Respondents are recruited through partner websites, emails with online partners, refer-a-friend, trade shows, client supplied lists of customers, TV advertisements, direct mail, telephone recruitment of targeted populations
As the children were recruited from the adults who initially agreed to take the study, the sample selection bias is found primarily in the characteristics of adults who chose to take this survey. Hence, only adults were propensity weighted and the propensity weights that were generated for the adults were applied to the child. The propensity score was derived from key questions in the survey that examined the attitudes and behaviors of the respondents as well as the demographic questions.
28% report 31-1hour 24% report half hour or less 22% report 1-2 hours
Why study? A lot of games are being sold and many kids are playing them.
… and the kids are often (more than half the time) playing violent video games that reward violent behavior
Data (Anderson et al. meta analysis in 2004) have shown that exposure to violent video game play is associated with increased antisocial behavior and decreased prosocial behavior
Realism – more graphically advanced, when blood is turned on, using a light gun Identification – sex of the character is consistent with player, player can ‘see’ their character Two studies 1) A study used a level editor to create a central play area that was mirrored, so as the player ran by, they’d see themselves 2) A study found that aggression by males and females was higher when their character was the same sex than when opposite sex
Despite the substantial research on exposure to violent media, in general, and violent video games in particular, relatively little is known about the extent to which exposure to violent video games is associated with serious, moderate, and mild forms and antisocial behavior.
28% report 31-1hour 24% report half hour or less 22% report 1-2 hours
Amount of game playing increases as level of aggression increases
Last 12 months
Murder: (i.e., stabbing or shooting someone); Aggravated assult: (i.e., threatening with a weapon; attack requiring medical care) Robbery: (i.e., using a knife, gun, or other weapon to get something from someone) Sexual assault: (i.e., unwanted kissing, touching, or anything else sexual) Note, these questions were randomized for each participant to adjust for response bias due to ordering
Breaking rule: (i.e., running away from home; ditching/skipping school) Threatening / Fighting: (i.e., threatened teacher; threatened someone with knife/gun; involved with gang fight) Burglary (i.e., broken into someone house; stolen something valuable) Animal cruelty (i.e., hurt animal on purpose) Based on factor analysis Note, these questions were randomized for each participant to adjust for response bias due to ordering
Bullying: (i.e., Excluding people from group; spreading rumors about people) Physical aggression: (i.e., shoving or pushing people; getting into fights) Vandalism: (i.e., damaged someone else’s property) Dishonesty: (i.e., lying to someone to get something or a favor or get out of work) Note, these questions were randomized for each participant to adjust for response bias due to ordering
Realistic = They are JUST LIKE or SOMEWHAT LIKE real life Identification = They are JUST LIKE ME or KINDA LIKE me
Violent video game = many/most/all of the games they play have violent content Realistic = They are JUST LIKE or SOMEWHAT LIKE real life Identification = They are JUST LIKE ME or KINDA LIKE me
Violent videogame play significantly predicted all three outcome variables (p<.05) Model fits the data, though, I do not have the exact fit statistics (they would help, yes?). But clearly, the model is not explaining A LOT of the variance in the 3 outcome variables. Controlling for age, sex, and income (these variables are not shown in the model) The 3 outcome variables are also correlated, but I cannot figure out how to draw the intercorrelations: SVB & ASB: r=0.65 ASB & DB: r=0.45 SVB & DB: r=0.30
Violent video game play predicted character identification (p<.01) and character identification predicted the 3 outcome variables (p<.05) Character identification mediates the relation between violent video game play…when character identification is added, the relation between violent video game play and the 3 outcome variables disappears, but character identification is significantly associated with the 3 outcome variables. Model fits the data, though, I do not have the exact fit statistics (they would help, yes?). This model explains more of the variance in the 3 outcome variables. Additional variables in this model include age, sex, income, STAXI, and exposure to community violence (these variables are not shown in the model) The 3 outcome variables are also correlated, but I cannot figure out how to draw the intercorrelations: SVB & ASB: r=0.51 ASB & DB: r=0.23 SVB & DB: r=0.19
You don’t know the answer unless you ask – it’s a strength that we have this data and are able to analyze it’s influence Note that to be at risk, you have to have the exposure – i.e., internet use; General population findings may yield different frequencies.
The current data highlight some of these unhealthy behaviors, including sexually soliciting others, seeking out violent pornography, and other violent exposures online.
Discounting violent video game content…and how we tend to overlook certain aggressive behaviors.
