Pages

Wednesday, December 30, 2015

Guns And Controllers: The APA Reviews, Kotaku Gets Aggressive, The Game Continues

Just about one year ago, I summarized the best research available on the topic of violent video games causing aggressive behaviour. Reviewing several meta-analyses and discussing the APA's statement on the subject, I concluded that violent video games almost certainly have a minimal effect and likely have a higher effect on aggressive cognition. For those of you who follow my posts, this was a long winded process, and left on a cliff hanger: the APA had been petitioned to review the literature again and revise their first resolution on the subject. This effort was led by Christopher Ferguson, one of the leading researchers in this field, and joined by approximately 228 media scholars and academics as well. That was in October of 2013.
Remember this picture? I know you all missed me.

In August of this year, the APA released a new resolution as well as a full report on video games and violence/aggressive behaviour. Keep in mind that the last APA resolution on this subject was from 2005, so this is big news since we have had just about 10 years to analyze and accumulate new studies and research. Surely after taking a new look at the data, the APA would come to a more "reasoned" conclusion, right? As promised in my first article on the subject, I'm here to report. That's right, you get tons of presents for the holidays: the queen returns, she's getting newsy, and she's actually keeping one of her promises for once!

Focusing exclusively on research conducted between the time of the first resolution and now, the new APA resolution concludes the following:
  • The link between video games and aggressive behaviour exists, is robust, and is backed by the majority of data.
  • The link is not just limited to aggressive behaviour, but aggressive affect and aggressive cognition as well.
  • In addition, violent video games are also related to decreases in prosocial behaviour, empathy, and sensitivity to aggression.
Needless to say, people flipped. Kotaku wrote an article in response to the new APA resolution attempting to point out the numerous flaws in the studies they cited to convince people that the studies are "nonsense," and that the resolution should be disregarded. Oh, Kotaku...

Let's take a look at some of their contentions:
"An outside observer might wonder—how can you tell whether someone is 'more aggressive'? Is there really a way to measure an emotional state like aggression? Well, some of the tests used in violent video game studies include:
A) The 'short story' test, where a subject is given the beginning of a writing prompt ('A driver crashes into Bob’s car. Bob gets out of his car and approaches the driver.') and told to fill in what happens next.
B) The 'noise' test, where a subject is asked to press a button that delivers a terrible sound to another subject, then evaluated based on how much noise they deliver and how intense it is.
C) The 'hot sauce' test, where a subject is asked to dole out hot sauce to another subject and is evaluated based on how much sauce they give and how spicy it is.
Other tests ask subjects to fill out questionnaires asking how aggressive they feel, and if all this has you raising an eyebrow, you’re not alone. 'Aggression' is an ambiguous psychological concept—if I get mad at a game and scream at my TV for a few seconds, am I being aggressive?—that can only be measured in subjective and often arbitrary ways."
Kotaku is correct to say that there is no single way to define or study aggression, and that it is often subjective and arbitrary, but it is meaningless to point this out here. There are swaths of psychological traits that we study every day that are largely arbitrary in their definition, such as intelligence, impulsivity, introversion versus extroversion, empathy, and so on. To say that the studies examining these things are useless betrays a lack of understanding regarding why we create these terms in the first place. Simply put, a term is only meaningful if it can practically be used to examine something. For example, the term "intelligence" is meaningful because it can be one word which describes several factors which are often correlated with one another: good grades, analytical skills, and spatial ability to name a few. It is true that you can define "intelligence" in a number of other ways too, but as it is defined here, it is quite useful and predictive.

"Aggression" - a useful definition for real, observable things.
Similarly, we define aggression a number of ways. In the APA resolution, they make their definition clear: "behavior that is intended to harm another." This is the practical definition that is used by most researchers in the field of psychology; thus, when a study claims to find an association between video game violence and aggression, it can be reread to claim an association between video game violence and behaviour that is intended to harm another person. It means the same thing.

