I am using TYPE = IMPUTATION (with 10 imputed data sets) for a SEM. This works fine for most analyses but in order to be able to do the Chi²-difference test with non-normal data for MLR (as described on statmodel.com), I would need non- aggregated results (especially for CHI²-tests) for each of the 10 data sets.
Is there a way to get the results for each data set and the aggregated results across all data sets by running only one analysis and not having to run 11 analyses (1 aggregated and 10 for 10 data sets)? I would love to get one results statement without having to track the outcomes of every analysis for each imputed data set.
I'm not sure research exists that shows how to do difference testing for multiple imputation. If you want results for each data set, you need to run the analysis separately for each data set and save the results.