Abstract
Prior studies have mainly focused on the controversial issue of whether violent video games lead to aggressive behavior in real life. However, data regarding antisocial behavior while playing online are still scarce. In this study, we examined the relationship between toxicity (a form of verbal aggressive behavior directed against other players) in multiplayer online video games and several potential predictors such as personality traits, emotion reactivity, and motivations to play. A large-scale survey (n = 816) was conducted among French-speaking adult gamers, assessing video game habits, impulsivity, empathy, emotion reactivity, and motivations to play (i.e., socialization, achievement, immersion). Results showed that younger age, being male, spending a lot of time playing per week, and being highly achieving increased the likelihood of reporting toxicity and change of behavior in game. High emotional reactivity and being high in two dimensions of impulsivity (negative urgency and sensation seeking) increased the likelihood of toxic behavior. In contrast, individuals who are more empathic tended to report nontoxic behavior and individuals who play to socialize reported little change in behavior in game. Future research should address the complexity of behaviors displayed during online games—notably through experimental studies allowing direct observation of gamers while they play—and the functional connection between toxicity and several psychological, emotional, and motivational dimensions in further detail. Scientific research would also benefit from the creation and validation of a scale measuring in game, the social and antisocial behaviors.
Introduction
Video gaming and the “violence” issue
Over the last decades, video games have become one of the most popular forms of entertainment, with >2.5 billion active gamers worldwide in 2020. But this trend has been accompanied by two major debates among the scientific community. Most of the literature has focused on the effects of video games on gamers through two scopes: pathological gaming and video game violence. Many popular games either depict violence in different forms (scenes of violence, provocative or offensive contents) or the game play involves violent acts, such as killing enemies, to move forward in the main scenario.
The California Civil Code defines a “violent video game” as “a video game in which the range of options available to a player includes killing, maiming, dismembering, or sexually assaulting an image of a human being”. 1 However, drawing the line between what is considered as violence in a video game and what is not can be tricky. While popular First-Person Shooter games such as Call of Duty, Overwatch, and the Counter Strike franchise are inevitably categorized as violent video games, the line is harder to draw when it comes to other types of games. For example, in multiplayer online battle arena (MOBA), a game genre popularized worldwide by Dota-2 and League of Legends in which the objective is to defeat one's enemies, the graphic representation of violence is less pronounced due to the fact that the game takes place in a fantasy world.
Forms of aggressive behaviors in multiplayer online games
Since the controversial issue over violent video games surfaced in the 1990s, the vast majority of scientific literature have been addressing whether or not violent games lead to aggressive behavior in real life. Thus far, scientific findings are inconsistent on that topic and no consensus has been reached regarding the influence of such games. Research has repeatedly concluded that violent video games increase the frequency of aggressive thoughts and sometimes of aggressive responses.2–6 However, recent studies have disputed this stance, arguing that the relation of game play and real-world aggressive behavior is at best overstated and at worst spurious.7–9 It appears that individuals adversely affected by violent video games have preexisting dispositions such as specific personality traits, notably psychoticism and aggressiveness. 10 There is also evidence for publication bias in video game violence effects literature, 11 and methodological concerns were raised.12,13
Yet, focusing on this debate have left aside others aspects of aggressive behaviors that occur during online games. Social interactions are an essential feature in multiplayer online games (MOGs). However, interactions between players can sometimes lead to the emergence of undesired and unintended behavior. 14 Video games, such as League of Legends, Call of Duty or Counter Strike: Global Offensive, are notorious among gamers for having highly “toxic” communities. 15 The term “toxicity” refers to perceived hostility between players in MOGs. Those antisocial behaviors can be categorized in two types: “flaming” and “griefing.” Flaming involves harassing and verbally assaulting other players. Verbal toxicity entails trash talk, sexism, racial slurs, and homophobia. This phenomenon has mostly been studied through linguistic analysis to assess communication patterns in online games.14,16 On the other hand, the practice referred to as “griefing” 17 by video gamers consists in counterplaying to sabotage games and do things to intentionally ruin the experience for other players. In the end, both of those pervasive antisocial behaviors break established norms and unsettle the wellbeing of online communities.
Explaining antisocial behaviors in MOGs
Several moderator effects help to explain why some gamers demonstrate antisocial behaviors when others do not. Although exposure to violent games seems to increase short-term aggression for women as well as for men, 18 there might be a potential sex difference. Playing violent video games apparently results in more aggressiveness for men than women.2,19 Nonetheless, the population of female players have received relatively little attention in scientific literature—except on the topics of gender representation and sexism in video games—leading to a lack of evidence to support the previous hypothesis.
