We are studying personality change from pre- to -post disaster in a group of subjects on whom we have a continuous measure of disaster exposure. We have performed invariance tests comparing the two time periods. There was little difference between the pre- and post models. Nevertheless, we want to test whether the degree to which participants were exposed accounts for the degree to which the pre-disaster model differs from the post-disaster data.
Is there a way to model the effects of exposure of model invariance?