Let’s talk about logic.
A Yahoo headline proclaims: “New study shows school sports improve grades.” The article says:
…LAUSD has discovered a striking correlation between students’ participation in interscholastic athletics and their performance in both attendance and in the classroom. According to the study, the 35,000 student athletes in LAUSD attended an average of 21 more days of school per year than their counterparts, while they also sported GPAs some 0.55 to 0.74 points higher than non-athletes.
How should we interpret this finding? The article makes it clear:
“[The study statistics] prove what has generally been assumed, that participation in high school athletics, on average, positively enhances the student’s academic progress in comparison with the rest of the student body,” LAUSD commissioner of athletics Barbara Fiege said…
(Emphasis mine.)
Simple, right? Students who play sports get better grades. So if you want to boost your kid’s GPA, sign them up for basketball.
Let’s apply this logic elsewhere. A recent study discovered a striking correlation between people opening their umbrellas, and rain falling. So if you don’t want it to rain, ditch the umbrella.
You see the problem.
Yes, it could be that sports improve grades. It could also be getting good grades makes a student more confident, and thus more likely to play sports. A third option, most probable of all, is that some extra factor (like self-discipline or strong parenting) causes both the sports commitment and the good grades.
You’d need more analysis, or more research, to find the underlying cause-effect relationship. But the article doesn’t even consider these other possibilities. It leaps straight from correlation to causation, unaware it is leaping.
This way of thinking is very common. News stories do it often. Even scientists do it. I did it myself, for a long time. The cure for this logical fallacy can be summed up in five words, which are enshrined in their own Wikipedia page:
Correlation does not imply causation.
This simple piece of knowledge will open your eyes to all sorts of misleading studies and claims. These days, writers say “studies show…” to justify almost anything, waving the phrase like a magic wand, conjuring in the name of science. Well, science doesn’t lie, but your brain is the ultimate con artist. Tread carefully.
To be fair to the media, here’s an NBC story – posted just yesterday – which is much more careful:
Over 3 hours of television a day may make kids more antisocial
British researchers…found those who watched television longer than three hours per day were more likely to develop antisocial behaviors such as fighting, stealing or bullying. […]
The study’s authors concede the link between conduct problems and television time may be indirect, and a factor of the added inactivity, sleeping difficulties or impaired development associated with too much screen time.
The findings…showed an association and not cause and effect. The researchers said more work is needed to determine a causal link, but in the meantime, they suggest a cautionary approach to heavy television use in young children.
(Emphasis mine.)
Kudos to the author of this article (Ryan Jaslow). Granted, these subtleties may be lost on the average reader, but the media can only do so much. Readers have to meet them halfway.
Yea, verily: go forth and be smart.
My recent favorite is that there is a correlation between countries whose citizens consume more chocolate and the amount of Noble Prize winners from that country.Basically chocolate makes you more likely to win a Nobel Prize.
Seems legit. I’m willing to test this hypothesis.
Yea, this is something that bothers me a lot in our media. We often get presented with correlations that are set up to imply causality without any proof.
Reminds me of a joke.
“I took a statistics class.”
“Was it good?”
“I used to think correlation implied causation. Now I don’t.”
“So the class helped?”
“Well, maybe.”
Pingback: Do You Believe in Dragons? | The Happy Logophile
Lol at the statistics joke.
This logical fallacy is an easy one to fall for and unfortunately pervasive throughout the media. I used to think most things with antioxidants in them were good for me until I read that it’s not quite so simple. I like this TED talk by Molly Crockett on debunking nuerobunk.
Unfortunately, WordPress puts comments in moderation hell if they have two or more links in them. So your comment didn’t show up till I got a chance to approve it.
Thanks for the TED link though! I love any excuse to watch a TED talk. I’ll check it out soon.