It is certainly possible to do that comparison. When you do the MI, a new variable is created (I think called "Imputation"). To compare, turn off the data split (under Data > Split) and then compare whatever variables you want by that new variable that was created. I hope that helps.
As a side note, this website has moved to: www.statsmakemecry.com
I have run multiple imputation using SPSS on a dataset using SPSS. This is on a cohort of subjects on health questionnaire scores over a year. Looking at the original data and pooled imputed data I think there is a difference in the later scores i.e. not MCAR (possibly a healthier survivor effect). Is it possible to run a comparison between the original and imputed data in SPSS?