Can someone enlighten me as to why these different model specifications output different values for select odds ratios (highlighted)? And which is correct?
proc glimmix data=deployments_and_dwells ic=pq; class randomid_adsaged dwellratio_cat(ref='low') officer2_begin marital_status_begin occupation_begin officer2_end marital_status_end occupation_end ; model ptsd (event="1") = dwellratio_cat dep_length_months count dwellratio_cat*count first_mos_begin waiver age sex / dist=binary link=logit ddfm=bw solution cl oddsratio; random intercept / subject = randomid_adsaged; covtest / wald; output out=nonmiss_data / nomiss; where prior_sub = 0; run;
proc glimmix data=deployments_and_dwells ic=pq; class randomid_adsaged dwellratio_cat(ref='low') officer2_begin marital_status_begin occupation_begin officer2_end marital_status_end occupation_end ; model ptsd (event="1") = dwellratio_cat dep_length_months count dwellratio_cat*count first_mos_begin waiver age sex / dist=binary link=logit ddfm=bw solution cl oddsratio(label at count=1.6832 at dep_length_months=6.8151 at first_mos_begin=18.594 at waiver=0.5048 at age=19.794 at sex=0.9767); random intercept / subject = randomid_adsaged; covtest / wald; output out=nonmiss_data / nomiss; where prior_ptsd = 0; run;
The most obvious thing is
model subabuse (event="1")
vs
model ptsd (event="1")
Given different models I would expect different odds ratios
Sorry that was a typo. The outcome is ptsd in both models.
Are you ready for the spotlight? We're accepting content ideas for SAS Innovate 2025 to be held May 6-9 in Orlando, FL. The call is open until September 25. Read more here about why you should contribute and what is in it for you!
ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.
Find more tutorials on the SAS Users YouTube channel.