I charge for a moderated mediation in lavaan.
The issue is that one variable correlates with the moderation interaction term.
I have thus included this correlation into the model.
Unfortunately, the interaction variance is now exceedingly high.
My supervisor recommended me to set the interaction variance to 1, however doing so absolutely ruins the model's fit.
Does anybody have suggestions about what to do?
The model in question:
model4 <- ' Med ~ a1UV + a2Controllvariable+ Mod
AV ~ bMed + c1UV + c2Controllvariable + Mod+ INT
INT ~~ 1INT
ind_AV := a1b