I have been asked by a reviewer to report the "proportion of the variance in scores accounted for by the CFA model". I assume that he or she is asking for the proportion of variance that is explained in the observed measures, which I understand is typically the item R^2. I have a CFA model with 12 categorical indicators, analyzed with WLSMV. There is one substantive factor (all 12 items) and one method factor (6 of the 12 items).
I would report the item R^2 values, however, I understand that the R-square for each indicator listed in the output is actually the proportion of variance in the latent response variable (for each item) that is explained by the latent factor. Is there a standard for reporting "proportion of variance explained" when using WLSMV with categorical indicators?