When crosstabing state with puma, in the 2015 5-year acs file, I’m getting a puma, lets say in CA, that it’s not included in the “2010 census tract to 2010 puma” file downlaoded frome the census bureau (https://www.census.gov/geo/maps-data/…).
HOWever, when I include county in the crosstabs, I’m finding the puma code from ipums acs is coded as ‘0’ (not identificable?).
Therefore, would I have to crosstab state-county-puma to get an accurate description of pumas (pending not identificables etc)?
The 2011-2015 5-year ACS file uses two different versions of PUMAs. The microdata cases from the 2011 survey have PUMA codes based on the 2000 Census of Population. The microdata cases from the 2012-2015 surveys have PUMA codes based on the 2010 Census of Population. The PUMAs changed between 2000 and 2010, and the cases not in the “2010 census tract to 2010 PUMA” file are those from the 2000 census.
If you restrict your cases to those from YEAR = 2012, 2013, 2014, and 2015, you should have PUMAs that match the “2010 census tract to 2010 PUMA”.