Then, the way people behave in MOGs strongly relies on their goals and their motivations to play. Playing violent video games to vent anger significantly reduces the relative accessibility of aggression. 20 Moreover, the mode of play is what changes the participants' behavior during a game. 21 The social context of game play influence behavior more than the content of the game that is played: people who play violent video games cooperatively engage in more prosocial and cooperative behaviors than those who play competitively. 22 Thus, playing cooperatively appears to be associated with less aggressive behavior, whether games are violent or not.23,24
Computer-mediated communication specificities influence online behaviors. Communicating behind a screen involves anonymity and invisibility. A previous study demonstrated that the lack of eye contact is the chief contributor to the negative effects of online disinhibition. 25 As empathy is the ability to understand and share feelings of another, one could expect that high levels of empathy would decrease aggressive behaviors driven by the lack of eye contact online. Although research suggested that the lack of nonverbal cues in online world contributes to overall lower levels of virtual empathy compared with the real world, 26 greater levels of empathy do foster prosocial behavior and reduce aggression even in the context of violent video gaming. 27
Last but not least, a number of investigators have demonstrated a relationship between impulsivity and aggression in general. In the context of violent video gaming, results indicated that participants who played violent video game with aggressive content exhibited significantly interpersonal aggression after they stopped playing. 28
Aim of the study
In brief, past research provided evidence that motivations to play, the lack of eye-contact, and some specific emotion-related and psychological traits were correlated to aggressive behaviors and they influenced positively or negatively the way people behave after having played violent video games. However, very few studies have focused on how gamers behave toward each other while playing online games, knowing how important communication is in those video games and that toxic behaviors while playing can have pernicious effects on individuals. The present study aims at exploring if empathy, impulsivity, emotion reactivity, and motivations to play may influence antisocial communication between players in MOGs, whether the games are categorized as violent or not.
Materials and Methods
Sample and procedure
In this study, we analyzed data from 816 French-speaking online gamers. Of those participants, 109 were female and 707 were male. Participant ages ranged from 18 to 54 (M = 22.81; SD = 4.81). Data were collected through an online survey. Links to the survey were put on various French-speaking online forums, popular online game sites, and several social media groups dedicated to video gaming. To achieve heterogeneity of the sample, no exclusion criteria other than being at least 18, understanding French, and playing online games were specified. Data were collected from June 2017 to October 2018, with the approval of the Comité d'éthique pour la recherche en santé (no. 2017-24), a French Ethics Committee specialized in health and social science research.
Materials
Participants provided their responses to a number of basic demographic information, including their age and gender. Gaming habits were explored through several items: number of hours played per week (from “0–9 hours” to “>60 hours”), frequency at which they played competitive games (from “never” to “always”), alteration of behavior when they played competitive games (from “not at all” to “totally”), and self-reported assessment of displaying regular “toxic” behaviors in game (yes/no), for which several examples were given to the participants. The survey also included four scales, listed below.
Online gaming motivations scale
Motivations to play were explored with the online gaming motivations scale (OGMS). The OGMS is a 39-item survey measuring online gaming motivations on a three-factor model. 29 The three factors correspond to Achievement, Social, and Immersion motivations. The instrument also assesses 10 underlying components. The first factor, “achievement,” is composed of advancement, mechanics, and competition motivations. The second factor, “social,” regroups socializing, relationship, and teamwork motivations. The third factor, “immersion,” encompasses discovery, role-playing, customization, and escapism motivations. Responses are rated on a 5-point Likert scale.
Emotion reactivity scale
Emotion reactivity was assessed with the emotion reactivity scale (ERS). The ERS is a 21-item measure designed to assess individuals' experience of emotion reactivity. 30 The items inquire about three aspects of emotion reactivity: sensitivity, arousal/intensity, and persistence. Each item is rated on 5-point Likert scale ranging from 0 (“not at all like me”) to 4 (“completely like me”). We used the French version of the scale. 31
Urgency, premeditation, perseverance, sensation scale
Impulsivity was evaluated with the short version of the urgency, premeditation, perseverance, sensation (UPPS) scale. The UPPS-P 32 is a 59-item scale measuring five impulsivity components: negative urgency, positive urgency, lack of premeditation, lack of perseverance, and sensation seeking. For the study, we used a validated short, 20-item French version of the UPPS-P. 33 All items are scored on a 4-point Likert scale ranging from 1 (“I agree strongly”) to 4 (“I disagree strongly”).
