Jackie, So cem() won't work quite like that. It will only match categories with each other not the kind of bins you are talking about ( where 1 can match 3 and 3 can match 4, but 4 can't match 1). That being said you can do a couple of things: Option1: Just match within strict bins Option2: Implement the matching yourself (particularly if you don't have too many variables this won't be very hard) Option3: Trick cem into doing what you want. How would you do that? Well lets make the problem a little simpler. Let's say you have 3 categories. 1 can be matched to 1,2; 2 can be matched to any; 3 can be matched to 2,3. Basically you replicate all the observations (giving each a unique number) and code them as every category they could be in. Then run cem matching exactly on the categories. Finally go back and remove any remaining duplicates (making sure not to let any strata be empty and adjusting the weights!). You'd have to be extremely careful not to leave any duplicate observations remaining. I think these are in increasing order of difficulty... Brandon On Tue, Apr 26, 2011 at 2:41 PM, Gary King <king(a)harvard.edu> wrote: _______________________________________________ gov2001-l mailing list gov2001-l(a)lists.fas.harvard.edu http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l -- gov2001 mailing list served by Harvard-MIT Data Center List Address: gov2001(a)lists.gking.harvard.edu Subscribe/Unsubscribe: http://lists.gking.harvard.edu/?info=gov2001