To: EDSTAT-L@JSE.stat.ncsu.edu Subject: Bell Curve surprise Date: Wed, 16 Nov 94 11:02:22
From: Bob Wheeler
Subject: Bell Curve surprise
Date: Wed, 16 Nov 94 11:02:22 EDT
> To me, the startling truth revealed in "The Bell Curve" is
> that intelligence test scores are not a bell curve! The only
> frequency distribution curve I could find in the book while
> perusing it was of SAT scores, and was skewed to the right.
> That is, most scores were "below average," pace Garrison
> Keillor, because there was a long tail to the right,
> causing the mean to fall to the right of the median.
You shouldn't overreact to this by attempting to draw strong
conclusions about human ability. It is in large part due
to the fact that test scores are bounded below -- one cannot
have negative scores, and a fortiori feeble minded people
seldom take IQ tests. The "bell curve," that is used as the
model is unbounded in both directions -- hence it is not
a completely satisfactory model. Many physical measurements
exhibit similar properties due to the fact that they are
only defined for positive values.
E-Mail Fredric L. Rice / The Skeptic Tank