Cold or Hot videogames or applying the Chinese videogame diet (Sestir & Bartholow, 2010)

My parents are immigrants from South East Asia and I never took their understanding on food seriously. They classify food as either hot or cold, fried food or anything with fat are called hot food (I think) and cold foods are… I don’t really remember, but they’re not necessarily uncooked food. But the point is that I was not allowed to eat anything that would upset the balance between hot and cold and if I did, I would get sick. They always serve soup for supper to take into account of the main course. Although, I hate some soups that I just drink the soup and dump the lumpy or disgusting stuff they put in, it always pisses off my dad.

This study published for the November 2010 issue of the Journal of Experimental Social Psychology drew my interest as it turned some of the research upside down with the fascinating results they got. “They” as in Marc Sestir (currently teaching at Hobart and William Smith Colleges) and Bruce Bartholow (University of Missouri-Columbia). They teased out aggressive outcomes from violent and nonviolent video games and the results seemed to suggest that I should follow my parents’ hot and cold diet to videogames.


Experimental studies routinely show that participants who play a violent game are more aggressive immediately following game play than participants who play a nonviolent game. The underlying assumption is that nonviolent games have no effect on aggression, whereas violent games increase it. The current studies demonstrate that, although violent game exposure increases aggression, nonviolent video game exposure decreases aggressive thoughts and feelings (Exp 1) and aggressive behavior (Exp 2). When participants assessed after a delay were compared to those measured immediately following game play, violent game players showed decreased aggressive thoughts, feelings and behavior, whereas nonviolent game players showed increases in these outcomes. Experiment 3 extended these findings by showing that exposure to nonviolent puzzle-solving games with no expressly prosocial content increases prosocial thoughts, relative to both violent game exposure and, on some measures, a no-game control condition. Implications of these findings for models of media effects are discussed.

If you are thinking into applying graduate school at Ohio State University and see my jaded face, please remember that you will be working 50-60 hours. I’m writing this post at my grad office on a Saturday.

The authors had some interesting points in their introduction and literature review section. They cited a criticism that videogame effects are short-lived stressing the need to consider the importance to test the temporal effect from the various psychological mechanisms of aggression. If videogame effects aggression comes from priming, then it cools down rapidly, if there are no opportunities to aggress. If the effects comes from aggression-related goal activation (I’m just citing them with a vague idea of what they mean), then it cool down more slowly until there’s an opportunity to punch someone’s lights out.

The authors considered the nature of nonviolent videogames and how it shouldn’t be treated as a control variable and more like a cold videogame (my words). The authors contend that no matter how many game features you try to match up with a violent equivalent, there’s going to be some differing qualities present than the violent content. Really? Then what about Przybylski’s study or what about Staude-Müller et al.’ study? Well, an analysis will detract away from the study in question. So, perhaps it would be a good idea to setup a content analysis study on nonviolent videogame and perhaps set up some guideline in modding games for nonviolent and violent gameplay. Returning on topic, they argue that nonviolent videogames are primarily problem-solving oriented, whereas violent videogames are primarily impulsive (read quicktime events) actions and “see-and-shoot” strategies. Whoa, I think they should clarify that they meant action games. Strategies games are a different breed, since they require planning and anticipating the opponent’s actions, much like chess. They reasoned that repeated in-game actions in nonviolent videogames would temporarily reinforce individuals to be more mindful of their behaviours, hence lowering their aggressive behaviours. Furthermore, they wanted to test nonviolent videogames with no explicit prosocial messages, because previous research used nonviolent videogames with explicit prosocial messages.

All of this is educated speculation and they formulated three hypotheses: Playing a violent videogame will cause increases in aggression and only after playtime is over. Videogames effects will be lower after a delay. Nonviolent videogames, those that lack any prosocial messages, will increase prosocial tendencies.

Experiment 1

Participants: 188 undergraduates, 131 are women and 57 are men. The gender bias is due to participants being psychology majors. Age not mentioned. Videogame experience was briefly mentioned and that nearly every men and over half of women had some recent experience with videogames.


Interpersonal Reactivity Index: measures trait empathy. I’m not familiar with that questionnaire.