Toward a better understanding of the relation between violent videogame play and different types of antisocial behavior
Toward A Better Understanding of the RelationBetween Violent Videogame Play and Different Typesof Antisocial BehaviorMerle Hamburger PhDaMichele Ybarra MPH PhDbJeffery Hall PhDaPhilip J Leaf PhDcMarie Diener-West PhDcaCenters for Disease Control; bInternet Solutions for Kids, Inc.;cJohns Hopkins School of Public HealthInternational Society for Research on Aggression,July 2008, Budapest, Hungary* Thank you for your interest in this presentation. Please note thatanalyses included herein are preliminary. More recent, finalizedanalyses may be available by contacting CiPHR for furtherinformation.
Growing up with Media(GuwM) Methodology Baseline data were collected August -September, 2006 1,588 households (one caregiver, onechild) were surveyed online Participants recruited from Harris Poll OnLine
GuwM Eligibility ADULT Be the most (or equally) knowledgeable of theyouth’s media use in the home Be a member of HPOL YOUTH Aged 10-15 years Use the Internet at least once in the last 6months English speaking
Harris Poll On Line HPOL is a double opt-in panel ofmillions of respondents. HPOL data are consistently comparableto data that has been obtained fromRDD telephone samples of generalpopulations when sampling andweighting is applied.
GuwM RR andWeighting Response rate was 26% Propensity scoring was applied Data were weighted to match the USpopulation of adults with childrenbetween the ages of 10 and 15 years
GuwM Youthcharacteristics (n=1,588) 48% Female Mean age: 12.6 years (SE: 0.05) 71% White, 13% Black, 9% Mixed, 7%Other 19% Hispanic Median time spent online on a typicalday: 31 minutes – 1 hour
Why Video Games? Video games are BIG BUSINESS ~268 million computer/video games soldin 2007 ~$9.5 BILLION in revenue (for 2007) Approximately 60% youth (8-18) playvideo games for about an hour on anygiven day
Violence in Videogames > 50% of the most popular videogames are rated ‘T’ or ‘M’ Teen/Mature rated games Almost all have violent content Most (90%) reward injuring characters Many (~69%) reward killing characters Youth (8-18) prefer ‘T’ and ‘M’ ratedgames
Exposure to ViolentVideo GamesExposure related to: IncreasedAggressive behavior,Aggressive affect, andAggressive cognitions Decreased prosocial behavior
Immersion as a MediatorA player’s sense of ‘presence’ in thegame Realism Effects more pronounced if game isrealistic Immersion Effects more pronounced if playeridentifies with characters
Problem StatementLittle is known about how exposure to violent videogames is associated with: (a) seriously violentbehavior; (b) antisocial behavior; and (c)delinquency.What is the association between playing violent videogames and concurrent reports of externalizing behavior;To what extent does immersion mediate thisassociation?
Characteristics of gamevideo players N=1,493 (video game players) 48% Female Mean age: 12.5 years (SE: 0.04) 79% White, 13% Black, 8% Other 12% Hispanic Median HH income: $50,000-$74,999
Game Playing Behavior Median # of days / week: 3-4 Median time playing/ day: 31-60 min Overall median exposure: 157 min / week Median exposure by violent video game None: 67.5 min / week Some: 157.5 min / week Many / Most / All: 287.8 min / week
Violent Video Game Play“When you play video, computer, orInternet games, how many showphysical fighting, shooting, orkilling?”Response alternatives:‘None’; ‘Some’; ‘Many’; ‘Most/All’
Outcome Variables… Seriously Violent Behavior Behavior likely resulting in murder Aggravated assault; Robbery; Sexual assault Alpha = 0.87
SummaryPlaying violent video games is common. Over a quarter of respondents reportplaying violent video games Weekly exposure significantly related toplaying violent video games
SummaryConsistent with previous literaturereporting associations between violentvideo games and aggression.. Frequent exposure to violent video gamesis concurrently associated with seriousexternalizing behaviors, Character identification is a mediator
Limitations ofGuwM Data Data are cross-sectional Reliance on self-reports It is possible that: Children were monitored by their parents 22% of youth indicated someone was closeenough to see the screen during data collection Parents completed the youth survey.
Implications Need to educate caregivers about thegrowing evidence for the impact ofviolent media / video games on seriousexternalizing behaviors Character identification appears to beimportant, over and above the violencein the game itself.
Contact InformationDr. Merle HamburgerCenters for Disease Control and Preventionmhamburger@cdc.govThis research was supported by Cooperative Agreement numberU49/CE000206 from the Centers for Disease Control and Prevention(CDC). The findings and conclusions in this presentation are those ofthe authors and do not necessarily represent the official position of theCenters for Disease Control.