Premium
Probleme der Erweiterung der Prüfverteilungen des Maximum‐Modulus‐Testes und des DUNNETT‐Testes
Author(s) -
Schmerling S.,
Renne J.,
Lenze W.,
Bätz G.
Publication year - 1981
Publication title -
biometrical journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.108
H-Index - 63
eISSN - 1521-4036
pISSN - 0323-3847
DOI - 10.1002/bimj.4710230106
Subject(s) - mathematics , statistics , combinatorics
Abstract SCHULZ (1976) gave formulas for the approximate computation of the percentage points needed for the comparison of sample means by the Maximum‐Modulus‐Test resp. by the DUNNETT‐Test. Our investigation shows that these formulas are also practicable if the number of the mean differences k is greater than 20. The maximum difference between the exact values and the approximate values of the percentage points amounts to 0,05 in the case of the Maximum‐Modulus‐Test with k >20 and degrees of freedom v > k for a level of error a = 0,05 resp. a = 0,01. In the case of the DUNNETT‐Test this maximum difference amounts to 0,02 for 20 < k < 50 and a = 0,05. Besides it can be here supposed that the approximate percentage points are also practicable for both k >50, a = 0,05 and k >20, a = 0,01.