Abstract

The purpose of this study was to explore whether video-game play can influence the antisocial and prosocial behaviors of the players. It was hypothesized that prosocial and antisocial behaviors would depend on the presence of violence in games. More prosocial behaviors were expected by those who played non-violent games; more antisocial behaviors were expected for those playing violent games. Male and female gamers from a wide range of age and racial/ethnic backgrounds at two arcades were observed for instances of prosocial and antisocial behaviors when playing against each other or teamed together against the computer. Observation took place for approximately 2 to 3 hours on each of three days. Chi-squared tests were used to compare for differences in behavior gamers exhibited in non-violent and violent games. Contrary to the hypothesis, no antisocial behavior was observed in the sample. Prosocial behavior was seen occasionally, but its appearance was not significantly different in violent and non-violent gamers, indicating that prosocial behavior was independent of game violence.

Introduction

The majority of researchers view video games in a negative light, often implying that they are one of the factors to blame for violence among youths. For example, Anderson and Bushman (2001) associated “violent video games” with antisocial behaviors such as aggression. Silvern and Williamson (1987) found that 4- to 6-year-olds’ aggression increased and prosocial behavior decreased after violent video game play.

Antisocial behavior is defined as socially undesirably behavior, including antisocial verbalization and antisocial action with or without verbalization. Prosocial behavior is defined as “socially desirable behavior that includes positive interpersonal behaviors … achievement-related behaviors … [and] imagination” (Calvert 1999). The three subcategories for prosocial behavior are: behaviors that benefit others (positive interpersonal), help the individual succeed academically (achievement-related), and promote creativity (imagination).

According to arousal theory, players of violent video games become physiologically aroused — increased activity in brain waves, heart rate, blood pressure, and skin conductance — through exposure to the level of violence presented. After habituating to that level, the players become desensitized — requiring higher levels of aggression to be aroused — from overexposure to aggression (Calvert 1999). Based on the principals of the arousal theory, Anderson and Bushman (2001) developed the “General Aggression Model,” which states that people who play violent video games develop aggressive beliefs and thoughts which predispose the gamers to future violent behavior in real-life.

Many studies, however, suggest that video games also have prosocial benefits. Kestenbaum and Weinstein (1985) noted that aggressive games calm players. Scott (1995) found evidence of catharsis: undergraduates had fewer aggressive feelings after playing moderately aggressive games but had an increase in aggressive feelings after playing highly aggressive games. This appears to resonate with the idea that violent games may have prosocial effects, such as providing a safe means of venting aggression.

Materials & Methods

Participants

Thirty-four game-play pairs (68 total gamers) of both male and female subjects of varying ages (pre-school/kindergarten to middle-age) were observed in video arcades on three different weekend days (Friday, Saturday, and Sunday) from the afternoons through late evenings. A “game-play pair” is defined as each instance two gamers played together against the computer or against one another. Each gamer of the pair was scored separately for three minutes and up, and the same person could appear in more than one pair. The gender of the gamers was recorded, along with a best guess of their race/ethnicity and age group.

Materials

Two observers (one primary scorer and a second to establish reliability) were given scoring sheets with spaces to identify the gamer (by gender, approximate age, and race/ethnicity), the name of the game, and level of violence, and check boxes for prosocial behaviors (positive interpersonal behavior, achievement-related behavior, and imaginative behavior) and antisocial behaviors (antisocial verbalization and antisocial action with or without verbalization).

Results

As explained above, games were categorized into non-violent and violent games, and gamers were scored for prosocial and antisocial behaviors. All major races — 32 caucasian, 16 Hispanic/Latino, 13 Asian, and 7 black — and age groups — 4 preschoolers (younger than age 6), 10 early grade-school children (age 6-11), 16 late grade-school children (age 12-17), 31 young adults (18-35), and 7 middle-aged adults (over age 35) — were represented in the sample. One of the arcades was located in a middle-class neighborhood with racial and ethnic diversity, while the other was located in a working-class neighborhood largely comprised of Latino Americans. Although there were far more male than female gamers (56 to 12), exactly half of the males (28) and half of the females (6) played violent games.

