, Volume 74, Issue 2, pp 268-285
Date: 20 Nov 2012

Many to one comparisons in a longitudinal binary data setup

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access

Abstract

Motivated by an example we discuss the problem of comparison to control in a scenario where we have repeated binary data from a large number independent individuals under each comparison group. We use the generalized quasi-likelihood (GQL) estimators for the comparison group effects including the effect of the control group to define a single step and step-down method for performing many-to-one comparisons. We discuss the asymptotic properties of our method and provide ways of computing the critical values. We follow up with simulations for Type I error and marginal power and show that our proposed method performs well in practice.