On the Large Sample Optimality of Sequential Designs for Comparing Two or More Treatments

Abstract The present paper deals with optimal (in Kiefer's sense) response-adaptive designs for parametric inference on v ≥ 2 treatments. Sometimes (e.g., for nonlinear models) a sequential estimation procedure combined with an adaptive experiment suggests itself as the “natural” best design. One of the questions is whether, since we proceed sequentially, we should infer conditionally on the design. Another question is whether such an adaptive design is really optimal for the chosen type of inference. The main purpose of this paper is to give proofs of the asymptotic optimality for inferring both conditionally and unconditionally of a large class of such designs, incorporating response-adaptive randomization as well. The asymptotic optimality of the Maximum Likelihood design, namely that based on the step-by-step updating of the parameter estimates by maximum likelihood, is proved for responses belonging to the exponential family. Under this procedure the MLEs retain the strong consistency and asymptotical normality properties. Furthermore, such properties still hold approximately for suitable inverse sampling stopping rules.

[1]  H. Robbins,et al.  A Sequential Analogue of the Behrens-Fisher Problem , 1967 .

[2]  H A Wendel,et al.  Randomization in clinical trials. , 1978, Science.

[3]  Jeffrey R. Eisele.  The doubly adaptive biased coin design for sequential clinical trials , 1994 .

[4]  G. Jasso.  Review of "International Encyclopedia of Statistical Sciences, edited by Samuel Kotz, Norman L. Johnson, and Campbell B. Read, New York, Wiley, 1982-1988" , 1989 .

[5]  C. C. Heyde,et al.  Estimation of Parameters from Stochastic Processes , 1980 .

[6]  D. Titterington,et al.  Inference and sequential design , 1985 .

[7]  Connie Page,et al.  An adaptive randomized design with application to estimation , 2001 .

[8]  Anastasia Ivanova,et al.  A play-the-winner-type urn design with reduced variability , 2003 .

[9]  William F. Rosenberger,et al.  Asymptotic normality of maximum likelihood estimators from multiparameter response-driven designs , 1997 .

[10]  Connie Page,et al.  Estimation after adaptive allocation , 2000 .

[11]  William F. Rosenberger,et al.  On the use of generalized linear models following a sequential design , 2002 .

[12]  D. L. Hawkins.  A proof of the asymptotic normality of the inverse sampling estimator in finite population sampling , 2001 .

[13]  L. J. Wei,et al.  The Randomized Play-the-Winner Rule in Medical Trials , 1978 .

[14]  A. Biswas,et al.  Test of Bernoulli success probability in inverse sampling for nearer alternatives using adaptive allocation , 2002 .

[15]  P. Mykland,et al.  Nonlinear Experiments: Optimal Design and Inference Based on Likelihood , 1993 .

[16]  M. Tweedie.  Statistical Properties of Inverse Gaussian Distributions. II , 1957 .