divisionbyzero: (Default)
divisionbyzero ([personal profile] divisionbyzero) wrote2003-04-09 12:09 am

Regressing now.

Given that the name of this journal is DivisionByZero, I figured I'd rant about something math-related that bugged the crap out of me in my statistics class tonight. (Oh, and if anyone can tell me how to make it just say 'DivisionByZero' up there, let me know.)

So tonight in my stat class, we talked about regression towards the mean. My stat professor is a thorough, if patently boring, guy who does a good job of explaining statistical concepts accurately. He seemed to really like the idea of regression towards the mean, instilling a vigor in his delivery usually reserved for his discussion of dog shows and basketball.

Dr. L had 2 pages worth of examples of where regression towards the mean happens in the real world. Most of them I liked, but I took issue when he said the following: "Between 1969 and 1988, no team had repeated as champions in the NBA. Regression towards the mean."

ARGH!

Now, I'm all for more statistical speak in daily life, and for combating the misuse of statistics and misinterpretation of numbers, but I'm sorry, this is a little much. If we reduce the idea of teams not repeating as 'regression toward the mean', how many factors do we ignore? Essentially, we're reducing all the other variables like player turnover null and void. He cited baseball too... do we really think the Yankees are subject to the same statistical processes that the rest of the world is? Try telling Steinbrenner that.

And just when I thought I was being overly critical, especially since this class doesn't seem too sports-savvy, I hear two women walking out of class talking about how the Pirates' hot start is destined to fade because of... regression toward the mean.