answered prayer, bell curves, ice cream, and commodes

"George M. Rimel III" (GRimel@compuserve.com)
Wed, 23 Jul 1997 23:47:10 -0400


>>which to me implies a Gaussian normal
distribution, yet you have a finite number of samples in your example,
and none can possibly be negative, and there is certainly an upper
bound also -- the tallest person in the sample space, so the normal
distribution is replaced by a binomial distr5ibution.  If you have an
infinite (or near infinite) number of samples, with a lower bound but
no upper bound (such as the speed of molecules in a gas), you would
use a Poisson distribution.  The computation of the first moment, or
mean (colloquialloy called the "average"), is no longer just the sum
of the samples divided by the number of samples.<<

"Up is down, left is right, and back is straight ahead"  Cain

>>Consult a good college statistics text for the details...<<

Can you say oxymoron??