Okay, it’s actually a hypothesis. But theory sounds better.
I do a fair amount of driving around and the other day I noticed something significant. Driving a convertible sportscar with the top down creates baldness in men. How do I know this? My amazing powers of observation tell me this is true. When I see a man driving a convertible with the top down he is usually bald. Ergo, driving with the top down causes baldness.
Flawed hypothesis, you say? Well yes, yes it is. On a number of counts.
I use my flawed hypothesis here to highlight the challenge of correlation versus causation.
In this example I have implied that driving a convertible with the top down causes baldness. And clearly, there are many examples to prove my hypothesis. Every third person driving a convertible is bald or balding so it must be true, right? At 33% we’re talking some pretty good chances here. I mean really, would you drive through an intersection if every third car coming from the right didn’t have to stop?
Yeah, not so much.
It is true there is a correlation between driving a convertible with the top down and baldness but I’m pretty sure the correlation has more to do with men at the typical age of balding are in a financial position to drive a convertible sports car.
But that’s just a guess.
My point is this. There are a lot of false assumptions flying around asserting that because X happened during Y that Y caused X. Maybe it did. But maybe they just happened at the same time for other unrelated reasons.
This is important to keep in mind with social media allowing us to propagate bad information just as quickly as good. (Does bad information actually spread faster or is that just how it seems to me?) As memes show up that agree with our biases and we’re tempted to like, share, retweet, and so on please give some thought to whether the ideas you may send are actually accurate and true. And remember, correlation does not equal causation.