I am attempting to use the interaction option with ipdmetan and determine if a treatment-covariate effect is present.
I use the command:
ipdmetan, study(study_id) re: logit depend trt##agegrp sex [pweight=weight], vce (cluster psu)
The error message that is returned is "an error occurred when ipdmetan executed logit".
This appeared to be the syntax in the help file as well as Stata journal article.
Including the poolvar() option with either trt##agegrp, trt, or agegrp results in the same error message.
The same logit regression with the interaction term, outside of ipdmetan runs fine.
Specifying the model with trt, agegroup, and trt#agegroup results in the same error.
Advice for testing the treatment-covariate interaction term in ipdmetal will be greatly appreciated.
I use the command:
ipdmetan, study(study_id) re: logit depend trt##agegrp sex [pweight=weight], vce (cluster psu)
The error message that is returned is "an error occurred when ipdmetan executed logit".
This appeared to be the syntax in the help file as well as Stata journal article.
Including the poolvar() option with either trt##agegrp, trt, or agegrp results in the same error message.
The same logit regression with the interaction term, outside of ipdmetan runs fine.
Specifying the model with trt, agegroup, and trt#agegroup results in the same error.
Advice for testing the treatment-covariate interaction term in ipdmetal will be greatly appreciated.
Comment