Date of Award
6-18-2014
Document Type
Thesis
Degree Name
Mathematics, MS
First Advisor
Hong Zhou
Committee Members
Debra Ingram; Ferebee Tunno
Call Number
LD 251 .A566t 2014 W5
Abstract
Multiplicity is a well established notion in statistics that arises when simultaneously testing n &ge2 individual hypotheses. To combat this issue, over time, numerous multiple comparison procedures have been developed. Such methods are relied upon heavily in confirmatory clinical trials, both in dose-finding and multiple-endpoint scenarios. The proposed procedure is a hybrid between the data-driven Weighted Holm procedure and the Fallback procedure which relies heavily on the a priori ordering of the hypotheses based on weights. Strong control of the familywise error rate is proven. Both general and previous clinical examples are performed utilizing this new method while comparing the outcomes to those of other techniques. Simulations test the proposed procedure against the Weighted Holm and the Fallback with varying weighting schemes and numbers of hypotheses. The results provide evidence that the proposed procedure outperforms the Weighted Holm in a multitude of scenarios while becoming increasingly competitive against the Fallback for larger n.
Rights Management
This work is licensed under a Creative Commons Attribution 4.0 International License.
Recommended Citation
Wolf, Jared Ray, "A Modified Weighted Holm Procedure for Multiple Comparisons" (2014). Student Theses and Dissertations. 762.
https://arch.astate.edu/all-etd/762
