1) the methodological approaches used to illustrate a relationship, (2) the existence of a longitudinal or causal effect, (3) the generalizability of results to other populations such as older children or teenagersand (4) the possibility of spurious relationships caused by confounding variables such as family violence and trait aggression (Browne & Hamilton-Giachritsis, 2005; Ferguson, San Miguel, Garza, & Jerabeck, 2012; Ferguson & Savage, 2012; Freedman, 2002).
Regardless of this debate, there is ample support for an approach focusing on theindividual susceptibility to the effects of media violence (Browne & Hamilton-Giachritsis, 2005; Valkenburg & Peter, 2013). A number of scholars have argued that particular types of players may be more prone to violent video game effects than others (Gentile et al., 2004; Markey & Markey, 2010; Markey & Scherer, 2009; Slater, Henry, Swaim, & Anderson, 2003). In other words,user characteristics such as gender, personality traits or family and peerenvironmentmay moderate the association between violent video games and aggressive and delinquent behavior. Several factors that mitigate the effect of violent media content on aggressive behavior have thus far been identified: (1) user characteristics such as gender, sensation seeking, psychoticism and aggressiveness(Bartholow & Anderson, 2002; Bushman & Anderson, 2002; Markey & Scherer, 2009; Slater, 2004), (2) aspects of their environment such as family violence and parent child communication (Fikkers, Piotrowski, Weeda, Vossen, & Valkenburg, 2013; Wallenius & Punamäki, 2008)and (3) content characteristics such as realism and consequences of the violence depicted(Huesmann & Eron, 1986; Huesmann & Taylor, 2006). Additionally, it has been argued that there is no evidence that any subgroup is completely immune to the effects of media violence, nor that there is a moderator that provides such immunity. In other words, while some variables might put some individuals more atrisk, in the end, everyone is somewhat at risk of behaving aggressively after exposure to media violence(Anderson et al., 2003; Anderson, Gentile, & Buckley; 2007).
A Risk Factor Perspective
Other scholars have arguedthat researchers should look beyond the individual moderation effects and to consider media violence as part of a cumulative risk model(Anderson et al., 2007; Boxer et al., 2009; Gentile & Sesma, 2003):“No single risk factor dominates an individual’s overall risk for behaving violently in the future, but the presence of multiple risk factors and the absence of resilience factors add up to a fairly accurate probabilistic prediction of future aggressive behavior” (Anderson et al., 2007, p. 138). Contrary to focusing on the predictive value of each factor independently, such an approach postulates that the total number of risk and resilience factors is more accurate in predicting the outcome (Farrington, 2000; Gentile & Bushman, 2012). Some of these studies have also noted a multiplicative effect when several risk factors are present (Gentile & Bushman, 2012; Herrenkohl et al., 2000).
This approach originates from the public health domain and has been coined the risk factor prevention paradigm(Farrington, 2000). A risk factor has been generally defined as a factor that predicts an increased probability of a negative outcome (Kazdin, Kraemer, Kessler, Kupfer, & Offord, 1997 p.377). However, there has been some inconsistency with regard to the operational definitions of a risk factor: it can refer to an extreme category of an explanatory variable (for instance, poor parental control), a dichotomous variable (poor vs. good parental control) or a continuous explanatory variable (a scale of parental control from poor to good) (Farrington, 2000, p. 3; Haines & Case, 2008; p. 7). In short, the paradigm is aimed at identifying the risk factors for a certain negative outcome, and to implement targeted strategies to prevent it (Farrington, 2000). The idea is often illustrated with the example of how doctors assess a patient’s risk of having a heart attack by checking off identified risk factors such as family history, diet, exercise level, and so on. By determining the presence of risk factors and protective factors, the likelihood of having a heart attack is estimated. The same approach has been applied to assessing an individual’s likelihood of engaging in delinquent behavior. It was first applied in criminal career research (Farrington, 2000; Loeber, 1990). The studies that have used such a risk factor paradigm emphasize the multicausal nature of aggression and delinquent behavior.
