And what doesn't seem right? Did you read my original response?   If you don't want to factor in the other measurement (interaction term or not) then you can run a model with each as the individual variable - two separate models, one with ala and one with aba.   @nottyprof wrote: @Reeza I tried this code below and the output doesn’t seem right.    Proc logistic data=Participants; Class Ala(ref= ‘1’) Aba(ref= ‘1’) / param=reference; model Cas (ref= ‘0’) = Ala aba / scale=none aggregate; run;     ... View more

Found it out later. option ps=10 1s=40; data in; input BMI; title 'Question 4 Respondents Body mass index (BMI)'; if 10 le bmi le 24.99 then bmiclass='Respondents BMI <25'; else if 25 le bmi le 29.99 then bmiclass='Res BMI 25 to 29.99'; else if 30 le bmi le 40 then bmiclass= 'Respondents BMI <30'; cards; 25.06 19.84 23.23 20.33 27.44 29.95 22.44 24.96 23.17 25.74 20.80 21.03 18.47 19.85 21.03 20.48 30.11 23.73 18.48 24.12 24.82 25.62 21.13 21.25 31.00 23.11 23.41 17.16 19.84 21.95 28.36 29.26 18.79 20.80 19.48 20.50 20.23 21.70 25.06 19.14 14.12 23.05 19.13 26.78 29.05 33.20 19.57 26.25 23.57 21.14 20.80 21.45 24.09 30.54 27.89 22.71 29.05 22.66 19.97 22.80 22.80 23.38 20.77 23.04 21.00 20.90 18.95 22.59 23.77 23.17 20.47 22.46 28.34 25.93 19.39 24.96 22.67 22.04 23.99 21.45 20.98 24.21 27.46 24.21 23.03 23.38 24.63 34.97 38.41 23.34 26.47 32.22 26.62 22.96 33.89 20.63 25.49 22.89 27.19 19.57 22.49 25.82 23.71 22.59 16.95 24.33 26.38 25.66 19.97 21.63 29.18 21.87 21.56 19.02 25.92 26.29 25.00 33.47 20.47 18.88 37.65 19.58 24.51 run; proc freq; tables bmiclass; run; Cards could be replaced with dataline ... View more