07-29-2015 12:14 PM
I have been trying to automate my glm using %do loops. However, when i try to include all variables initially when &i=1, the loop did not complete and there are no parameter estimates and type3 wald results output (due to non-convergence). I was just thinking if there is any additional codes I should include in my macro to allow SAS to drop off variables (which have extremely large p-values) immediately when there is non-convergence and continue to proceed on looping the GLM.
Your help will be appreciated!