Effects of Playing Non-Violent Video Games Cooperatively or Competitively on Subsequent Cooperative Behavior
Caitlin Li, Sean Kotta, Natalie Espinoza, Sarah Viscosi, Nataly Aguilar, Liz Kyonka ([email protected])
presented at the Association for Behavior Analysis International, May 23 2026, San Francisco, CA
presented at the Association for Behavior Analysis International, May 23 2026, San Francisco, CA
Rocket League Experiment Explanation
by Natalie Espinoza
Our Rocket League experiment was inspired by the Ewoldsen et al. (2012) Halo experiment which concluded that gamers who played violent video games cooperatively engaged in more prosocial behavior than gamers who played competitively. We asked whether participants who competed against each other in rocket league cooperate less often in a 20-trial Iterated Prisoner’s Dilemma (IPD) compared to participants who played as a team.
To simplify the experimental process and avoid logistics issues such as participant no- shows, the study included a confederate rather than pairing two participants. The first step of the experiment was the IPD task. Participants played a computerized game against what they believed was another person, but was actually a computer programmed to use a tit-for-tat strategy. During the task, participants chose between up and down arrows. Mutual cooperation earned both players the greatest long term reward, while defection resulted in one player receiving 5 points and the other receiving fewer points instead of both earning 3 points consistently per round.
The next step was the Rocket League portion which consisted of either a cooperative or competitive game. In the cooperative condition, the participant and confederate played together against the bots. In the competitive condition, the participant and bot played against the confederate and a bot. After the gameplay, participants completed the IPD task again, followed by a competitiveness questionnaire, and then received a $20 amazon gift card for participation.
Participants who played cooperatively chose to cooperate in the IPD 75% of the time, significantly more often than participants who played competitively, who cooperated only 48% of the time, F(1, 35) = 5.72, p = .022. However, this difference was attributable to a difference in Pregame cooperation during the first IPD rather than group assignment.
Although the change in cooperation by some participants supported previous research suggesting that cooperative gameplay may encourage greater prosocial behavior, the results were more complex than hypothesized. Ideally, there would have been little to no difference between groups in the first IPD task, with behavior differences emerging after gameplay. Instead, group mean cooperation by participants in the cooperative condition decreased slightly, while group mean cooperation by participants in the competitive condition increased over time. This raised the possibility that factors beyond the gameplay style may have influenced the results.
Participants scored about the same number of goal regardless whether they cooperative or competitive condition. The rocket league bots scored many more goals compared to the confederates, so participants in cooperative conditions usually lost and participants in competitive conditions usually won.
For an additional measure of prosociality, participants rated confederates' favorability on a 5-point Likert scale after each IPD. With few exceptions, ratings increased over time. The score differential (how many points the participant won or lost the Rocket League soccer game by) was a better predictor of change in confederate favorability rating than group assignment or trait competitiveness score from the questionnaire. The relation between score differential and change in favorability rating lead us to the conclusion that, in this case, winning is what induced prosocial behavior.
Overall, our findings suggest that the relationship between video game play and prosocial
behavior may be more complex than simply cooperation vs competition. Winning and losing play an important role in influencing cooperative behavior after gameplay.
by Natalie Espinoza
Our Rocket League experiment was inspired by the Ewoldsen et al. (2012) Halo experiment which concluded that gamers who played violent video games cooperatively engaged in more prosocial behavior than gamers who played competitively. We asked whether participants who competed against each other in rocket league cooperate less often in a 20-trial Iterated Prisoner’s Dilemma (IPD) compared to participants who played as a team.
To simplify the experimental process and avoid logistics issues such as participant no- shows, the study included a confederate rather than pairing two participants. The first step of the experiment was the IPD task. Participants played a computerized game against what they believed was another person, but was actually a computer programmed to use a tit-for-tat strategy. During the task, participants chose between up and down arrows. Mutual cooperation earned both players the greatest long term reward, while defection resulted in one player receiving 5 points and the other receiving fewer points instead of both earning 3 points consistently per round.
The next step was the Rocket League portion which consisted of either a cooperative or competitive game. In the cooperative condition, the participant and confederate played together against the bots. In the competitive condition, the participant and bot played against the confederate and a bot. After the gameplay, participants completed the IPD task again, followed by a competitiveness questionnaire, and then received a $20 amazon gift card for participation.
Participants who played cooperatively chose to cooperate in the IPD 75% of the time, significantly more often than participants who played competitively, who cooperated only 48% of the time, F(1, 35) = 5.72, p = .022. However, this difference was attributable to a difference in Pregame cooperation during the first IPD rather than group assignment.
Although the change in cooperation by some participants supported previous research suggesting that cooperative gameplay may encourage greater prosocial behavior, the results were more complex than hypothesized. Ideally, there would have been little to no difference between groups in the first IPD task, with behavior differences emerging after gameplay. Instead, group mean cooperation by participants in the cooperative condition decreased slightly, while group mean cooperation by participants in the competitive condition increased over time. This raised the possibility that factors beyond the gameplay style may have influenced the results.
Participants scored about the same number of goal regardless whether they cooperative or competitive condition. The rocket league bots scored many more goals compared to the confederates, so participants in cooperative conditions usually lost and participants in competitive conditions usually won.
For an additional measure of prosociality, participants rated confederates' favorability on a 5-point Likert scale after each IPD. With few exceptions, ratings increased over time. The score differential (how many points the participant won or lost the Rocket League soccer game by) was a better predictor of change in confederate favorability rating than group assignment or trait competitiveness score from the questionnaire. The relation between score differential and change in favorability rating lead us to the conclusion that, in this case, winning is what induced prosocial behavior.
Overall, our findings suggest that the relationship between video game play and prosocial
behavior may be more complex than simply cooperation vs competition. Winning and losing play an important role in influencing cooperative behavior after gameplay.
References
Herbranson, W. T., & Schroeder, J. (2010). Are birds smarter than mathematicians? Pigeons (Columba livia) perform optimally on a version of the Monty Hall Dilemma. Journal of Comparative Psychology, 124(1), 1–13. https://doi.org/10.1037/a0017703
Mazur, J. E., & Kahlbaugh, P. E. (2012). Choice behavior of pigeons (Columba livia), college students, and preschool children (Homo sapiens) in the Monty Hall dilemma. Journal of Comparative Psychology, 126(4), 407–420. https://doi.org/10.1037/a0028273
Herbranson, W. T., & Schroeder, J. (2010). Are birds smarter than mathematicians? Pigeons (Columba livia) perform optimally on a version of the Monty Hall Dilemma. Journal of Comparative Psychology, 124(1), 1–13. https://doi.org/10.1037/a0017703
Mazur, J. E., & Kahlbaugh, P. E. (2012). Choice behavior of pigeons (Columba livia), college students, and preschool children (Homo sapiens) in the Monty Hall dilemma. Journal of Comparative Psychology, 126(4), 407–420. https://doi.org/10.1037/a0028273
