Conservativeness of global tests for bivariate binomial and Poisson data
The global tests of Bonferroni, Rüger, Hommel and Simes are compared concerning deviation between actual and nominal significance level, when individual hypotheses are tested via sign, Fisher or Poisson tests. Despite remarkable conservativeness due to discreteness the results confirm applicability of Simes-based multiple test procedures to discrete data.
Year of publication: |
1999
|
---|---|
Authors: | Krummenauer, Frank |
Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 41.1999, 4, p. 407-412
|
Publisher: |
Elsevier |
Keywords: | Global hypothesis Improved Bonferroni procedures Actual level |
Saved in:
Saved in favorites
Similar items by person
-
Representation of multivariate discrete distributions by probability generating functions
Krummenauer, Frank, (1998)
-
A note on Rüger's test for discrete data
Krummenauer, Frank, (1999)
-
Limit Theorems for Multivariate Discrete Distributions
Krummenauer, Frank, (1998)
- More ...