Eysenck's impulsivity inventory
Empathy was measured by using the subscale “empathy” from Eysenck's Impulsivity Inventory. 34 We used the French version of the scale, 35 but we decided not to use the full Inventory, which also measures Impulsiveness and Venturesomeness, as we already had a measure of several impulsivity components with the UPPS. The subscale “empathy” is composed of 19 items. Responses are dichotomous (“yes” or “no”).
Results
Gender, age, motivations to play (achievement, socialization, immersion), impulsivity (negative urgency, positive urgency, lack of premeditation, lack of premeditation, lack of perseverance, and sensation seeking), empathy, emotion reactivity, and total played time per week scores were entered in a binomial logistic regression model with “toxic behavior” as dependent variable (i.e., 1 = yes, 2 = no, Model 1). We then conducted a linear regression model with “behavior change while playing competitive games” as dependent variable (i.e., Model 2) and the same predictors as in Model 1. All scales had acceptable to excellent reliability, with all reliability coefficients >0.70 (Table 1). Detailed results and estimates can be found in Tables 2–5.
Scales Reliability Statistics
Model 1: Odd Ratios
Model 1: Likelihood of Displaying ‘‘Nontoxic’’ Versus ‘‘Toxic’’ Behavior While Playing
SE, standard error.
Model 2: Odd Ratios
Model 2: Likelihood to Change Behavior While Playing Competitive Games
Model 1: Toxic behavior
We conducted a binomial logistic regression analysis with toxicity as the dependent variable, such that a higher odds ratio indicates a higher probability not to be toxic. Older gamers were less prone to being toxic, χ 2 (1) = 4.19, odds ratio (OR) = 1.04, p = 0.05, as well as female gamers compared with male gamers, χ 2 (1) = 20.82, OR = 4.67, p < 0.001. Gamers who reported playing for longer time per week were more likely to report being toxic than gamers who played less time, χ 2 (1) = 4.46, OR = 0.89, p = 0.035. We found a significant effect of empathy χ 2 (1) = 9.34, OR = 1.11, p = 0.002, showing that the higher gamers are empathic, the less they are likely to be toxic. We also found a significant effect of emotion reactivity, χ 2 (1) = 18.67, OR = 0.97, p < 0.001, as well as achievement, χ 2 (1) = 39.48, OR = 0.94, p < 0.001, showing that highly reactive and highly achieving gamers are more likely to engage in toxic behaviors. Lastly, results indicated a significant effect of two dimensions of impulsivity: negative urgency, χ 2 (1) = 4.61, OR = 0.92, p = 0.032, and sensation seeking, χ 2 (1) = 3.99, OR = 0.94, p = 0.047, showing that gamers high in negative urgency and in sensation seeking are more likely to be toxic than gamers who have low scores in those two dimensions.
Model 2: Behavior change while playing competitive games
We conducted a linear regression analysis with behavior change while playing as the dependent variable and the same predictors as in Model 1. Again, we found a significant effect of age, such as older gamers tended less to change behavior while playing, χ 2 (1) = 4.10, OR = 0.99, p = 0.043, and male players tended to report changing more their behavior while playing compared with female players, χ 2 (1) = 6.10, OR = 0.77, p = 0.014. Highly achieving gamers, χ 2 (1) = 34.01, OR = 9.51, p < 0.001, and players who reported playing more time during the week, χ 2 (1) = 8.59, OR = 1.07, p = 0.003, tended to report changing more their behavior while playing. On the contrary, gamers who play to socialize tended to report changing less their behavior while playing compared with other gamers, χ 2 (1) = 5.70, OR = 0.99, p = 0.017.
Discussion
This aim of the study was to propose two explanatory models of toxic behaviors displayed by gamers during multiplayer online video games. Younger age, male gender, achievement motivation, which includes competition, and spending a lot of hours to play per week were found to increase the likelihood of toxicity and behavioral changes while playing. A prior study found that younger generations tend to prefer action games, 36 a game genre that includes a large variety of subgenres, ranging from shooter games to MOBA, and which often imply competition between gamers. In contrast, older adults demonstrated poor adherence to this type of games and preferred intellectually stimulating games, strategy games and simulator games. 37 Preference for MOGs such as role-playing games and shooter games was found to be a significant predictor of daily gaming time. 38 Moreover, playing competitively predicts higher levels of aggressive cognition in gaming context24,39,40 just as competition does in other activities such as sports and gambling. 39 The results in the present study demonstrated that males motivated by achievements in video games were more likely to be toxic than others. Hence, gamers not only have greater levels of aggressive thoughts when they play competitively, they also display aggressive behaviors in game accordingly. This imply that playing violent video games might not be sufficient to elevate aggressive behavior. In line with past research,23,41 this means that competition, not graphic depiction of violence, may be the video game characteristic that has the greatest influence on aggressive behavior.
