D ecision makers in any industry constantly function as information processors.They have limited amounts of information which must be combined in some way to make intelligent decisions. The process which fills the interval between information gathering and the actual decision has long been a challenge to research ers: What exactly does the decision maker do with the available information? One approach to studying and describing the decision-making process is the method known as "policy capturing."Policy capturing is a statistical strategy which analyzes actual decisions, ulti mately providing a mathematical description of the judgment policy that was used. Its major objective is to develop models of the specific processes decision makers use to weight and combine different pieces of information. CHOOSING THE MODELWithin the regression approach to policy capturing (the other approach is the Bayesian, and will not be considered here), there are two paradigms: the correlational and the analysis of variance paradigms. Each technique builds a model which weights the available pieces of information (known as cues or predictors) according to their influence on the decisions (criteria).The correlational approach employs the statistical procedure of multiple regression analysis. Although it may involve any of several mathematical models, one model in particular has received a great deal of empirical investigation and support. This is the linear model, introduced by Hoffman (1960). Use of the linear model involves the notion that a judgment is accurately described as some linear combination of the available cues. Thus, in a linear model, each piece of information is weighted (or Laura Donnelly is a third-year graduate student in psychology at Virginia Polytechnic Institute and State University. She recently received her master's degree there and is currently working towards her doctorate, with a focus in industrial/organizational psychology. Her priamry areas of interest are per sonnel selection and management issues. David Bownas is Assistant Professor of Psychology at Virginia Polytechnic Institute and State Uni versity. His major interests are in job analysis, personnel selection, and applications of quantitative methods to industrial psychology. Public Personnel Management Journal 81 by guest on July 25, 2015 ppm.sagepub.com Downloaded from

A technique was developed which provided a quantitative index of the fit between training curriculum content and job task performance requirements. The procedure also generates a listing of tasks which receive undue emphasis in the training curriculum, those which are not being trained, and those which instructors intend to train, but which course graduates report being unable to perform. The procedure is illustrated for three training programs in the U.S. Coast Guard Class 'A' schools.

No abstract