Are people who have usernames based on popular shonen animes prone to have a higher frequency of antisocial behaviour / rage in comparison to the norm? How many mothers (or so they say…) have fallen victim to that twelve-year-old on the other side of the mic? Online games have grown to a point where players are capable of challenging each other through matchmaking to compete with people from all over the world while establishing a sense of team spirit with those they’ve never met before.

With billions of gamers across the world, millennials are a big bulk of this modern gaming world. Researchers Kokkinakis, Lin, Pavlas & Wade (2015) have found that millennials scored higher on narcissism than any other age-matched cohorts from previous generations, so it is no surprise that some of the most popular online competitive games out there such as Defense of the Ancients 2 (DotA2), League of Legends (LoL), and Counter-Strike: Global Offense (CS:GO) have implemented systems to discourage “toxic” in-game behavior. Most would agree they turn to gaming as an enjoyable activity to unwind after a stressful day. So let’s ask ourselves — why are some players more prone to flaming others, intentionally feeding in competitive matches, and participating in communication abuse?



Truly, if you simply thought of online trolls as those who would typically indulge in toxic behavior, you would not be wrong. Golbeck (2014) describes an internet troll as someone whose purpose in the online community is to deliberately upset others. In a recent study conducted by Buckels, Trapnell & Paulhus (2014), the notion where online trolls will exaggerate, lie and offend others just to get an emotional response was put to the test. The researchers conducted two online studies with 1,215 participants who completed a series of personality tests along with a survey about their internet commenting behavior. These personality tests assessed the participant’s scores in relation to the Dark Tetrad, which consists of four personality traits:

  1. Narcissism, which can be described as a maladaptive form of self-obsession that manifests itself as a massively egotistical person who is very often self-centered.
  2. Machiavellianism, which can be described as a manipulative, detached egotist who typically lacks a conscience. A person with this trait scores high in all aspects of the Dark Triad.
  3. Psychopathy, which is typically a pattern defined largely by the absence of empathy and the lack of the emotional aspects of a conscience.
  4. Sadism, which is the tendency to derive pleasure from other people’s suffering.

Unsurprisingly, the online trolls scored exponentially higher in the Dark Tetrad traits in comparison to the norm. Buckels, Trapnell & Paulhus thereby concluded that online trolls are most likely prototypical everyday sadists. To quote the authors, “Both trolls and sadists feel sadistic glee at the distress of others. Sadists just want to have fun… and the Internet is their playground!”



How is antisocial behavior manifesting in the gaming community?

A rather interesting study conducted by Kokkinakis et al. looked at online antisocial behavior in regards to username selection and age using a multiplayer online battle arena (MOBA) game called League of Legends (LoL). A MOBA can be defined as a large virtual battleground which requires complex problem solving and social interaction in order to be successful. In order to ‘git gud’ at MOBAs, Drachen et al. (2016) states that players and teams need to have strong tactical knowledge, as well as a substantial investment of time. Kokkinakis et al. was curious to explore whether there was a correlation between usernames chosen by players and their in-game behavior. In general, they found both the age of the players and their use of highly anti-social words to be relatable with how they acted in-game. This suggested that the players’ real-world characteristics have some degree of influence over their behavior and interpersonal interactions within online games.

The researchers described antisocial tendencies as the tendency to engage in behavior that deviates from societal norms and is likely to cause offense to a large proportion of people. They were convinced that there was a solid correlation between player ages and their tendency to use foul and/or offensive language. Kokkinakis et al. used the North American server of LoL to obtain a sample of 2,198 users whose names were designated to be antisocial. This meant their username contained some kind of blatant racial, sexual, or obscene label. Similarly, they also randomly selected a control group of equal size with non-antisocial names. They found that players with antisocial names had sent a significantly higher number of reports, as well as having received them. With no surprise, the control group had significantly lower sent and received ‘honor ratings’, which reflect in-game altruistic & prosocial behavior.



