Dear All, 

I have model in GENLIN with both age and smoking status (dichotomous - no = 0, yes = 1) added to the covariate box. The EMmeans  produced are fixed for age at 50 (mean across the groups) and smoking status of 4.5. The latter does not make sense given it can only be 0 or 1 (no or yes). I believe that it is treating the 0s and 1s as continuous data rather than categorical. Here is the code:

GENLIN Max_ini BY Group (ORDER=ASCENDING) WITH AAE Ev_smo
  /MODEL AAE Ev_smo INTERCEPT=YES
 DISTRIBUTION=NORMAL LINK=IDENTITY
  /CRITERIA SCALE=MLE COVB=ROBUST PCONVERGE=1E-006(ABSOLUTE) SINGULAR=1E-012 ANALYSISTYPE=3(WALD) 
    CILEVEL=95 CITYPE=WALD LIKELIHOOD=FULL
  /EMMEANS TABLES=Group SCALE=ORIGINAL COMPARE=Group CONTRAST=SIMPLE(1.00) PADJUST=SEQBONFERRONI
  /MISSING CLASSMISSING=EXCLUDE
  /PRINT CPS DESCRIPTIVES MODELINFO FIT SUMMARY SOLUTION.

My questions are:

1. Should smoking go into the covariates box as it is categorical?
2. If so then how can I make the model appreciate it is categorical data?
3. In UNIANOVA I have been able to state using with WITH function the level of the covariate that I want to report the EMmeans, but this does not seem to work in GLM.  Is there a way of asking SPSS to return the EM means for a desired level of a covariate (i.e. smoking status set to 0)?

Any help would be much appreciated.

Phil