A Comparison of Three Approaches for Constructing Robust Experimental Designs
Vincent K. Agboto1, William Li2, Christopher Nachtsheim2
1Family and Community Medicine, Meharry Medical College, Nashville, TN, United States; 2Operations and Management Sciences, University of Minnesota, Minneapolis, TN, United States

While optimal designs are commonly used in the design of experiments, the optimality of those designs frequently depends on the form of an assumed model. Several useful criteria have been proposed to reduce such dependence, and efficient designs have been then constructed based on the criteria, often algorithmically. In the model robust design paradigm, a space of possible models is specified and designs are sought that are efficient for all models in the space. The Bayesian criterion given by DuMouchel and Jones (1994), posits a single model that contains both primary and potential terms.

In this article we propose a new Bayesian model robustness criterion that combines aspects of both of these approaches. We then evaluate the efficacy of these three alternatives empirically. We conclude that the model robust criteria generally lead to improved robustness; however, the increased robustness can come at a significant cost in terms of computing requirements.

Keywords: Bayesian designs; Model-robust design; D-optimality; Supersaturated design.

Biography: Dr. Agboto is an Assistant Professor in the Department of Family and Community Medicine at Meharry Medical College and the Director of the Biostatistics at the Meharry Medical College in Nashville (Tennessee, USA). His current areas of research interests include the development of new Bayesian methodologies to design experiments and novel ways to design health disparity epidemiological studies. He is an adjunct assistant professor in the department of biostatistics at the Vanderbilt University Medical Center. Dr. Vincent Agboto currently teaches biostatistics, research design and data management courses at the Meharry Medical College and ivolved in health disparity research with faculty at Meharry, Vanderbilt, Emory. He has published peer reviewed articles in Bayesian optimal designs and health disparities and has coauthored technical reports.

He is a member of the American Statistical Association, an elected member of the International Statistical Institute, a member of the American Public Health Association.