Caprara Irritability Scale: measures aggressive impulsiveness, not familiar with that too.

Aggression Questionnaire: measures trait aggression. That’s the Buss & Perry Aggression questionnaire which is a 29-item on a 7-point likert scale where the participants rate how characteristic the item/statements are to them.

Aggressive Affect: 32 mood statement on a 5-point agreement likert scale. Isn’t the State Hostility Scale?

Aggressive Cognition: The word completion task. Participants must fill in word fragments with whatever to pops up in their mind. Participants are given 5 minutes to complete 98 word fragments.

Other: 7-point likert scales on how frustrating, aroused and interesting the game was.

Delay measure: a map-drawing task where participants are asked to draw a map of their campus in 15 minutes. A neutral time filler used in previous studies. Sounds like something I could use for my studies.

Videogames used: Quake 3 and Unreal Tournament: Game of the Year edition are the violent videogames. Zuma and the Next Tetris are the nonviolent videogames.


Participant are tested individually and are assigned to play either a violent or nonviolent videogame and whether they complete the aggression measures immediately after play or after a 15 minutes delay via the map-making task. And that’s done through the flip of a coin, I wonder whose idea was that, the coin-flipping that is. Participants are given a brief tutorial and playtime is 30 minutes. I’d like to comment about the playtime, but I’d defer this discussion to another study. After that, they completed the word completion task or the aggressive affect questionnaire first and then the other one for counterbalancing purposes.


They analysed the data through a 2 (game) X 2 (gender) X 2 (delay) factorial ANOVA. Covariates such as arousal, interest and frustration did not affect the results when entered or omitted from the analyses.

It’s text-heavy, so I made this table for you.

So, they got some interesting interaction effects in aggressive words. Participants in the violent game showed increased aggressive cognition in the 0 minute delay in comparison to the rest. However, those in the 15 minute delay showed no differences with the nonviolent group and significantly had a smaller percentage than the 0 minute violent game condition. This showed a cooldown effect that the authors contend that’s predicted from priming.

Participants in the 0 minute delay nonviolent condition showed a smaller percentage than the 15 minute delay nonviolent condition. The authors cautioned against the conclusion that the nonviolent videogame reduces aggression because there were no pre-test aggression measures. They argued against the use of a pre-post measure design because measuring one variable (i.e.  pre-test) would influence the measurement of other variables (post-test), sounds logical, especially if the participants would start wondering why they are being repeatedly tested on their aggression.

The author had found a single interaction for gender. The delay effect is much stronger among men than women in both game conditions. The authors contend that it might’ve been due to videogame experience and violence exposure.

Experiment 2

This experiment builds upon the first experiment by including a no-game control measure and an aggressive behaviour measure.

Participants: 389 undergraduates, no age measure or gender ration mentioned. Most likely psychology majors or from a psychology class.

Measures and videogames are the same from the first experiment, except that there are no aggressive cognition or aggressive affect questionnaire which they are replaced by a aggressive behaviour measure called the Competitive reaction time task.

The competitive reaction time task is basically asking participants to blast a painful noise to another ostensible participant who is also tasked to blast at the participant. Participants can set the noise level and duration of the noise blast. There are 25 rounds or trials.

They also used a different delay task, they used images from the International Affective Picture System and participants are to rate the colorfulness of the images. The images used in the experiment were neutral and of low arousal value. There’s 120 pictures for the participants to rate.

Procedure is similar to the first experiment.


43 participants were excluded from the experiment, half of it due to equipment failure or participants not following procedures, the rest of it due to not showing variability in their aggressive behavioural response or being suspicious of the study.

So that’s what they found.

At the 0 minute delay, the violent game condition was more aggressive than the nonviolent condition, a statistical trend in regards to the no game control condition (p = .07). The authors suggested a bigger sample size to verify the statistical trend. At the 15 minute delay condition, none of the conditions were different from each other.

The analysis comparing the delay conditions have found that the violent game condition showed less aggressive than those in the 0 minute delay. As for the nonviolent game condition, the 15 minute delay condition showed more aggression than the 0 minute delay condition. Finally, the control group is being true for their purpose, no differences.

