A Note on Two-Sided Distribution-Free Treatment versus Control Multiple Comparisons
Journal of the American Statistical Association · 1984 · Volume 79 · Issue 385
DOI: 10.1080/01621459.1984.10477086
Abstract
Distribution-free treatment versus control multiple comparison procedures using joint ranking and pairwise ranking are compared. The pairwise ranking method is extended to allow for all contrast comparisons among the treatments and the control. The author shows by counterexample that the joint ranking method does not control the maximum type 1 error rate.
Enable JavaScript for citation tools and additional paper interactions.