Abstract
The Cronbach-Gleser utility model provides an integration of the costs and benefits associated with personnel selection programs. Crucial to the appli- cation of this model in the selection programs of the Canadian Forces is the demonstration that the standard deviation of job performance in dollars, SD, can be estimated reliably for complex military occupations. In this study, 206 senior officers who held supervisory positions estimated SD,, for the naval occupation of Sub-Lieutenant using the Schmidt-Hunter and other proce- dures. Results of the study showed that the rank and experience of the senior officers affected the between-judge variance of the SD,, estimates. A modifi- cation of the Schmidt-Hunter technique that provided information on the context of performance and the performance dimensions being assessed did not reduce the between-judge variance. It did, however, significantly affect the judges' perception of an underlying performance distribution on which their estimates were based. Approximately 50% of the officers using the modified technique fulfilled the assumption of normally distributed perfor- mance compared to 33% using the original procedure. Estimates made using these two methods showed little convergence to those made using the superior equivalents and 40% procedures. These latter two procedures produced similar estimates that were higher than the first two. We found the results suggest cautious use of the Cronbach-Gleser model in estimating the utility of military selection programs.