So what about these tests that these researchers use? Do they help us examine "aggressive" behaviour, affect, or cognition (note: this is an important distinction that Kotaku neglected to make; these tests were used for different psychological traits entirely, and so may not make sense when examined exclusively through the lens of "aggressive behaviour")? The answer is yes. Imagine the answers one might receive to the question in example A. Bob gets out of his car and approaches the driver, they exchange information, and they go about their way because they both have jobs to get to. Someone else might answer, Bob is shot and the driver flees. Aggressive cognition.

Example B, a button which delivers a noise which hurts another person's ears. Person 1 presses the button quickly and flinches. Person 2 slams down on it and holds it until the air runs out. Aggressive behaviour.

Example C, intentionally giving someone spicier hot sauce, more hot sauce, or both, is intended to hurt them. Aggressive behaviour.

Kotaku also seems to have chosen, randomly or otherwise, those tests which appear to have the least merit. Some other tests include parental reports, peer reports, and teacher reports of explicitly aggressive behaviour such as kicking, hitting, threatening, hair pulling, insulting, biting, pushing, and much more. There is also the Implicit Association Test, which is a direct test of cognitive processes based on subject response. The tests used to determine aggressive behaviour, cognition, and affect are numerous, which is why I like to tell people to defer to concordance among sources. Like climate change denialists, those who refuse to accept the causal connection between violent video games and aggressive behaviour will either throw out broad generalizations about how these studies are conducted or will nitpick every single one. Disregarding calculations for effect size in meta-analyses such as these, the fact is that the vast majority of research is in agreement, regardless of the particular method each study uses. People using multiple different methodologies and coming to the same conclusion is evidence supporting the hypothesis in question, not taking away from its credibility.

So regardless of whether or not it's "arbitrary," it's consistent and useful. For research purposes, this is all that matters.
"One major problem with the tests used by these studies is that they all measure their subjects’ aggression directly after they’ve played violent video games. Even if you assume the tests are good ways to measure aggression, this is not particularly useful information for practical purposes. If you’re a parent who wants to know how violent video games might affect your children, the bigger concern is how their behavior will be impacted in the long run.
But there aren’t enough studies on the long-term effects of violent video games. Admits the APA in their report: 'However, the meta‐analyses we reviewed included very few longitudinal studies, and none of those that were included considered enough time points to examine the developmental trajectory of violent video game use and associated outcomes.'
So the APA’s conclusion—that there’s a consistent relation between violent games and aggression—is misleading at best. What they’ve actually concluded is that there’s a consistent relation between violent games and short-term aggression."
The statement Kotaku is referring to comes from the full APA report, which is still in the correction stage (i.e. why I waited so long to talk about this). This is found on page 4 in response to questions about whether or not violent video games have particularly harmful effects for children and adolescents who are susceptible to developmental harms. Given this context, it is easy to see why the APA would respond in this way: there simply isn't enough evidence to suggest that violent video games significantly impact a child's development. But do they really believe in the lack of evidence overall from longitudinal studies? This is not so. In summarizing the outcomes of the research on page 10, they say the following, emphasis my own:
"Since the earlier meta-analyses, the literature has broadened in some directions. For example, there are more longitudinal studies and multi-exposure studies. The literature has also broadened in terms of populations studied, including a limited number of children, high-risk populations, and non-U.S. samples, although more similar research is needed. Several longitudinal studies, using both experimental and naturalistic approaches, have helped establish that the effects of violent video game exposure last beyond immediate effects in the laboratory."
The emphasis on concordance between experimental and observational studies is important, because it shows that even those longitudinal studies which directly assess violent video games as a cause for aggression found an effect. For aggressive cognitions:
"Numerous laboratory and longitudinal studies have assessed the impact of violent video game use on aggressive cognitions, which includes both self-reports and direct measures of cognitive processes. ... Of the 31 studies reviewed, 13 included aggressive cognitions as an outcome. All of these studies showed an effect of violent video game use on increased aggressive cognitions, replicating the finding in the pre-2009 research."
That moment you find out you didn't read the whole thing.
This is not unusual for a research report to do. When given a specific question, they respond in a specific way. When asked about the overall findings of the research, the water gets clearer.

