Angel Arias posted on Friday, January 25, 2019 - 7:43 am
I have followed the two-step procedure to compare CFA nested models with categorical indicators by saving the derivatives of the H1 (least restrictive) model and then fitting the H0 model. However, the value of the chi-square difference test (0.372) does not equal the subtraction of the chi-square values of the H0 (878.143) - H1 (877.891) = 0.252, which is NOT 0.372. Why is this the case? Is it because of the estimation method, WLSMV?