Dear Stata list,
I am working with the command xtpoisson (re) and xtnbreg (re) and have a few problems/questions. my DV is count data and I have an unbalanced panel data set with 74 clusters.
1) When using xtnbreg the iteration procedure often takes a long time to iterate. Often producing hundreds of iteration results. Is this something to do with the missing data used in estimation? In this case, is it best to rely on the results from xtpoisson?
2) How do I obtain robust standard errors. From previous posts I see that there may be a problem with using the bootstrap options with clustered counts.
