@burtsm  An alternative approach is to use a hash table lookup. You first load the hash table with the id's matching the condition you don't want. You then delete all records who have a match with the hash table over id. data have; input id firstbirth birthnum; datalines; 1 1 1 1 1 2 2 1 1 2 0 2 2 0 3 3 1 1 3 1 2 3 1 3 4 1 1 4 0 2 4 0 3 ; data want; if _n_=1 then do; dcl hash h1(dataset:'have(where=(FIRSTBIRTH=1 and BIRTHNUM ne 1))'); h1.defineKey('id'); h1.defineDone(); end; set have; if h1.check()=0 then delete; run; proc print data=want; run; Using hash tables has the big advantage that your source data doesn't need to be sorted. ... View more

The average marginal effect and its standard error can be computed with the Margins macro. The standard error if determined using the delta method.  ... View more

Thank you PGStats. ... View more

You could use a doubled set reading the dataset as usual and opening the same dataset again using a pointer and renamed versions of the variables. The effect it is you can read some data back and forward in the pointered dataset (direct access). The complexity to understand can be hurdle but when understanding it .....   It has a marvelous performance on big data. Interested? SAS(R) 9.4 Statements: Reference, Third Edition ... View more

Thank you Jon--your response was very helpful! ... View more