06-08-2014 10:31 AM
Hiya, not sure how to calculate a confidence interval, particularly the standard error
My teacher said it's "exp(B1 + B4 +/- 1.96* sqrt(se(B1) + se(B4) + 2cov(B1,B4) )"
where B1 is the factor and B4 is the interaction term (estimates).
06-08-2014 03:47 PM
You haven't provided enough info. What proc are you using, what code, options?
If its survival analysis and you need WALD CI for hazard ratio you can use the CL=WARD option on the hazard ratio statement.