The Dimensionality of the Swedish Scholastic Aptitude Test

Abstract

The paper investigates the dimensional structure of the six subtests in the Swedish Scholastic Aptitude Test, using data from 10 different versions of the test administered to in all some 50,000 examinees. In the analysis several lisrel models of the factorial structure of the subtests are fitted and evaluated. The modelling shows a simple 1‐factor model to be inadequate for the data and two alternative, hierarchical, models are investigated. One of these includes 3 first‐order factors and 1 second‐order factor. The other model includes 2 orthogonal factors: a general factor (G) with relations to all 6 subtest, and a verbal‐educational factor (Gc') with relation to 4 subtests. On the basis of evaluations of estimated factor loadings and statistical criteria of fit it is, tentatively, concluded that the model with 2 orthogonal factors (G and Gc') is to be preferred.