Stepwise Procedure for Toxicological Assessment Based on Ratio of Mean Differences for a Normally Distributed Data
Michael J. Adjabui *
Department of Mathematics, CK Tedam University of Technology and Applied Sciences, Navrongo, Ghana.
Jakperik Dioggban
Department of Statistics, CK Tedam University of Technology and Applied Sciences, Navrongo, Ghana.
Nathaniel K. Howard
Department of Statistics, University of Cape Coast, Cape Coast, Ghana.
*Author to whom correspondence should be addressed.
Abstract
We propose a new stepwise confidence set procedure for toxicity study based on ratio of mean difference. Statistical approaches for evaluating toxicity studies that properly control familywise error rate (FWER) for difference of means between treatments and a control already exist. However, in some therapeutic areas, ratio of mean differences is desirable. Therefore, we construct stepwise confidence procedure based on Fieller's confidence intervals for multiple ratio of mean difference without multiplicity adjustment for toxicological evaluation. Simulation study revealed that the FWER is well controlled at prespecified nominal level α. Also, the power of our approach increases with increasing sample size and ratio of mean differences.
Keywords: Balanced design, fieller, FWER, MED, stepwise procedure, power