Despite careful observations, there was not a single observed instance of antisocial behavior among the 34 game play pairs. Although most of the gamers displayed little verbal or visual contact with their playing partners, prosocial behavior was observed in 11 out of 34 pairs, ranging from complements and consolation to sharing of game strategy.

Table 1 compares the number of prosocial behavior occurrences between non-violent and violent gamers, with total number of minutes observed written in parentheses. The rate of occurrences was calculated by dividing the number of occurrences by the total number of minutes observed. Contrary to the hypothesis, violent gamers appeared to exhibit a higher rate of prosocial behavior; a 2 x 1 chi-squared test rejected the null hypothesis, χ2(1) = 12.48, p < 0.001. This unexpected finding may result from the presence of two very social gamers who played only violent games and were extremely positive outliers (rate of prosocial behavior occurrences were 1.00 and 2.00 per minute vs. 0.11 for entire sample). After excluding the two outliers (shown as adjusted occurrences in Table 1), the null hypothesis could no longer be rejected, χ2(1) = 0.20, p > 0.05, χ2(1) = 0.20, p > 0.05. Therefore, there was no significant difference between the amount in the rate of prosocial behavior occurrences between in non-violent and violent gamers.

In order to better account for prosocial behavior in gamers, numbers of prosocial gamers—those who showed at least one instance of prosocial behavior—were analyzed for non-violent and violent games in place of total occurrences of prosocial behavior (Table 2). A 2 x 2 chi-squared test of independence was done to test for the independence of prosocial behavior from violence in games. Yates’ Correction was incorporated to account for the small sample size. The chi-squared value was not significant; therefore, the null hypothesis could not be rejected, χ2(1) = 0.29, p > 0.05.

It is possible that there might have been a confound with the modality of the game type (competitive versus teamwork games). The majority of violent games were teamwork games (28 of 34), while 32 out of 34 non-violent games were competitive. It may have been that the teamwork modality facilitated greater prosocial behavior in the gamers who played violent games, which were mostly teamwork games. A 2 x 2 chi-squared test of independence (see Table 3) was performed to test for the independence of prosocial behavior from modality of game type. Yates’ Correction was incorporated to account for the small sample size. The chi-squared value was not significant; therefore, the null hypothesis could not be rejected, χ2(1) = 0.04, p > 0.05.

Based on this study, the experimental hypothesis — more prosocial behaviors expected in gamers who played non-violent games, and more antisocial behaviors for those playing violent games — was not supported (Table 2); the null hypothesis could not be rejected, possibly indicating that prosocial behavior was independent of video game violence. Originally, the rates of prosocial and antisocial behaviors were to be measured. Non-violent games were predicted to have a higher rate of prosocial behavior and lower rate of antisocial behavior than those for violent games. However, during the period of preliminary and actual studies, antisocial behavior was not once observed. One possible explanation could be that antisocial people do not go to arcades and video games may not be strong enough to make arcade-goers display antisocial behaviors.

In preliminary studies, games were categorized into four categories (levels I through IV): level I games with no violence; level II with cartoon/animated violence; level III and IV were graphically violent games with some killing involved. Level I, II, and III games had goals to win or to achieve goodness of humanity, whereas level IV games had no coherent goal. Since it turned out that difference between the games of level II and III was not very distinct, but only quantitative, both were combined to form “violent games” and level I games were renamed to “non-violent games.” There was only one game that made it to the level IV: a shooting game without any coherent goal except to mutilate grossly mutated monsters from a circus. This may have been the only game that instigated antisocial behaviors. Unfortunately, however, only one person, not a pair, briefly tried to play the game. It would be interesting to obtain sample pairs for this game in future studies, the hypothesis being that people who would act like characters in these types of games gravitate toward them and reinforce their own tendencies toward antisocial behavior.

