A study quoted in the Times Online this week suggests Facebook use is “damaging students’ academic performance.” The study, from the education department at Ohio State University, finds that students who report using Facebook on a daily basis tend to have “significantly” lower grade point averages than people who don’t regularly use the service.
Here’s the problem, though: The research — at least, as it’s quoted in the Times Online — makes the age-old mistake of assuming causation from correlation. Case in point, taken straight from the original article:
“The findings will confirm the worst fears of parents and teachers. They follow the ban on social networking websites in many offices, imposed to prevent workers from wasting time.”
All the study actually found was that, within a small pool of about 200 people, students who said they spent a lot of time on Facebook also tended to spend less time on their academic work. That’s a correlation. A goes along with B.
Jumping to the conclusion that A causes B is a result of assuming information that we don’t know. There’s no reason to believe the Facebook use actually caused the lower grades, as the story suggests (direct quote: “Research finds the website is damaging students’ academic performance”). If Facebook weren’t a factor, the people with the lower grades might very well spend their time on other tasks, online or off, rather than on studying.
The more likely conclusion, then, is that some people simply aren’t motivated to spend the time needed in order to get high grades in their academic work. Whether they spend their time surfing Facebook or eating French fries instead of studying isn’t terribly significant — not based on the information we have here, anyhow. If the survey asked how much time those same students spent watching television or reading blogs, odds are, the ones with the lower grades would also respond with higher amounts.
So, the better headline: “Study mistakes correlation for causation.” Maybe not as sexy of a sell, but it’s the more accurate truth in this instance.