I am running into a wall on applying a predicted class of a state (all 50 states have one of four predicted classes) in 2001 to the rest of the years in the dataset. The data is set as long with the variables state, year, and predclass. Any tips or tricks on how to either (a) change the predclass variable within a state to the 2001 level or (b) create a new variable from these three to apply the 2001 predicted class to the other years?

Thanks in advance.