Studies show correlations, not causations
October 16, 2012
Have you ever heard of a study you can’t believe is true or that seems completely random?
Scientists are researching everything these days, trying to find a relation between two things and claim they found it. The way studies are progressing, scientists will soon relate how the amount of paper you print shows you’re more intelligent, the more favorites former Gov. Mitt Romney’s tweets get directly relates to the amount of natural disasters during his years, or how the more coffee you drink the more likely you’ll go blind.
Oh, wait, Harvard did do that.
Harvard University recently released a correlational study through Investigative Ophthalmology & Visual Science (IOVS). The study states researchers found a “positive association between heavier coffee consumption with risk of [exfoliation glaucoma],” or basically there seems to be a likely relationship with coffee and developing glaucoma.
The thing is, these two just have a correlation. There is no proof that coffee is the cause of glaucoma.
“We can’t say that one thing like heavy consumption of caffeinated coffee caused the risk of glaucoma to be greater, especially when there was a history of glaucoma in the person’s family,” said Beverly Henry, associate professor in NIU’s College of Health and Human sciences.
But not everyone understands how studies work, and I wanted to find out first hand how people would react to this one.
In a rare moment of being openly social, I went around the Holmes Student Center cafeteria and talked to people about Harvard’s study. I said that a study was recently released that suggests a relationship between the more coffee you drink and the more likely you’ll get glaucoma, which leads to blindness, and then I asked for my listener’s reactions.
Over half of the students, upon hearing this information, said they’d stop drinking coffee completely or cut back the amount they drink.
Could you imagine if I had said that Harvard released a study on how drinking water increases the radiation in your blood, or that living in light for longer than four hours a day will increase the chance for lung cancer?
I can say any of the above examples and there may be a positive association, but correlation is not causation. We can’t assume coffee will cause people to go eventually go blind if there is no concrete evidence.
But what does evidence matter when people read the study and confuse correlation for causation?
This can really damage a person’s way of life.
So, before a study is published showing how the more steps you take is positively correlated with the amount of failing grades you’ll get in a year, researchers need to pause and see if this information is actually relevant and can actually help someone do better with their life.
On the side of the general audience who gets these studies, we really need to make sure these make sense before we buy into everything published.