Okay. I made that change. The MIANALYZE now produces output for 4 (gender,length,regions,status) of 8 of my predictors - in other words, no output for "ethno", "sab", "reasons", or "d8lgbtcom". These are all of my dichotomous variables, so I wonder if it's an issue with the coefficients, I selected. I changed the code from param=ref to param=glm as per your instructions. I'm not sure of the implications of this. Based on your explanation of coefficients above, does this mean that a variable with multiple levels will need multiple estimates lines? For example, the b5gp variable has 3 levels: Estimate 'gender' b5gp 1 -1/exp;* Estimate 'gender' b5gp 1 /exp;* Additionally, the mianalyze output provides one single-row LBetaEstimate Variance table and one single-row LBetaEstimate Parameter Estimates table). I'm not sure what this tells me that the original mianalyze output wasn't telling me before. Thank you! here is the log for mianlayze. NOTE: The data set WORK.EST_DS has 80 observations and 12 variables. NOTE: The data set WORK.REGPARMS has 740 observations and 9 variables. NOTE: PROCEDURE GENMOD used (Total process time): real time 44.16 seconds cpu time 27.37 seconds 21008 proc sort data=est_ds; 21009 by label _imputation_; 21010 run; NOTE: There were 80 observations read from the data set WORK.EST_DS. NOTE: The data set WORK.EST_DS has 80 observations and 12 variables. NOTE: PROCEDURE SORT used (Total process time): real time 0.01 seconds cpu time 0.01 seconds 21011 proc mianalyze data=est_ds; 21012 by label; 21013 modeleffects LBetaEstimate; 21014 stderr stderr; 21015 run; NOTE: PROCEDURE MIANALYZE used (Total process time): real time 0.07 seconds cpu time 0.07 seconds
... View more