Predictive validity of a management aptitude test battery

This study had two major objectives: (1) to determine predictive validity of the portion of the Standard Oil of New Jersey management aptitude test battery given to short service employees and (2) to measure the degree of consistency between the original Standard Oil keys and item data for the follow-up sample for the three empirically keyed tests. The sample members were 146 marketers employed by the same company in 1957-61 and tested in 1963-64; test scores were compared to rates of progression and estimates of future progression of this group within the organization. More specifically, two types of variables served as criteria in this study, A complex success index comparable to the one constructed in the original study, combining a graphic scale rating of overall performance, salary, and organization level with corrections for age and service was computed for the years 1966-68 individually and combined. Also, management's estimate of potential organization level of responsibility was available for the years 1966-68 separately and combined. Performance evaluations and estimates of potential were not influenced by test scores. Total test battery scores, with individual tests weighted according to the original standard Oil formula, generally showed correlations with the criteria in the upper .30's contrasted to the original concurrent validity of the middle .50's; test scores reweighted and combined to maximally predict each criterion in turn showed multiple correlations in the upper .40's. These correlations were consistent in magnitude for the three years for which criteria were developed. The success index proved to be a more useful criterion than the estimate of potential due primarily to greater year-to-year reliability, more consistent patterns of relationship with the individual tests, and finer differentiation among sample members. Scoring keys derived from follow-up sample item data for the three empirically keyed tests showed considerable consistency with the original keys; however, there were a few reversals in scoring for each test. The consistently keyed items as well as those showing reversals in scoring were theme analyzed for each test to determine the types of items demonstrating stability or instability as predictors through time.