Don't Trust 50,000.



If someone told you that 50,000 people were murdered in ritual Satanic sacrifices every year in America, would you think that sounded frightening, or ridiculous? This weekend's On the Media featured a pair of interesting stories about the use of numbers in journalism, showing how fuzzy they are. Reporters love to put statistics in their stories -- it gives the stories an authority and precision they wouldn't have otherwise. And reporters almost never have the time or expertise to investigate whether the source of the numbers produced them in a defensible way; what they tend to do is evaluate the source (i.e., "This study was done at a major university") and use that to decide whether to use the figures the source produced.

That's not a bad way of judging whether you can trust a number, since it would be impossible for each of us as news readers to make an independent evaluation of every statistic we encounter on a daily basis. Unfortunately, some numbers take on a life of their own and end up getting repeated so often that not only do people stop questioning them, they don't remember where they first heard them. The second of the radio show's stories is about the strange number 50,000, which seems to have kind of a magical power to find its way into crime stories when the real figure is difficult to determine. One person quoted in the story calls 50,000 a "Goldilocks number" -- it's not too small, like 200, but not absurdly large, like 10 million. It's just about the right size to sound both plausible and significant on a national basis. So it pops up as an estimate in a lot of places, when describing both things we can get a precise read on (e.g. how many people die in traffic accidents) and things we're basically guessing at (e.g. how many sexual predators are online at any given moment).

So the next time you hear the number 50,000, you might want to exercise your skepticism. You can listen to the On the Media stories here and here.

-- Paul Waldman

You may also like