A Comparison of the Performance Characteristics of Two Sequential Procedures for Ranking Means of Normal Populations

摘要

This paper reports on the results of a Monte Carlo study of the performance characteristics of three statistical procedures for selecting the normal population with the largest population mean, when the population variances are known and equal. The procedures are the Bechhofer-Kiefer-Sobel (open-ended) sequential procedure, the Paulson (closed) sequential procedure, and the Bechhofer single-stage procedure. Particular emphasis is given to the expected number of stages, and the expected total number of observations required to terminate experimentation for the sequential procedures. In addition to comparisons between procedures, the performance characteristics of the Paulson procedure (actually a family of procedures) are studied as a function of the parameter defining the family. (Author)