I have the following problem: I have a data set with a panel structure and I want to estimate a regression for every cross-section individually. The time dimension varies significantly between the cross-sections. However, for comparison reasons I want to use for all cross-sections the same time period. E.g. I want to use for the cross sectional regression for individual i the same time observations that are available for individual j. Does anybody have an idea, how to program this?