biology statistics

Jan Vorbrueggen jan at neuroinformatik.ruhr-uni-bochum.de
Wed Oct 4 19:01:50 EST 1995


In article <cdrake-0110951606400001 at ppp17.tcs.tulane.edu>
cdrake at mailhost.tcs.tulane.edu (Chris Drake) writes:

   I was wondering if there is a fundamental reason why seemingly all physical
   data are normally distributed.  It seems that all biological measurements
   are normally distributed data: height, weight etc.

That's an effect of the central limit theorem of statistics. It says that the
sum of a sufficient number of arbitrary probability distributions will always
look like a normal (Gauss) distribution. For instance, 3-6 iterated
rectangular (stepwise) distributions are quite sufficient to give a very good
approximation of a Gaussian.

	Jan




More information about the Comp-bio mailing list