Higher-Order Factors in Structure-of-Intellect (SOI) Aptitude Tests Hypothesized to Portray Constructs of Military Leadership: A Re-Analysis of an SOI Data Base

Abstract

Taken from a 1959 research report by Marks, Guilford, and Merrifield, a correlation matrix of 21 structure-of-intellect (SOI) tests intended to reflect aptitudes required in military leadership was re-analyzed through use of relatively objective exploratory factor analytic procedures (varimax and promax solutions for orthogonal and oblique structures, respectively) and through employment of confirmatory maximum likelihood factor analyses. For the sample of 204 Marine officers, only modest correspondence was found in the exploratory analyses between the solution by Marks et al. based on the orthogonal rotation of factor axes from a centroid matrix and the factor solutions obtained from application of the varimax and promax methods of rotation. An evaluation of selected alternative models comprising first-order and higher-order factors involving use of LISREL methodology revealed that a two-factor oblique model affording a separation of semantic and figural contents yielded the greatest promise in accounting for the covariation among the 21 aptitude tests. A higher-order oblique factor model furnishing a differentiation among three psychological operations of cognition, divergent production, and evaluation ranked second in its capability of reproducing the original correlation matrix. The conclusion was reached that the covariance among these aptitude tests can be meaningfully conceptualized within higher-order factor structures. A short discussion highlights possible difficulties and limitations likely to be realized in establishing the construct validity of the hypothesized first-order dimensions within the structure-of-intellect model.