ABSTRACT
Data collected from groups of relatives, such as MZ and DZ twins, provide an opportunity to test alternative hypotheses about factor structure and the origins of covariation across time. Such data require multiple–group modeling and constraints to equate parameters across the family members, providing the opportunity to test various hypotheses about growth. One set of hypotheses addresses familial resemblance for growth curve components, and if so whether these patterns are consistent with genetic or environmental factors. A second set of hypotheses addresses whether variation in growth is associated with genetic factors and whether the residual variance in a growth curve model has reliable as well as unreliable components of variance. The full information maximum likelihood model–fitting framework provided by Mx is described. Genetic modeling of dynamical systems is also described and illustrated with application to twin data on blood pressure collected in a physical and mental stress experiment.