When the researchers looked at age as a variable in the study, they had found a significant relationship between age and online interaction rates. In general, there was a pattern where the rate of negative interactions decreases with age. To specify, older players between the ages of 22-26 years old were significantly less likely to send or receive reports for antisocial behavior in comparison to younger players between the ages of 11-15 years old. For simplicity’s sake, the important findings thus far include:

  • Antisocial naming tendencies are associated with antisocial gameplay, which led to being reported more often and having received lower honor rates than the norm.
  • In comparison to the control group, players with antisocial names were often the instigators of negative interactions and were more likely to criticize their teammates than to praise them.
  • There is a correlation between in-game antisocial behavior and losing games.
  • As people age, they become more prosocial. It can be expected that negative interactions will decrease while positive interactions increase. Kokkinakis et al. describe this finding to be small at an individual level, but very robust and significant at a group level.

Why is it the case that the bulk of negative interactions are instigated by adolescents?

Adolescence can be defined as a period where an individual undergoes significant changes in important brain structures that oversee decision-making, such as the amygdala and frontal lobes. While the limbic system is maturing quicker than frontal lobe structures, the adolescent is more prone to reacting to emotionally provoking stimuli. This can be seen as what drives the overall higher negative interaction rates in younger players.

Researchers also suggested that, due to the increased ratio of negative to positive interactions in younger players, there also may be an issue of reduced cognitive control in this cohort. In a separate study (see Dreyfuss et al., 2014) that assessed emotional sensitivity towards potentially threatening stimuli in adolescents, it was found that even when the adolescents were told to ignore threatening stimuli, there was still an increased sensitivity toward it. Adolescents are more emotionally sensitive, which may cause them to misinterpret a neutral message as a hostile one. Therefore, implications can be made that adolescents are generally less capable of inhibiting possibly threatening stimuli, which tends to lead to communicational escalations.



The Dunning-Kruger Effect in competitive games.

Most would agree that group collaboration is the way for individuals to exchange ideas and arguments in order to build upon a better decision than if they were to work solo. This implies that the individuals should be relatively competent — but this is bad news, because confidence can be misaligned as their actual competence. Unsurprisingly, the team tends to under-perform when this is the case. In this context, confidence can be defined as a self-estimation of your own performance, whereas competence is defined as your actual observed performance in a given task. Thus, the Dunning-Kruger effect is a cognitive bias where people of low ability overestimate their own abilities.


What are the effects of confidence-competence misalignment?

Individuals who are the most confident generally have the larger impact on the group’s decision, even when they’re at the same level of competence as everyone else. Fu, Lee & Danescu-Nichulescu-Mizil (2017) were interested in the effects of mismatched competence-confidence in a group collaborative setting. They presented an online collaborative geography puzzle game that was played by over 10,000 teams, where players would have potentially different “views” of a given location when they tried to determine that location by pooling their information. The system then records each individual’s guess before they interacted with their teammates, as well as their self-reported confidence in the answer, the entire team conversation, and the accuracy of the final answer that the team settled on.

The researchers notably mentioned that more than half the users indicated a confidence level that was mismatched with their competence level. As a result, there was a substantial number of instances for under- and over-confidence occurrences in the context of group decision-making. When the most competent individual displays less confidence than the least competent member, the group generally performs worse compared to teams with the same position in terms of correctness but in which individuals have accurate confidence estimates. In other words, teams will fail to reach their potential in terms of performance.



The key to successful teams in consonance to science?

According to researchers Kim et al. (2017), the key to building strong teams is collective intelligence (CI), which they define as a factor that can measure a group’s ability to perform together on a wide range of different tasks. In other words, if a group of people does well in one type of task, it should also be expected that they perform equally as well on other tasks. CI is supposedly built upon each individual member’s combination of attributes, processes, and norms (just like a game, who would have known?!). But most importantly, the researchers found CI capable of predicting the competitive performance of teams during online MOBA matches, where self-organized, time-pressured, and intense collaborations occur.

MOBAs are easily characterized by their intensity, necessity for fast decision-making, and competitiveness. As virtual teams are becoming more prevalent, Kim et al. deemed that studying CI could shed new light on how it operates in other environments, such as the modern business world. The similarities lie in the necessary quick-mindedness required by business leaders that need to make and modify decisions in response to rival companies. Research from previous studies indicates that groups containing individuals with the ability to perceive subtle emotional and interpersonal cues tend to have higher levels of CI than those who cannot. As such, women tend to score higher on these such tests. Understandably, groups with more women have higher levels of CI.

In the end, it all comes down to communication — whether verbal or non-verbal, being able to work together and communicate important information to each other in dire situations is the game changer.