Again, the authors cautioned against making any conclusions that nonviolent videogames cause a reduction in aggression since there were no pre-post test measures. Additionally, they noted that the nonviolent game isn’t exactly a prosocial game either and it is possible that the nonviolent videogames in the experiment only affect aggression and not prosocial tendencies.

Experiment 3

Of course, they went and investigated whether nonviolent videogames have any appreciable effect on prosocial tendencies.

Participants: 111 undergraduates, 68 are women and 43 are men drawn from a psychology class.

Measures are the same as experiment 1, except there are no aggression outcome measures, but two measures that could measure both aggression and prosocial:

Story completion task where participants are to read three ambiguous stories, such as having a disagreement with a friend, and were asked to provide 15 items detailing “what could have happened next”. I’m not sure what they meant. Since their responses are open-ended, they are judged and coded into either prosocial, neutral or aggressive.

Word Completion task: a modified version where they included prosocial word fragments. A total of 90 words, 30 words that consist of aggressive, neutral and prosocial words.


Same as experiment 1. The outcome measures are administrated at random for counterbalancing purposes. Additionally, there are no delay conditions. Why? I don’t know.


So, what they got is what is expected. The violent game condition are more aggression and less prosocial than the other two groups. With the exception in the story completion where there’s statistical trend (p = .06) between the violent game condition and the no-game control on aggressive completions. The authors contend that the correlation would be stronger with a bigger sample. The nonviolent game condition scored higher on the prosocial words, but didn’t differ on the aggressive words. As for the story completions, well there are the least aggressive and more prosocial of the conditions. The authors suggested that the words limited the participants’ accessibility or maybe opportunity to complete more prosocial words in contrast to the more open nature of the story completion task.


Back to the idea of a Chinese videogame diet therapy of cold and hot videogames, it seems like a wonderful idea to test with the results that Sestir and Bartholow have found. Parents might find it easier to handle their children’s violent videogame play by including some nonviolent videogame in their media diet, like romance visual novels, and it doesn’t have to be some game with a prosocial message, but any cold game would do or I quote the authors: “explicitly prosocial content in video games is sufficient, but not necessary to increase prosocial responses”. But let’s not get ahead as we really need more research funding and manpower because what I suggested is based on one study and we really need to be sure that in the probability that I’m in the wrong.

There are good reasons to be a bit pessimistic of the Chinese videogame diet. The authors noted that the nonviolent videogames are puzzle solvers that focused on logical and reasoning thinking which is contrasted with the action-pace of the violent videogames. Thus, there are limited generalizations to other genres and gameplay mechanics. Here’s a few factors that I could think of:  The narrative elements is absent from the experiment, however, I have this sinking feeling that somebody did a study on aggressive versus prosocial gaming narrative. But who? Second, a comparison between aggressive versus prosocial games in similar genres, especially in regard to real-time strategy games, since they require planning skills, would be great to study. At the top of my head, a resource management RTS game would be a fine example of a nonviolent videogames whereas any game from Relic Entertainment would do since they adopt a simple resource management style, but is compensated to focus on the violence and actions. Third, pacing control may be important, it relates a bit to narrative, but this is different conceptually. Take the example of Left 4 Dead where the pacing is analyzed and directed by an A.I. director. Perhaps controlling or structuring the game flow can designers calm their players by introducing other nonviolent activities, such as jumping puzzles or, like in most RPGs, shopping (get that hidden item on someone’s roof, poking at Far Cry 2). Or, reducing the number of enemies and uncertainty at various intervals and perhaps preferably near save points, just so the players would take that opportunity to save and quit.  Furthermore, there are existing gaming settings that allow players to cool off, such as the post-match screens, loading screens, post-match chat rooms, etc. Unfortunately, anyone who ragequit wouldn’t reap those benefits and who knows what they would do. Actually, I need to know what people do when they ragequit. Do they punch a wall or go join another game?


That's what my dad would say


Sestir, M. A., & Bartholow, B. D. (2010) violent and nonviolent video games produce opposing effects on aggressive and prosocial outcomes. Journal of Experimental Social Psychology, 46 (6), 934-942. doi: 10.1016/j.jesp.2010.06.005.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s