Violence in Teen-Rated Video Games

Kevin Haninger; M. Seamus Ryan; Kimberly M. Thompson, MS, ScD

In This Article


We identified 396 T-rated video game titles released on the major video game consoles in the United States by April 1, 2001, which at that time included Nintendo 64 (N64), Sega Dreamcast (DC), Sony PlayStation (PS), and Sony PlayStation 2 (PS2).[14] Using several video game Web sites,[12,15,16,17] we verified the game titles and ESRB content descriptors, and we classified each game title by 1 of 10 primary genres (ie, action, adventure, fighting, racing, role-playing, shooting, simulation, sports, strategy, trivia, or other.) We stratified the 396 video game titles by genre and randomly selected 20% (n = 80) of the game titles to play, but we played 81 games because one game title (Final Fantasy Anthology) included 2 separate games (Final Fantasy V and Final Fantasy VI).

Our random sample did not include video games for Microsoft Xbox (Xbox) and Nintendo GameCube (GCN) because these consoles entered the US market after we generated our random sample and began collecting data. However, in the interest of including all current consoles, we played 9 additional video games on Xbox (n = 5) and GCN (n = 4). Overall, we played 90 T-rated video games.

For consistency, one author with considerable video gaming experience (M.S.R.) played the entire random sample of video games (n = 81) and recorded all game play on videocassettes for later coding.[14] Two additional students played the video games for Xbox and GCN (n = 9). We purchased or rented all video games with their original game manuals intact, then the game player read the manual and played each game for several hours to become familiar with the game features. Following this familiarization, the player restarted the video game from the beginning and recorded at least 1 hour of game play on an individual videocassette. We recorded all game content from the moment the player turned the console on, including any introductions and setup. This step differed from our prior study of violence in E-rated video games,[13] in which we started recording from the first scene when autonomous movement began. We observed in that study that game introductions sometimes depicted content that could warrant an ESRB content descriptor, and that game setup often allowed the player to select weapons. Consequently, in this study, we included those elements in our hour of game play. Our method of playing each game for at least 1 hour yields approximately the same amount of game play for each game and allows us to play a larger number of games while still obtaining a reasonably representative sample of game play.[14] However, it captures only a small amount of the game play for those games designed for much longer play. For example, we may miss some content in those games that become more difficult as the player advances. We may also introduce some bias if we consistently miss more aggressive, faster, and more violent game play or new types of weapons introduced later in the game.

With the game play recorded on videocassettes, one author with considerable video gaming and prior coding experience (K.H.) reviewed and manually coded the content related to the depiction of violence and blood using standard coding instruments (available upon request), then entered the data into a database constructed with Microsoft Access (Version 2002, Microsoft Corporation, Redmond, Washington).[14] All authors discussed all parts of games that presented difficulty in coding.

Recognizing the importance of consistent definitions, we defined violence as intentional acts in which the aggressor causes or attempts to cause physical injury or death to another character. We did not include unintentional actions that led to physical harm, the effects of natural disasters, or the presence of dangerous obstacles not attributed to a particular character. We defined characters broadly, including humans and nonhumans (eg, monsters, animals, and personified robots) that attacked the player or other characters. We did not code as violence intentional acts of physical force that represent normal play in a sports game (eg, tacking in football), because the intention is to stop the opposing player without necessarily causing injury. We did code all punches and kicks as violence, including those in boxing and wrestling games, in which the intention is to cause injury.

To quantify the amount of violence, we divided the videocassettes into 1-second intervals and noted whether each second of game play contained acts of violence. We used a similar coding method in an earlier study in which we obtained good agreement between coders.[13] However, to assess the consistency of our method between coders in this study, a research assistant with considerable video gaming experience but with no other involvement in this study independently coded a randomly selected subset of 12 video games (ie., 2 video games were used for training and 10 video games were used for comparison with a kappa statistic). In addition to coding whether each second of game play involved a character committing acts of violence, we further coded for both precursor acts of planning violence (ie, selecting a weapon or aiming prior to committing violence) and depictions of injury after violence (eg, a character lying on the ground dead or wounded). Thus, for each second of game play not already coded as "committing violence," we coded the second as "planning violence" if it showed a character planning acts of violence, as "depicting injuries" if it depicted injuries from violence, or as "not violent" if it failed to meet either of these criteria.

To further characterize the overall portrayal of violence, we used categories to rate how the video game depicted injuries to human and nonhuman characters. We designated the violence as relatively "mild" when the most severe depiction of violence in the game involved minor auditory or visual representations of injury and pain that primarily served to notify the player that a character was injured (eg, characters grunted or turned red when injured, but did not scream or bleed). We designated the violence as relatively "moderate" when the most severe depiction involved more realistic representations of injury and pain (eg, characters screamed, fell over, or bled when injured). We designated the violence as "strong" when the most severe depiction involved graphic representations of injury and pain that exaggerated or focused attention on suffering (eg, characters screamed in agony or bled excessively when injured or when otherwise physically tortured). For each video game, we distinguished between human and nonhuman characters and coded the highest (and most severe) category we observed. Given this new aspect of our coding method, the author who played the games (M.S.R.) coded a randomly selected subset of 20 video games for comparison with the categories assigned by the author who coded all of the games (K.H.). We calculated intercoder reliability using the kappa statistic.[18]

For each video game, we noted the types of weapons used for violence, whether the player could select these weapons, whether violence resulted in injury, whether injuring or killing human and nonhuman characters was rewarded or was required to advance in the game, the number of human and nonhuman deaths from violence, and the depiction of blood. We also noted whether destroying objects was rewarded or was required to advance in the game, although we did not code the destruction of objects as violence.

Finally, following the release of the R-rated film The Matrix: Reloaded associated with same-day release of the T-rated video game Enter the Matrix (Xbox),[19] we included this video game as one of 9 additional games we played after we drew our random sample, and we purchased and coded the R-rated films The Matrix and The Matrix: Reloaded using the same coding scheme. We also quantified the content not related to violence and blood that could motivate the T-rating in the 9 additional games we played on Xbox and GCN, using the same definitions and methods as in a prior study.[14] We performed statistical tests using SAS (Version 8.2 for Windows, SAS Institute Inc., Cary, North Carolina).


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.