Abstract
Simple models in the general spirit of Coonet al. (1990) are fit to the Colorado Adoption Project height and intelligence data as the second step in a two-step process. In the first step, over-time data on height and IQ are reduced to level and slope parameters. In the second, these are incorporated in path models along with parental data and fitted using EQS in a multiple-group design. Some comparisons are made between EQS and LISREL for this kind of modeling.