A contemporary method for developmental-genetic analyses of age changes in intellectual abilities

The purpose of this article is to describe a methodology for the evaluation of biometric genetic hypotheses in the context of a developmental model of growth and change. Linear structural equation models are described for longitudinal and twin data, including aspects of subject attrition and practice effects. These models are applied to 2 variables measured at several points in time in the New York Twin Study (Jarvik, Kallman, & Falek, 1962; Kallman, Feingold, & Bendy, 1951). The patterns of psychometric and biometric changes are different for the 2 intellectual variables. Substantive results are discussed in relation to g(f)/g(c) theory (Cattell, 1971; Hem, 1988), and some methodological limitations are emphasized.