I am trying to run a two factor confirmatory analysis in AMOS (version 20). However, the factors theoretically do not correlate. Should I keep in the two way arrow between the two factors and constrain the covariance to 0? Or should I take it out? Or just keep it in and don't do anything to it?
I am also having the issue where if I take it out or constrain it, I have the same degrees of freedom as the single model CFA, making it impossible to calculate the chi-square difference test.
Thank you so much for any advice or help!