The study evaluated the accuracy with which one could estimate General Aptitude Test Battery (GATB) scores on the basis of performance on the Career Ability Placement Survey (CAPS) in a sample of 147 psychiatric patients who had received both procedures. Multiple regression analysis was used in which CAPS scales were used as independent variables and GATB scales as dependent variables. Multiple correlation coefficients (Rs) were found to be statistically significant for all twelve GATB scales. A cross-validation achieved by splitting the sample in half and performing separate multiple regression analyses for each subsample was successful, indicating that the regression coefficients were stable. Stepwise analyses indicated which CAPS scales were the most powerful predictors of the various GATB scales. Analysis of the residual scores indicated that predictive accuracy increased as scores approached average levels. It was also noted that the CAPS predicted more accurately to the GATB paper-and-pencil tests than to the performance tests. It was concluded that the CAPS can predict GA TB scores reasonably accurately, but caution should be exercised when predicting scores of individuals with extremely high or low CAPS scores, or when estimating GATB performance test scores.
The stability of validity coefficients over time: Ackerman's (1988) model and the General Aptitude Test Battery.
