Reply #37: these posts are great

[Neil B Forzod]

I love that kid, he's got spunk. Doesn't know jack about statistics, but if we had a few more people with his persistence and enthusiasm it'd go a long way towards fixing things.

I'm not going to point out all the bad assumptions and their effects on the math; that's already been done (and very effectively) time and time and time again. If someone doesn't know (for example) the difference between sampling and non-sampling error, he should probably just keep any analyses to himself and enroll in a first-year stats course before setting up a web site.

Making sarcastic jabs at people who disagree with you is actually kind of an art form. I do admit I get a guilty kick out of TIA's propensity for talking down to people in a lengthy string of one-to-three-word sentences as though his audience is comprised of small children who are too stupid to understand his vastly superior wisdom. The danger in doing that, however, is that when you make stupid statements yourself, you come off worse than you otherwise might.

Here's a simple, random example. And as a bonus, it's 100% math-free (or as TIA might put it, 100.00000000000000000000000% math-free, because significant digits are apparently something to be ignored, like sampling error), to boot:

Were ALL these pre and post-election polls BIASED?
They ALL confirm that Bush lost.

Well, no. Obviously no pre-election poll can "confirm" that Bush lost, since a pre-election poll is (by definition) taken before a single vote is cast. (Never mind that he picks only polls that support his pre-determined answer, and dismisses all others -- that's beside the point being made here.)

Anyway, I do admire TIA's enthusiasm even if I think his analyses are severely misguided. I wish we had a few more like him and a few less people who concede at the first hint of adversity.

Neil