Abstract

Costing human performance has long been a goal in industrial and organizational psychology in evaluating the effects of organizational interventions on the benefits expressed in economic terms like money. One procedure for estimating the utility of job performance has been developed by Schmidt, Hunter, McKenzie, and Muldrow(1979). The purpose of this study was to investigate the applicability of the Schmidt-Hunter global estimation method to a broad range of jabs as a function of the Data, People, and Things variables of work. The applicability of the method was evaluated by the the degree of normality of economically-expressed distribution of job performance(the index of normality) and the degree of agreement in the utility estimates among raters(inter-rater reliability). Raters in this study were 807 assistant managers in a large motor manufacturing company. Raters provided the degree of manipulating things, interesting people, and dealing with data in performing their jobs. Also they estimated the economic contribution of employee at the position of 15, 50, and 85 percentile of economically-expressed distribution of job performance. It was find that the Data, People, and Things variables of work can explain the applicability of the global estimation method across a variety of jobs. More specificically, the People variable was highly related to the normality of economically-expressed performance distribution and the degreee of agreement among raters. While the jobs dealing less with people had more statistically normal distribution of job performance and produced higher agreement, the jobs dealing mare with people had non-normal(positively skewed) distribution and produced lower agreement in the three estimates among raters. Therefore, it was concluded that the jobs, which require more frequent contacts with other people have low applicability of the global estimation method, and vice versa. The implications of these findings and limitations of this study were discussed.