ABSTRACT

Latent growth modelings (LGMs) strongly resemble classic confirmatory factor analysis models. However, because they use repeated measures raw-score data, the latent factors are interpreted as chronometric common factors representing individual differences over time. To interpret some of the model parameters, it is necessary to review the basic equations for expectation, variance of difference scores, and covariances of difference scores with initial status. All five of the model parameters can be expressed as functions of the measured means, variances, and covariances. The same manipulations are used for the variances and covariances, taking advantage of the fixed values for the factor loadings. In testing a more parsimonious model, such as a two-factor LGM, sufficient degrees of freedom are available for the evaluation of model fit. McArdle has termed the univariate single-factor LGM a curve model. The general two-factor LGM approach outlined in this chapter has many advantages for use in the testing and evaluation of developmental models.