Reading performance and general cognitive ability: A multivariate genetic analysis of twin data

Abstract

Confirmatory factor analyses of three reading performance measures and full-scale Wechsler IQ data obtained from 146 pairs of twins of average or above average reading ability were conducted. A simple factor model which contained one general reading factor, three specific factors, and a single set of IQ loadings was employed to assess the extent to which phenotypic correlations among the reading measures are due to IQ. The significance of the general reading factor was determined by constraining its loadings to be zero and then computing a χ2 goodness of fit for the reduced model. The resultant change in χ22 of 114.09 (P 0.5), the removal of the genetic covariance component did (χ210 = 20.26, P < 0.05). Subsequent confirmatory factor analyses of the genetic and unique environmental covariance matrices indicated that the reading factor could not be removed from either matrix without causing a highly significant deterioration in fit. However, the similarity between the genetic and phenotypic factor loadings suggests that the phenotypic factor structure is largely genetic in origin.