Hello. I can't seem to find documentation anywhere about this; I know that the released PUMS datasets are approximately two thirds of the total ACS in any given year, but how does Census they go about subsetting this?
The main thing that I'm concerned about is: is it possible that the PUMS includes partial households? As in, some people in a household were removed going from ACS to PUMS, yet other people on the same household are kept in. It seems like this would seriously limit many applications of the PUMS, but I've long since learned to not assume that the PUMS data work how I want.
In reply to Schwa: