K Williams posted on Thursday, September 20, 2007 - 8:04 am
I am trying to model change in relationship quality (over a 15-year period) and change in distress (over the same period) as parallel growth processes. My central hypothesis is that the effect of change in relationship quality on change in distress varies with age at time 1. Is there a way to regress the slope of distress on an "age X slope of relationship quality" interaction term? Or do I have to do multiple group analysis of the parallel growth process for different age groups?
You can regress the slope growth factor of one process on the slope growth factor of the other process and the interaction between the slope growth factor of the other process and age. You would create the interaction between the other growth process and age using the XWITH option. This model may be difficult to estimate.