Problems of Optimum Selection

Abstract

This study examined the selection of different sources of information made by multiple regression, cutting scores, and factor analytic techniques and investigated these procedures in terms of their comparable predictive efficiency. Prior to their job training, 220 Naval winch operator trainees took a battery of seven McQuarrie subtests for mechanical ability. Criterion scores were derived from ratings given subsequent to training. No differences were found among the three methods. However, it is pointed out that the judicious use of factor analysis can provide additional information about the relationships and complexities of the predictors and criterion not available through multiple regression or cutting scores.