Re: test of normality

Bob O´Hara said
(Posted: Jul 12, 2005 3:06 AM)


<You can ignore the K-S test: it'll be significant, because of <the huge sample size.

<Bob O'Hara
<Department of Mathematics and Statistics
<P.O. Box 68 (Gustaf Hällströmin katu 2b)
<FIN-00014 University of Helsinki
<Finland


With ALL due respect to everybody I would like to add a these remarks:

1_Let be H0 the hypotheses: The population is normal (under the K-S test): H0 acceptance.

2_From a very large sample of a no-normal Population H0 will be rejected almost certainly (the skewed Population of the wages in a Country – very few earning very much, the remainder much lesser.

3_ Mixtures
If to the normal Population is added values that are of a different Population the greater the size the lesser the probability to accept H0.

This is a rather poor contribution. I expect that other Readers post in this matter. We will learn a lot, surely.


_____________licas_@xxxxxxxxxxx
.

Loading