Looks like you might want:   proc sql; create table counts as select id, shortIPI, sum(case when PTB="yes" then count else . end) as PTB_yes, sum(count) as PTB_all from meta group by id, shortIPI; quit; proc glimmix data=counts; class id shortIPI; model PTB_yes/PTB_all = shortIPI / solution; random intercept / subject=id; run; or maybe exchange the roles of  PTB and shortIPI.   ... View more

Yes, @Astounding is correct, you have a semicolon where you need a / to specify options instead.   model prevhyp (event="1") = stroke diabetes sysbp educ BMI diabp totchol glucose prevchd /selection=stepwise ... <rest of your options>; @taylor_metz1 wrote: I am trying to use a backwards selection process to create a parsimonious model using proc logistic.   when I attempt it my code is this:   proc logistic data=test2; class stroke(ref="1") / param=ref; model prevhyp (event="1")= stroke diabetes sysbp educ BMI diabp totchol glucose prevchd; selection=stepwise; slentry=0.05 slstay=0.20 details lackfit; run;     and my SAS log says this:   248  proc logistic data=test2; 249  class stroke(ref="1") / param=ref; 250  model prevhyp (event="1")= stroke diabetes sysbp educ BMI diabp totchol glucose prevchd; 251  selection=stepwise;      ---------      180 ERROR 180-322: Statement is not valid or it is used out of proper order. 252  slentry=0.05 253  slstay=0.20 254  details 255  lackfit; 256  run;   I am not sure what I am doing wrong as I have never used the SLSTAY function before.   ... View more