I'm testing what I thought was a fully recursive, cross-lagged model with longitudinal data, but now I'm not sure. Are within-time correlations the same as direct pathways between the variables at the same time point?
Also, a member of my dissertation committee thought that the fully recursive model should have perfect fit. I know this is the case if the model is saturated, but my model has something like 20 df. Should a fully recursive model have perfect fit or is it just if it's saturated? And if my model is fully recursive, why is it not saturated?