Robustness issues for cluster randomised trials
Department of Methodology & Statistics, Faculty of Social Sciences, Utrecht University
Promotores: Prof. dr. J.J. Hox (Utrecht University), Dr. ir. M. Moerbeek (Utrecht University)
Project running from: 1 September 2005 – 1 September 2010
Project financed by: N.W.O.
Cluster randomised trials randomise complete groups to treatment conditions. The estimates of the model parameters and their standard errors are only correct if the chosen statistical regression model includes all necessary fixed and random effects, and if the model assumptions are satisfied. Furthermore, optimal designs for cluster randomised trials depend on the values of certain model parameters, of which the true values must be specified in the design stage. This study researches two questions: What is the robustness of optimal designs and estimation methods? What should be done to correct for an incorrect model or an incorrect guess of the model parameters?
Date of defence: 8 June 2012
Title of thesis: Robustness and optimal design issues for cluster randomized trials