Closed test procedures for the comparison of dose groups against a negative control group or placebo

TL;DR

This paper introduces two variants of closed testing procedures for comparing dose groups to a control, offering alternatives to the Williams trend test, with accompanying R software implementations.

Contribution

It proposes novel closed testing methods using global Williams tests and pairwise contrast tests for dose group comparisons, expanding statistical tools in this area.

Findings

Two variants of closed testing procedures are developed.

R software implementations are provided.

The methods offer alternative approaches to the Williams trend test.

Abstract

Dose groups are compared with a control assuming an order restriction usually by the Williams trend test. Here, as an alternative, two variants of the closed testing procedure are considered, one where global Williams tests are used in the partition hypotheses, and another where pairwise contrast tests are used for this purpose. Related R software is provided.