While the virtual environments of online games can foster healthy relationships and strong communities, some online games are also marred by antisocial and offensive behavior. Such behavior, even when relatively rare, influences the interactions and relationships of users in online communities. Thus, understanding the prevalence and nature of antisocial and offensive behaviors in online games is an important step toward understanding the full spectrum of healthy and unhealthy interactions and relationships in virtual environments. Extensive research has explored video game content produced by game developers, such as violence, profanity, and sexualized portrayals, but much less research has systematically examined potentially problematic content produced by players in online games. While potential effects of antisocial and offensive online game content are not well understood, a first step toward exploring this concern is systematic documentation of offensive usergenerated content in online games. To that end, two large-scale content analyses measured a range of 1 This paper contains strong language which may be offensive to some readers. offensive user-generated content, including utterances, text, and images, from a total of more than 2,500 users in popular first-person shooter video games. Findings indicated that some content, such as profanity, was frequent among users who spoke during games. More offensive and potentially harmful content, such as racial slurs, was proportionally very rare, but frequent enough to be encountered often by regular players. Results of this initial investigation should be interpreted tentatively, do not suggest that relationships in online shooter games lack healthy elements, and should not be generalized to other online game communities until further research is conducted.User names. For player names, coded variables assessed presence or absence of sexual references, references to drugs and alcohol, overall profanity, "seven dirty words" and other strong profanity, mild profanity, and slurs related to race, gender, religion, or sexual orientation.
User emblems.Coded variables for customized player emblems assessed images including the same variables as for player names, but with sexual reference categories including overall sexual references, mild sexual references, and explicit sexual references, and with a coded category for insults also included.Coding and reliability. For utterance variables, the unit of analysis for the study was the online speech from each of the 1,866 individual players appearing in the study, and the unit of enumeration was the total number of utterances by each player meeting the criteria for each of the measured utterance-related variables. For user name and emblem variables, the unit of analysis was the presence or absence of content in each of the 1,866 individual players' user names and emblems.The matches randomly assigned to the two coders included an overlap of 59 matches (yielding 633 individual player cases) for assessment of intercoder reliabil...