Over time, criminological research has extensively expanded the knowledge on risk factors for aggressive and anti-social behavior (see Shader, 2001 for overview), but has generally neglected exposure to media violence in this context. This is surprising, given that media effects studies have repeatedly demonstrated that exposure to media violence in general and violent video games in particularis significantly related to heightened aggression. With regard to video games, countless studies have examined violent video gaming as a predictor of aggressive behavior, which has amounted to observed effect sizes that are comparable or larger than those of well-established health risks such as the effect of second-hand smoking on lung cancer (Anderson, 2004; Huesmann & Taylor, 2006).
The use of a risk factor approach in media effects studies (i.e., to consider media violence exposure as part of a larger risk model for aggression) has gained more attention in recent years with studies that used various risk factors either as control variables or in an explicit risk factor approach. For example, in three studies, Anderson, Gentile, and Buckley (2007) examined the relationship between violent video games and aggression in the presence of other known risk factors, such as hostile attribution bias, gender and adult involvement in media habits. Overall, they highlighted the importance of situating violent video games as a risk factor in a developmental risk and resilience approach. A recent study by Delisi and colleagues (2013) tested the risk factor approach in a sample of juvenile delinquents. They found that violent video gaming was associatedwith delinquency in a clinical sample, after taking into account multiple correlates of juvenile delinquency (i.e.,gender, race, age, delinquency history, psychopathic trait and years playing video games). Finally, other studies have used a composite score of exposure to violent content on different media outlets in a risk factor approach, and have concluded that media violence exposure contributes significantly to the prediction of aggressive behavior (Boxer et al., 2009; Gentile & Bushman 2012; Graber, Nichols, Lynne, Brooks-Gunn, Botvin, 2006). However, it has been argued that the use of composite measures might “obscure the distinct purposes that different types of violent media content, particularly as found in interactive media, may serve for different adolescents” (Slater, 2003 p.113). Additionally, it has been suggestedthat, compared to more passive media such as television and movies, the interactive nature of video games may increase the likelihood of aggressive behavior by creating the unique experience for players of engaging in behavior that is not allowed in the offline world (Anderson et al., 2008; Glaubke et al., 2001; Vorderer, 2000). Therefore it is important to include violent video games as a separate risk factor in the risk factor paradigm: this is the central aim of this study.
Present Study
The current study examines video gaming as part of a risk model where we expect a significant contribution of the gaming risk factor over and above the other risks for predicting delinquent behavior. The relationship between video games and individual delinquent behaviorwill be examined using a multifactorial approach that considers media (violent and non-violent video gaming), individual(gender, age, sensation seeking, aggressiveness, alienation), and social (peer delinquency) influences. In addition, we consider the contribution of violent video games in the context of other types of games, more particularly non-violent games. In this way, it is possible to examine the differential impact of violent video games in predicting delinquent behavior. We expect that violent video gaming plays a significant role in predicting delinquent behavior when applying the risk factor paradigm and that a model that includes violent video gaming as a risk factor will perform significantly better in predicting delinquent behavior than the model without. In short, this paper will examine the following research question:
RQ 1: Is violent video gaming a significant risk factor within a larger risk matrix for adolescents’ delinquent behavior?
In total, we will take into account five potential risk factors (sensation seeking, prior victimization, alienation, peer delinquency, video gaming), in predicting adolescent delinquent behavior, while controlling for participants’ gender, age and overall gaming volume. These risk factors have all been identified as significant predictors of delinquent behavior in previous literature.
Gender & Age
Studies haveshown that boys played video games significantly more often and longer than girls did (Olson et al., 2007; Vandercammen & Vandenbrande, 2011)and were more attracted to violent video games (Lemmens et al., 2011; Olson et al., 2007). Moreover, boys generally display more aggression (Archer, 2004; Card, Stucky, Sawalani, Little, 2008) and engage in more delinquent behavior than girls do(Steffensmeier & Allan, 1996). With regard to age, it has been found that media consumption in general and the use of video games in particular increases during adolescence(Rideout et al., 2010). Additionally, the engagement in delinquent behavior also follows a developmental path: During adolescence, delinquent behavior increases, whereas aggression decreases(Barnow, Lucht, & Freyberger, 2005; de Haan, Prinzie, & Deković, 2010; Stanger, Achenbach, & Verhulst, 1997).