Regression to the mean may have been demonstrated by the Yankees. But they were
still at the top of the heap. This was also true of the home run totals of Sosa
and McGwire which were still above the Ruth record. Somewhat less than last year
but still extreme scores. Pretty good evidence that these entities are from a
different populations than the rest.
I think it also has been interesting that the press has not claimed that any of
the above had an "off year." This is a good example for the archive of stat
examples for use in class.
"Paul C. Smith" <[EMAIL PROTECTED]> on 10/05/99 07:23:46 PM
Please respond to [EMAIL PROTECTED]
To: "'T I P S'" <[EMAIL PROTECTED]>
cc: (bcc: William Wozniak/PSY/UNK/UNEBR)
Subject: Regression to the mean
All -
You may remember at the beginning of the year I mentioned that the
NYYankees' 1999 season would present an obvious example of regression to the
mean. They won 125 games last year, and not surprisingly, will do worse than
that this year. They've won 98 during the regular season, and cannot win
more than 11 more (3 in the current best-of-five series, and 4 in each of
the next two best-of-seven series). That would give them a total of 109, at
the very most (please pardon my math if I miss by a couple - there's a man
on second in the first, Palmeiro is batting, and I'm a little distracted).
No great prognostication, just good old regression in action.
Paul Smith
Alverno College
Milwaukee
*************************
William Wozniak, Ph.D.
Professor of Psychology
Department of Psychology
University of Nebraska at Kearney
Kearney, NE 68849
Ph: 308-865-8235
Fax: 308-865-8980
[EMAIL PROTECTED]
**************************
