Because of my fondness of mathematics, it pains me to my very soul when I see it misrepresented and abused by people in the media. A recent “Teens On Facebook” study has got me hopping mad, and I honestly think it is making people stupider to have articles like this out there.
The article, entitled Teens On Facebook And Social Media Sites More Likely To Drink, Smoke And Use Drugs: Study, tells us about a study that was done that shows that “teens who use Facebook and other social media outlets are five times more likely to smoke cigarettes, three times more likely to drink alcohol and twice as likely to smoke pot than teens that don’t use social networks.”
Before I continue, let’s take a moment to understand what that kind of language means, when talking about statistics. It means that knowing that one factor exists allows you to make the prediction that the other factor is more likely to be true. It DOES NOT mean that doing one thing makes you more likely to do the other. All that this result means is that if you know that a person uses Facebook, then you can predict that the odds are greater that they also drink and smoke. Is that difference clear? Doing A doesn’t cause you to be more likely to do B. It’s simply that knowing that A is true means that the odds are greater that B also happens to be true. That’s what “five times more likely” means.
Alright, now back to the news article. It spends one paragraph talking about the fact that “some researchers on the study” admit that it reveals nothing at all about causation. One paragraph! And whereas you might suppose that the article STOPS SPECULATING after that admission is made, you would be wrong: it spends the next seven paragraphs speculating about possible theories about how using facebook might cause drinking (e.g. seeing pictures of drunk people online makes it seem ok, etc.).
In my mind, this kind of gross mis-representation is tantamount to lying to the public. At the very least, it is actually harming education in America.
Anyone who has taken the most basic Statistics 101 class knows that when A is correlated with B, there are three possibilities:
- A might cause B
- B might cause A
- there might be a third factor C which causes both A and B at the same time
And even if you haven’t been taught this formally, where is the common sense? Where is the little voice in the head of whomever is writing this article, that says…
“Wait a minute! Maybe people who are generally extrovert will both be more likely to go on facebook and be more likely to drink! Maybe the two don’t actually influence each other at all!”
It’s just one article, I know. Nothing to get worked up about.
But it’s just one more grain of sand on the hill, one more case where it has been folded into our public consciousness and perpetuated a misunderstanding of what statistics mean. It has re-enforced that idea that it “might” be valid (even though “some” researchers disagree) to conclude that just because drinking and facebook using go together, some person’s speculation about facebook causing drinking has become “validated” through science. It goes on the junk-heap of such similar agenda-motivated “proofs” as the idea that looking at fashion magazines makes girls anorexic. (Nobody apparently ever thought that anorexic girls might be drawn to read the magazines because of their pre-existing obsession with their weight.)
And in the end, this whole problem becomes one more prick, one more prod, in the public consciousness that “science can’t be trusted” or “studies don’t really prove anything.” It just ends up being more fuel for the anti-science fire in our broader society.
The problem isn’t science. The problem is articles like this, that completely mis-read what the science means, for the sake of sensationalism and propaganda.
And Duskdemon, the SA article is another weird one. Maybe the article was garbage and didn't provide the relevant details from the research, but they didn't give you much to go on there. So narcissists and those with low self-esteem both use Facebook an hour a day, post pictures, write on their walls. I bet one would find a pretty good correlation between, say, Facebook employees and those with low self-esteem on that basis then too. Or 20-somethings and narcissists. Or whatever and whatever. Seems to me that if you're trying to establish a linkage between narcissism and low self-esteem it'd be a lot easier to, oh I don't know, just collect data from the psychologists who are diagnosing these people as having narcissism or low self-esteem and calculate the statistics on both traits being found in the same person.
[ sorry 2 posts b/c blogspot didn't care for how much I was going on and on in 1… 😉 ]
>Wow. Well, first, nice article, Greg – as usual! Agreed completely that articles (and TV, etc!) frequently are written this way and encourage the reader to make the jump to infer the causation, even if the article itself doesn't call out A as causing B. In this case it's even worse since the source article from Columbia actually *does* claim that Facebook causes drinking:
"Continuing to provide the electronic vehicle for transmitting such images [on Facebook of teens drinking] constitutes electronic child abuse." Images causes child abuse. Wild stuff. Way to go, Columbia! In doing a little digging, I found this interesting article on Columbia's own site:
http://www.columbia.edu/cu/norml/cproject.html
So 12 years ago, another organization within Columbia wrote to the University's Senate to urge them to investigate this oddball CASA organization for hiding politically-motivated messages behind scientific "evidence." Not that NORML isn't politically-motivated themselves, but still Columbia apparently doesn't mind allowing their respected name to be associated with faux research. If any organization should hold themselves to a higher standard…
As a mentally decaying member of America, I'm glad to have read your thoughts because they slapped some of the sense I used to have back in me. They also remind me of my admiration for mathematics.
Well said. I read a study in Scientific American http://www.scientificamerican.com/article.cfm?id=status-update-im-so-glamorous that dovetails. Is it incorrect in the same manner when it's said in the end "the question becomes, ‘Can Facebook help raise self-esteem by allowing patients to talk to each other and help each other in a socially interactive environment?’ I don’t think it’s necessarily a bad thing that people with low self-esteem use Facebook."??