Usage Note 23002: In Distribution(Y), why do the p-values from "Test for Specific Distribution" and "Test for Distribution" differ when the same distribution is tested?
The p-values from Curves : Test for Specific Distribution and Curves : Test
for Distribution differ because the tests are meant to be used under different
circumstances. For the Test for Specific Distribution, the parameters of the
distribution are assumed to be known. For the Test for Distribution, the parameters are
unknown and estimates are used. As discussed by D'Agostino and Stephens (1986),
although the same statistics are used, the distributions of the statistics for
these two tests are very different, resulting in different p-values. In
the SAS/INSIGHT User's
Guide, different references are given for
these two tests. Do not use Curves : Test for Specific
Distribution when you are providing only estimates of the distribution
parameters.
D'Agostino, R.B. and M. A. Stephens. 1986. Goodness of Fit Techniques. New York: Dekker.
Operating System and Release Information
| SAS System | SAS/INSIGHT | All | n/a | |
*
For software releases that are not yet generally available, the Fixed
Release is the software release in which the problem is planned to be
fixed.
| Type: | Usage Note |
| Priority: | low |
| Topic: | Analytics ==> Distribution Analysis
Analytics ==> Exploratory Data Analysis
|
| Date Modified: | 2007-10-11 16:52:09 |
| Date Created: | 2002-12-16 10:56:44 |
