Saturday, August 13, 2011

Mourning Morning Smokers

Since people will be reading the Tea leaves of the Ames straw poll today, it’s a good time to remind ourselves that people are godawful at thinking about statistics.

This Gawker piece is a relatively simple example of one mistake people make with stats all the time – sometimes in a subtler and more complex fashion than we see here – namely, thinking that something that is merely an indicator of a broader pattern is the specific cause of that broader pattern (e.g., people who subscribe to fitness magazines live longer – so give everyone fitness magazines and you’ll see lifespans increase!).

People who can’t resist a morning smoke are more likely to get lung cancer, the article explains – so Gawker concludes that waiting a few hours before the first cigarette is the key to survival, even though it is far more likely, of course, that needing a cigarette (or anything else for that matter) immediately is simply a good indicator that you are more addicted or less self-disciplined than other users (and future attempted-quitters).  It’s not as if the time of day is extra-deadly. 

As is typical with Gawker and all writers in this hip-irono-liberal vein (which is pretty much the norm in NYC), they sound like they sort of think they might maybe could be joking but aren’t probably sure so much, the important thing presumably being that they’re cooler than you are and know things that you don’t even when they sound stupid, so there (but in fact, they likely know less – about a great many things – so don’t be intimidated by the hipness).

Similarly, I trust I’ll be OK even if late night trips to Crif Dogs are revealed to be a good statistical indicator of obesity.

No comments: