Hi,

I would like to know how I could deal with an omitted variable bias in a survival analysis. My research involves the duration of a particular strategy up until the exit. I think variables measuring performance or size are relevant for the results of the research, although they are not included in the analysis. I have used, based on fitting criteria, a log-logistic parametric accelerated failure time survival analysis. Is it true that with the use of the 'frailty' function in Stata, I prevent or overcome the omitted variable bias?

Kind regards,

Michael