Another major limitation of the study was the focused sample: the vast majority of the sample was observed in one game room in Rockville, Md. Although there was good diversity in age, gender, and ethnicity, a more diverse socioeconomic and regional background would be needed for greater external validity. In additional studies, it would be useful to include a larger sample size coming from more diverse socioeconomic and regional backgrounds, more female gamers, and longer durations of observation. During the span of this study, there were no observations of any antisocial behavior. Thus, future studies could also include conditions that might promote antisocial behavior among players by focusing on more grotesque games (level IV games) in arcades, or by studying gamers at home.

If violence of video games was not a determinant influencing prosocial or antisocial behaviors of game players, would modality of the game (competition between gamers versus playing together as a team against the computer) be one? It is possible that the modality might have acted as a confound, since the majority of the violent games allied the two gamers against the computer while the majority of the non-violent games involved competition between gamers. There was no significant difference between the proportions of gamers who exhibited prosocial behavior in competitive games versus those in teamwork games, indicating that the prosocial behavior is not directly affected by modality of games in this study. However, it may be worthwhile looking further into the effect of the modality of the game, collecting larger samples.

Media researchers have tried to establish a causal relationship between playing violent video games and violent behavior in people. Utilizing foundations of the arousal theory and portraying the shooters of Columbine High School in Littleton, Colo., as representatives of video game players, Anderson and Bushman (2001) developed the “General Aggression Model” to connect violent content in video games to antisocial behavior (including violent school massacres), with aggressive thoughts provoked by violent video games serving as the intermediary between violent video game content and aggressive behavior. Although Anderson and Bushman (2001) measured physiological responses to video game play to connect arousal with antisocial behavior, they failed to show a clear association of increased physiological activities, such as electrical brain activity and sweating, and aggression in video game players. Although the arousal theory seems to have some validity with respect to video game players and, indeed, the players appeared to become noticeably stimulated, this did not make them systematically desensitized and seek increased stimulation. If that were the case, gamers would have successively played more and more violent games, but that did not happen in this study. Contrary to the General Aggression Model, gamers in this study behaved prosocially — not antisocially at all — in their interactions with their playing partners, exhibiting signs of increased prosocial (not aggressive) behavior.

Based on the findings of this study that violence of video games did not affect antisocial or prosocial behavior of game players, the social influence of video games may be explained by temperament. Temperament is defined as a stylistic component of behavior that explains how a person reacts or responds to situations, and can be determined as early as infancy (Chess and Thomas 1996), which is considerably before anyone plays video games. It is possible that players’ behaviors are not altered by the video game — non-violent or violent — unless they have a “difficult temperament.” According to Chess and Thomas (1996), only 10% of the population has a “difficult temperament.” It is quite possible that only players with a difficult temperament will manifest antisocial behavior by playing violent video games.

While researchers continue to search for the negative aspects of video games, they continue to ignore the possible benefits that can be accrued from video games. With respect to the increased stimulation, gamers do not necessarily become desensitized to violence. Many gamers find playing video games as a way to improve and practice their mental focus and attention. The increased stimulation also provides gamers with a way of actively meditating — relaxation from diversion of stresses and focusing on a fun game. Kestenbaum and Weinstein (1985) noted how aggressive video games can calm gamers. As with anything — music, movies, and even research — there exist some repugnant extremes. However, by exaggerating the prevalence of these extreme cases, researchers unfairly discriminate against the majority of games — violent or not — by associating them with these worst cases. As a result, researchers have portrayed the worst possible aspects that video games have to offer.

Contrary to Chambers and Ascione (1987), this study found that video games do, in fact, provide a means to demonstrate prosocial behavior. The findings in this study also offer a more promising view of video games; while Anderson and Bushman (2001) declared that “violent video games” are linked to antisocial behavior, not a single instance of such was observed.