I know that a one factor model with only 2 indicators is not identified and I receive that error when running my model using WLSMV. However, when I run the same model using MLR (I have categorical indicators) I don't receive an error and instead receive output with factor loadings when I constrain the factor to 1 but do not constrain each of the indicators. Are the factor loadings provided by MLR usable in this scenario? I have to run CFA on each factor separately before entering my full model with all factors. Thanks.