A maximin criterion for the logistic random intercept model with covariates

  • Ouwens M
  • Tan F
  • Berger M
  • 3

    Readers

    Mendeley users who have this article in their library.
  • 19

    Citations

    Citations of this article.

Abstract

A maximin criterion is used to find optimal designs for the logistic random intercept model with dichotomous independent variables. The dichotomous independent variables can be subdivided into variables for which the distribution is specified prior to data sampling, called variates, and into variables for which the distribution is not specified prior to data sampling, but is obtained from data sampling, called covariates. The proposed maximin criterion maximizes the smallest possible relative efficiency not only with respect to all possible values of the model parameters, but also with respect to the joint distribution of the covariates. We have shown that, under certain conditions, the maximin design is balanced with respect to the joint distribution of the variates. The proposed method will be used to plan a (stratified) clinical trial where variates and covariates are involved. © 2004 Elsevier B.V. All rights reserved.

Author-supplied keywords

  • Covariates
  • D-optimal design
  • Logistic random intercept model
  • Maximin design
  • Relative efficiency

Get free article suggestions today

Mendeley saves you time finding and organizing research

Sign up here
Already have an account ?Sign in

Find this document

Authors

  • Mario J.N.M. Ouwens

  • Frans E.S. Tan

  • Martijn P.F. Berger

Cite this document

Choose a citation style from the tabs below

Save time finding and organizing research with Mendeley

Sign up for free