Rolling Stone

I gather analysis isn't their strong suit. The sex scandal at Duke story has a few oddities that suggest that the reporter hung around with campus elites. Duke has 6500 undergrads, and they can't all hang out at Shooters. The author hung out with frats and sorors and talked about athletes. It makes for a juicy story, but I'd be more convinced if he'd talked to some chemistry majors too.

And they've an innuendo story about stealing elections . I judge that voting without a paper trail is a recipe for trouble. But I'm afraid that wishful thinking and skewed polls are not useful evidence of wrongdoing.

Both stories use biased sampling to try to reach conclusions (or more accurately, to give impressions and let us reach conclusions). I suppose reporters (not generally the best-educated among us) are never taught anything about statistics, and the editors seem to be no better. And I suppose a magazine devoted to the entertainment world isn't so used to looking past impressions to look at the details: impressions are a lot of what entertainment is about.

Rolling Stone is hardly the only offender (they're just handy). Think about the recent stories about the number of lives hospitals saved with more careful procedures in the last year and a half--and ask how many accidental deaths there were before, and start getting suspicious.