That’s a big head-scratcher. I mean, really? You’d think that the more a teen is plopped down in front of their gaming console the less energy they’d expend and the less likely they’d get physical exercise in the future. Nonetheless, there it was.
The claims come down to three main results. Five years after the start of the study, heavy gamers are less likely to smoke marijuana, less likely to have a high BMI, and more likely to suffer from depression.
Umm… those don’t really go together.
Looking closer at the study, there’s a lot of problems to be found with the statistics. First off, while the release touts that the study examined more than 10,800 youth in the United States, the actual number of data points for teens playing a lot of games was rather low. The researchers split them into three groups: 21+ hours per week (136 teens), 35+ hours per week (49 teens), and 42 hours per week (27 teens).
Only 136 data points is not a lot to be able to draw many conclusions from, let alone 27.
Next, we can dig down a bit more into the actual statistical analysis. First off, a p value is the likelihood of a researcher finding the same results due to random fluctuations rather than an actual effect. The closer a value is to 0, the more likely there’s an actual effect there. Typically, a p value less than .05 is considered to be decently suggestive of a real affect.
So in this study, we have the following p values:
Marijuana 21=0.019 35<0.001 42<0.001
Lower BMI 21<0.000 35<0.001 42<0.001
Depression 21=0.342 35=0.703 42<0.001
So while it looks like the results regarding marijuana and lower BMI might be a real affect, for some reason the depression rate only becomes unlikely to be a statistical anomaly once fewer data points are taken into account. Okay, so maybe there isn’t a linear relationship and only those playing 42 hours or more per week show any type of effect.
Let’s take a closer look at the effect sizes. After all, even if something does definitely affect a person, if it only changes things a tiny amount, who cares? In statistics, an effect size close to 1 or -1 is huge while one close to 0 is tiny. In general, you need to get above a 0.50 to consider it a large effect.
Marijuana 21=-0.129 35=-0.591 42=-0.468
Lower BMI 21=-0.251 35=-0.496 42=-0.194
Depression 21=0.052 35=0.703 42=0.731
So let’s tackle these one at a time.
For marijuana users, while the effect of gamers smoking less always seem to be genuine, it only becomes large for those playing more than 35 hours of games per week, and actually decreases once they start playing more than 42. So that extra hour a day means you suddenly are more likely to light up a joint.
Doesn’t make much sense.
For the lower BMI score, there is again a high chance the effect is real, but it only comes anywhere close to being a large effect after 35 hours of game play and all but disappears at all other levels. So once again, if you play 35 hours of games per week, something about that correlates to a lower BMI, but add just an hour a day or subtract two hours a day and the effect size diminishes greatly.
Doesn’t make much sense.
For depression, there is absolutely no correlation between playing a lot of games and becoming depressed for those playing 21 or 35 hours per week. But once you get up to six hours or more per day, there’s suddenly a very good chance that there’s a real effect and that it’s a big one.
That does make sense.
You might expect teens shut in their room for more than one-third of their day to develop depression. Whether the gaming is a cause or a symptom is, of course, up for debate. But typically people need real human interactions in a face-to-face manner to stave off mental health issues.
With only 27 teens represented, however, this study is by no means definitive. What it means is that, hey, there might actually be something to this, maybe some other folks should look into this more by designing a study to look at this specific question.
And seeing as how the data was taken from participants who were, on average, 15 in 1994 (making them 35 today), I’m guessing they could easily find enough participants in today’s youth who spend more than 42 hours per week playing games to conduct a study.
As for the university’s news office making claims about lower marijuana use and BMIs for teens that game a lot, I don’t buy it. I do, however, buy their use of the term “mouse potato.” That’s clever. But it also dates the study.
Who uses a mouse to play games anymore?
The study, “Long Term Effects of Video and Computer Gaming Heavy Use on Health, Mental Health and Education Outcomes Among Adolescents in the U.S,” is the thesis of Chennan Liu of the University of Illinois, who is now a professor of social work at Renmin University of China, and will be presented at a conference of the Society for Social Work and Research early next year.