Designing selection experiments with bernoulli populations |
| |
Authors: | Guido Giani |
| |
Institution: | Department of Medical Statistics and Documentation , Technical University of Aachen , D-5100, Aachen, FRG |
| |
Abstract: | This paper is concerned with a fixed size subset selection problem for Bernoulli populations in the framework of the indifference zone approach. The goal is to select s populationswhich contain at least c of those with the t largest success probabilities. In order to control the probability of correct selection over the preference zone extensive tables of exact minimum sample sizes have been prepared to implement the single-stage procedure generalized from the well-known Sobel-Huyett procedure. It is shown how the tables can also be employed to design certain closedsequential procedures. These procedures curtail the sampling process of the single-stage procedureand may differ in their sampling rules. Two procedures working with play-the-winner rules are described in detail |
| |
Keywords: | Bernoulli selection problem Fixed-size subset selection Indifference zone Statistical tables Seqential selection procedures |
|
|