Premium
TWO‐STAGE PROCEDURES FOR TESTING HOMOGENEITY OF MEANS AGAINST ORDERED ALTERNATIVES IN ONE‐WAY ANOVA WITH COMMON UNKNOWN VARIANCE
Author(s) -
Marcus Ruth
Publication year - 1980
Publication title -
australian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.434
H-Index - 41
eISSN - 1467-842X
pISSN - 0004-9581
DOI - 10.1111/j.1467-842x.1980.tb01174.x
Subject(s) - homogeneity (statistics) , percentile , analysis of variance , statistics , variance (accounting) , monte carlo method , one way analysis of variance , mathematics , multiple comparisons problem , stage (stratigraphy) , econometrics , computer science , paleontology , accounting , business , biology
Summary In this paper we present two‐stage procedures for testing the equality of normal means against ordered alternatives in one‐way ANOVA with common and unknown variance. The procedures proposed are analogous to Williams's and Bartholomew's single‐stage methods. A table of percentiles needed for implementation is provided. Some Monte Carlo results for estimating the power of both procedures are given.