In contrast, women are less involved in video games than men and when they do play, they often prefer different games. 42 Notably, women appear to be less interested in game genres featuring competition than men. 43 Added to the fact that women were found to have greater empathic disposition in comparison with men,44,45 this might account for the sex difference regarding toxic behaviors in the present study. Higher age and employment predict lower gaming time. 38 Moreover, older generation focuses more on strategic gaming offerings 36 and solitary playing with a particular fondness for casual personal computer games, 46 thus being less interested in competition and violent or action games than younger males.
Empathy is the basis of much prosocial interpersonal behavior 47 and is a likely contributor to the inhibition of aggression and antisocial behavior. 48 So it makes sense that the results of the present study demonstrated that gamers who have higher levels of empathy were found to be less likely to be toxic. On the other hand, low empathic concern has been strongly associated with lower prosocial behavior toward strangers 49 and with proviolence attitudes toward other players. 50 Sensation seeking and negative urgency, two components of the UPPS assessing impulsivity, predicted higher in-game toxic behaviors, which is similar to the findings of a meta-analytic review, which tested the hypothesis that sensation seeking would be positively related to aggression 51 and of several studies which found that negative urgency was associated with aggression in both children and adults.52,53
There are several limits to this study. First, the sex ratio of 1–7 might be a possible bias, but the sex ratio difference is also similar to the findings of previous studies showing that sex differences continue to exist between men and women in video game experience, favoring males: they play video games more often 54 and they overwhelmingly play more violent video games than women. 55 Another limit regarding the sample was the high percentage of young adults. Indeed, only 7.96 percent of the participants were 30 or older, although results from various polls concur on the fact that the average age for gamers is ∼34 (Refs.56,57). Moreover, even if several significant effects were found, the use of a single item with a dichotomous answer to assess toxic behaviors in the Model 1 was probably not representative of the variety of gamers' interactions while they play. The study would have been more accurate if the toxic behaviors had been better differentiated and categorized. Likewise in Model 2, “behavior change while playing competitive games” might have been interpreted very differently from one player to another and would have benefitted from more specifications. Furthermore, these two measures are only indirect measures of in-game behavior. Last but not least, the study relied on self-reported data. Therefore, caution should be taken with regard to generalizing the results.
Conclusion
The study sought to examine users' toxic behavior (a form of verbal aggressive behavior) and behavioral changes in MOGs, by exploring the influence of personality traits (empathy and impulsivity), emotion reactivity, and motivations to play on in-game behaviors. Past research mainly focused on assessing the influence of violent video games on aggression, leaving aside others aspects of aggressive behaviors which occur during online games. Moreover, the methods of the present study differ from those that have examined toxicity in online games, a phenomenon which has mostly been studied through linguistic analysis to assess communication patterns.
Overall and despite aforementioned limits, the findings provide support for the influence of personality traits, emotion reactivity, and motivations on the way people behave in video games. These factors might help to understand the occurrence of aggressive behaviors during online gaming, regardless of the content (i.e., violent or not) of the game. The study results imply that individual factors play a major role in how gamers behave toward each other when they play, all the more so in competitive contexts. MOGs often display various forms of violence. Yet, all gamers do not display verbal aggression when they play, which means that playing video games containing violence does not automatically lead to in-game aggression between gamers. Toxicity seems to be driven by individual characteristics, such as young age, male sex, being highly achieving and being high in sensation seeking, negative urgency, and emotion reactivity. On the contrary, empathy seem to be a protective factor against toxicity and socialization motivation against changing behavior in competitive contexts.
Future research on video gaming would benefit from in-depth analyzes of toxic behaviors displayed in MOGs: the creation and validation of a “toxicity” measurement scale would provide more accurate results regarding toxic behavior specificities and their links with personality, emotions, and motivations. Moreover, experimental studies with direct observation of gamers while they play would bring meaningful insights to better understand in-game behaviors. Lastly, employing the fuzzy-set qualitative comparative analysis method might help to look for complex relationships among variables.
Footnotes
Author Disclosure Statement
No competing financial interests exist.
Funding Information
The authors received no specific funding for this work.
