Latent Structure Analysis of Longitudinal Data on Relations Between Intellectual Abilities and School Achievement

Abstract

Covariance structure modelling by LISREL was applied to analyze the relations between intellectual ability, as measured by ITPA, and achievement in reading, writing, foreign language and mathematics in longitudinal data with 8 to 10 years' time interval between measurements. The subjects' (N = 234) ability scores were obtained at 5-7 years of age, and their achievement was measured at 14-16 years of age. The effect of ability to achievement was studied by analyzing structural equation models on both first-order ability and achievement factors and by developing a two-stage second-order factor estimation method for structural equation parameters between first-order ability and achievement factors. Squared multiple correlations and coefficients of determination as indices of explained variance were derived for reduced forms of structural equations. The general latent intellectual ability explained 49% of variance in school achievement. The results showed the high validity of ITPA in explaining school success as well as the great flexibility and usefulness of covariance structure modelling by LISREL in analyzing longitudinal data.