A CLASS OF BAYES-OPTIMAL 2-STAGE SCREENS

Items are characterized by a set of attributes (T) and a collection of covariates ((X) under bar) associated with those attributes. We wish to screen for acceptable items (T is an element of C-T), but T is expe nsive to measure. We envisage a two-stage screen in which observation of (X) under bar is used as a filter at the first stage to sentence mo st items. The second stage involves the observation of T for those ite ms for which the first stage is indecisive. We adopt a Bayes decision- theoretic approach to the development of optimal two-stage screens wit hin a general framework for costs and stochastic structure. We also co nsider the important question of how much screens need to be modified in the light of resource limitations that bound the proportion of item s that can be passed to the second stage. (C) 1996 John Wiley & Sons, Inc.