Using data available from the published reports of factor analyses performed by the Aptitudes Research Project at the University of Southern California, investigations were made into the possibility of demonstrating higher-order factorial abilities latent in those data. Since the policy in those analyses had been to rotate axes orthogonally rather than obliquely, there was no direct information regarding intercorrelations among first-order or basic factors. By combining salient tests for each basic factor, and using their loadings on the obtained factors, correlations were estimated among representatives of those factors. Several methods were devised for factoranalyzing those intercorrelations, applying experimental controls that seemed to be needed, in view of the three-faceted nature of each basic ability. Evidence was found for a large number of second-order abilities and a smaller number of third-order factors, connected with many parts of the structure-of-intellect model, within the limits of available data. In general, the results provide excellent support for the model and for its oblique nature, and for relative discriminability of its categories. Implications for higher-order factorial procedures, for the testing of intelligence, for research, both bivariate and multivariate, and for general psychological theory are suggested.
Download full-text PDF |
Source |
---|---|
http://dx.doi.org/10.1207/s15327906mbr1604_1 | DOI Listing |
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!