Wednesday, August 24, 2005

A travesty of Statistics

Although rants about statistics are more rightly the province of my sister Rachel or possibly Mike, I could not help but have an apopletic fit over a radio broadcast I heard yesterday.
Dennis Prager, a well known radio host, decided that he would not discuss the news today. (Apparently, he does this sometimes) Instead, he wished to share with the audience a scientific observation that he had recently made- that well-behaved children may be worse off psychologically when they grow up because they never had a chance to vent all that naughtiness. This was not just an idle theory- it was evolved based on his "thinking about him and his children and people he knows". I do not presume to comment on the theory, although I suspect it was evolved based on there being a couple of people who turned out worse or better as adults, without regard to all the people who grew up perfectly predictably, as the latter don't stand out so much. But that is neither here nor there.
Mr. Prager, being a man of science, was not satisfied with his own narrow sample of evidence. No, he decided to strenuously test his theory by having his listeners who had experiences that fit with his idea call up and talk to him about it. And sure enough, a man called up and said that he had been a naughtier child than his wife and now he was more psychologically stable.
I could not stomach much more of the show and anyway I had places to go, but I am pretty sure he continued to discuss this issue for over an hour.
Even if he had picked a random sampling among the callers, and even had the callers called not only because their experiences supported him, and even had they not been totally oblivious to cases that would contradict them because those are too obvious to notice, and even were they capable of truly evaluating people's psychological stability accurately, and even had there been enough callers to be statistically significant, this would have been a stupid system that at best could have established a correlation between obedience as a child and later psychological problems, failing to notice any outside causations (I bet abused children are obedient and later psychologically scarred- so Mr Prager's little theory has at least one strand of truth.)
The more disgusting part of all of this was how Mr. Prager, no doubt an intelligent man, did not notice the stupidity of his own system (Unless he was just doing it to fill up air time), nor did a single one of his imbecil callers. It makes you wonder, it truly does. What is wrong with the mathematics education in this country?


Rachel said...

Why didn't you call up and tell him? By the way, I think they've found personality very early in life is associated with adult personality. And it's positively correlated in virtually everything. So his statistics are wrong.

MiriK said...

if she had called him up and told him, he wouldn't have put it on the air unless he could've found someway to make her sound stupid and wrong. radio hosts can be like that...and hey Tobie, I've been pretty well-behaved. do you think I'm mentally stable, or unstable?

Tobie said...

Miri- I think we all know the answer to that one. And anyway, I wouldn't have wanted to go on air becuase unfortunately my voice on the radio, tapes, or phone sounds like that of a two-year-old, making it hard to wither anyone with my intellectual/statistical scorn.