Hi all,
I have a strongly unblanced panel dataset, i.e. for some sections there might exist 2 time periods, for others there exist 10, etc. and the respective time periods might not even be overlapping.
Nonetheless, I would like to perform a fixed effects analysis (both one-way and two-way), to analyse certain relations among variables. I am familiar with the "go-to" approaches in SAS for OLS and balanced panel models. i.e. proc reg and proc panel. But which procedure is preferableto perform panel regressions for such strongly unbalanced data? Is proc panel still a valid option?
