3 Fitting a Phenotypic Factor Model

Martin et al. (1985) obtained data on arithmetic computation from male and female twins who were measured once before and three times after drinking a standard dose of alcohol. To illustrate the use of a confirmatory factor analysis model in Mx, we analyze data from MZ females (first born twin only). The observed variances and correlations are shown in Table 10.1. The confirmatory model is one in which a single latent factor is hypothesized to account for all the covariances among the four variables. The Mx script below shows the model specifications and the $4\times 4$ input matrix.

Table 10.1: Observed correlations (with variances on the diagonal) for arithmetic computation variables from female MZ twins before (time 0) and after (times 1 - 3) standard doses of alcohol.

	Time 0	Time 1	Time 2	Time 3
Time 0	259.66
Time 1	.81	259.94
Time 2	.83	.87	245.24
Time 3	.87	.87	.90	249.30