An exact Cochran-Armitage test for trend when dose-response shapes are a priori unknown

authored by

Markus Neuhäuser, Ludwig A. Hothorn

Abstract

The Cochran-Armitage test is a widely used test for trend among binomial proportions of a dose-response relationship. This test requires preassigned fixed dose scores. Equally spaced scores were usually suggested if the dose-dependent shape of the binomial proportions is a priori unknown. Another approach is the construction of a maximin efficiency robust test. We recommend a combination of Cochran-Armitage tests based on different scores and the use of the maximum of the test statistics for a new test. Simulation results suggest that this combined test is superior to both the single test with equally spaced scores and the maximin efficiency robust test. The methods are applied to data of a toxicity and a tumorigenicity study with a stratified design.

Organisation(s)

Department of Biostatistics

External Organisation(s)

Altana Pharma

Type

Article

Journal

Computational Statistics and Data Analysis

Volume

30

Pages

403-412

No. of pages

10

ISSN

0167-9473

Publication date

07.07.1999

Publication status

E-pub ahead of print

Peer reviewed

Yes

ASJC Scopus subject areas

Statistics and Probability, Computational Mathematics, Computational Theory and Mathematics, Applied Mathematics

Electronic version(s)

https://doi.org/10.1016/S0167-9473(98)00091-7 (Access: Closed)