Using EARNWT for pooled analysis over multiple years and select states

This method does not change because you are only looking at a handful of US states. You can limit your pooled sample to only include observations from New York, Iowa, and Arizona and then follow the methods for approximating the samping weights as discussed here.