For those who are interested in the actual findings of the APA report, I would read it in its entirety (even in the correction stage). The one-liner that Kotaku cited has been noted abundantly by people who were outraged by this report, using it to essentially write the whole thing off. I personally saw this not only on the original Polygon report but on Reddit reaction threads as well. Truly comical to see so many people stop at 4 pages in to boast about the report's lack of power. Of course, this isn't to say that we couldn't use more longitudinal studies. We can always use more research, but at present, studies of varying designs have mostly come to the same conclusion. What else does Kotaku have to say?
"Few people are thinking about one of the most important factors: competition."
This is because the research is so lacking (or as the APA put it, "nascent") that nothing could be meaningfully drawn from examining the literature on competition. One of the only studies to date to isolate the effect of competition from violence on aggressive behaviour had a whopping sample size of 42 and 55 subjects respectively for pilots 1 and 2, with study 1 having an effect size of zero. They defend this by calculating the power of their test (~0.775), but the power of a test does not tell us if the effect actually exists, only the test's ability to detect it.

The lead author of the aforementioned study, Paul Adachi, submitted his doctoral thesis in 2013 in a similar realm. He took care of his sample problem, annually surveying 1,492 adolescents from grade 9 to grade 12 about their video game play and aggressive behaviour. He found that playing competitive games such as sports or racing games has a positive, moderate effect on aggressive behaviour, and that playing violent video games did not have an effect on this relationship. This effect also increases with how often the subject plays competitive games. So what's the problem here?

For one, the study was longitudinal, not experimental - it was a cohort. As explained a moment ago, longitudinal, experimental studies show that violent video games have an effect on aggressive behaviour and cognition. In addition, the assessment of violent video games was through a moderator analysis. All this means in the context of this study is that the correlation between competitive video game play and aggression does not go away when you account for violent video game play. This does not bring into question whether or not violent video game play has an effect on aggression; it only shows that there is an independent correlation between competitive gaming and aggression. And again, this is a correlation based on a cohort study.

The author of the Kotaku article, Jason Schreier, is not new to this discussion, and his review of the literature has not changed much over the years. Now that even more research has been conducted and the APA has weighed in once again on the subject, he still is not convinced. "It’s all of these questions," he writes, "and the subjectivity of scientific studies, most of which can be used to draw any number of conclusions—that have convinced me to avoid reporting on these violent video game journals every time we get a new press release or meta-analysis. There just isn’t enough research or
proper methodology to draw much from most of this science."

They're just sick and tired of it. They are just so, so done!
I think it's an interesting phenomenon that people have come to take to this "any conclusion can be drawn from a study" sort of thinking. It's a sweeping generalization that really seems lazy to me, when one can look at the individual merits of a study as I've done here and decide qualitatively whether or not the study constitutes any merit, or can be used for the claim that is being made. What's ironic enough is that Jason failed to see how the study he continuously cites has no bearing on this discussion, because he failed to apply such an analysis.

This just goes to show that the battle is far from over. The APA report will go largely ignored by those who have become disengaged and are tired of hearing the same thing come up every few years. Personally I'm tired of it too, but for different reasons. As I mentioned in my first article on the topic, this is something I came into being completely skeptical of the view that video games can cause aggression or violence. An honest review of the literature leads one to abandon the former view, but still leaves room for some skepticism of the latter, which is the good news: these studies can't just be interpreted any way you want. They can only truly be interpreted one way, and that is what they actually say.

Thank you all very much for reading, and have a happy New Year.


ResearchBlogging.orgReferences:

Adachi, P., & Willoughby, T. (2013). Demolishing the Competition: The Longitudinal Link Between Competitive Video Games, Competitive Gambling, and Aggression Journal of Youth and Adolescence, 42 (7), 1090-1104 DOI: 10.1007/s10964-013-9952-2

Adachi, P., & Willoughby, T. (2011). The effect of video game competition and violence on aggressive behavior: Which characteristic has the greatest influence? Psychology of Violence, 1 (4), 259-274 DOI: 10.1037/a0024908