How to run diagnostics for logistic regression

[coollord]

Proc glimmix data = cool plots=all;

Class a1 a2 a3 a4;

Model a1 = a2/ dist = binomial  s ddfm = kr;

Random a3/ subject = a4 residual type = cs;

run;

Is there a way you can get the deviance and goodness of fit from proc glimmix?

[UrvishShah]

Try with Proc Logistic as follow...

Proc Logistic data = cool descending;

       model a1 = a2 / DEVIANCE LACKFIT;

quit;

If you want to predict the probability of success based on catogrical response variable then i think Proc Logistic is best to use, you dont need to specify BINOMIAL option on MODEL statement...

-Urvish