I'm working on a SEM linking parental (M=mothers; F=fathers) autism and ADHD symptoms to perceived stress in the upbringing of their affected (strsFA/strsMU) and unaffected (strsFU/strsMU) offspring, with parental depressive mediating herein.
My model is as follows:
MODEL: deprF ON adhdF ASDF; deprM ON adhdM ASDM;
strsFA ON deprF; strsFU ON deprF; strsMA ON deprM; strsMU ON deprM;
strsFA ON adhdF ASDF; strsFU ON adhdF ASDF; strsMA ON adhdM ASDM; strsMU ON adhdM ASDM;
adhdF WITH ASDF; adhdM WITH ASDM; strsFA WITH strsFU; strsMA WITH strsMU;
ASDM WITH ASDF; adhdM WITH adhdF; deprM WITH deprF; strsFA WITH strsMA; strsFU WITH strsMU;
My main problem is that I can't get Mplus to equalize the correlation between strsFA WITH strsFU on the one hand, and strsMA and strsMU on the other hand. Several of my colleagues working with Mplus also did not get this to work. is this because these are dependent measures and also capture error variance?