* t. n. southton > Gauss's Curve (for lack of a better word for it) is described by this:
"The normal distribution"? > > What assumptions did Gauss make in fitting this curve to random > numbers, and when are Gaussian distributions valid? For example, if I > did a frequency plot of the digits in Pi expanded out to a million, it > probably won't be Gaussian, but a line with slope zero.
I am a complete layman with respect to this, but I once asked a statistican and he answered something like (any errors are mine):
1. Your data are normal distributed because you decided they were in your research model.
2. The distribution of mean values are provable normal distributed.
3. When the outcome of an experiment is depended on N independent factors and N is large, the random variable measuring the outcome is normal distributed.
(Point 1 is like the drunk standing at night under a lamp looking for something. "What are you looking for?" asked a passer by. "My keys." "Where did you lose them?" "Over there," the drunk answered. "But why are you looking here then?" "Because it is more light here.")
-- Jon Haugsand Norsk Regnesentral, <mailto://email@example.com> Tlf: 22852608/22852500, Fax: 22697660, Pb 114 Blindern, 0314 OSLO