Friday, October 13, 2006

Invalid Statistics

I get really annoyed when people make a point using invalid statistics.

For example, I used to get this newsletter from a psychologist author about taming your inner brat. She cited a survey where 80% of motorists said they'd witnessed road rage, but only 20% admitted to committing it. She said the math doesn't compute, and either implied or stated outright that people must be committing road rage without admitting it. She was very gracious when I pointed out that the 80% could have been seeing the same 20%. I mean, when someone flips someone else off on the highway, three or four people could witness it, right?

Well, here we go again. Po Bronson had a really good column at the back of Time magazine last week. It was about how the media blows certain things all out of proportion, leading us to believe that all of our kids are spoiled brats who are overscheduled and on panic attack medication about college selection and go back to loaf and live with their parents after college (I'm paraphrasing, of course). [The funny thing is that every example he cited was from past Time articles.] His point was that there are actually very few parents who do too much for their kids, and the media is highlighting a problem in a small portion of the population but applying it to the nation at large.

The problem was, he made his point by saying Baby Einstein only made $200 million last year, while Barbie made $3 billion. This was meant to illustrate that parents are letting their kids play, instead of trying to instruct them into insanity. This false illustration ignores the facts:

  • Baby Einstein targets a range of ages 0 to 2, maybe up to 3 or 4
  • Barbie's target demographic is three to six times that age range, not to mention adult collectors
  • Baby Einstein sells toys, books, DVDs, etc., and that's pretty much it
  • Barbie's name is all over clothes, lunchboxes, bicycles, thousands of products
Extrapolating from statistics in this manner undermines a person's point and makes all their assertions suspect. It drives me nuts.

And I never even took a probability and statistics course.

No comments: