| Although employers usually hire with the hope employees will successfully contribute to the organization for years, test validation studies typically use short follow-up periods. Unfortunately, of those that assess later performance, a significant number indicate that cognitive tests predict initial performance better than later performance. Many such studies were conducted in experimental contexts, but even those from organizational contexts indicate a decline over time of the validity of cognitive tests. In contrast, the validity of non-cognitive predictors may increase over time (e.g., Helmreich et al., 1986).;The hypothesis that the validity of cognitive and performance predictors would decrease over time and the validity of non-cognitive predictors would increase over time was tested with a sample of 253 air traffic controllers. Various aptitude measures and initial simulation scores served as cognitive predictors, whereas NEO-PI scores and biographical data served as non-cognitive predictors. Criterion measures included time to train, and supervisor ratings of training outcomes at early, middle, and late phases (up to two years) of training.;Relatively few predictors of either type were significantly associated with performance criteria. However, regression analyses indicated that the validity of the non-cognitive predictors did increase over time but the validity of cognitive predictors was inconsistent. Similarly, the incremental validity of non-cognitive over cognitive predictors increased over time. Results also indicated that non-cognitive predictors were significantly associated with change in performance from early to middle and middle to late phases. These data support the use of non-cognitive predictors and highlight the need to engage in long-term test validation. |
