I used the PWPUMA10 variable provided by staff in a previous question posted here. After merging the file with my existing 2012 1-year file, everything seems to match correctly. However, once I weight the dataset using PERWT, the PWPUMA10 variable doesn’t seem to correctly calculate. The estimates appear to be way too high for the geography I am interested in. Any help you could provide is much appreciated.
When I merge the PWPUMA10 with the 2012 1-yr estimates, the calculations are incorrect. Could be the weights?
There are a number of potential causes for overly large PWPUMA populations. First, make sure to use the variable PWSTATE2, as PWPUMAs are state dependent. Also make sure you are calling on the correct PWPUMAs, as PWPUMA codes differ from PUMA codes. PWPUMA codes and boundaries can also differ by year. the 2012 ACS was the first year to use new geographies based on the 2010 Decennial Census so the PWPUMA codes and boundaries in the 2012 ACS are different from those used in the 2000-2011 samples. The Census Bureau provides a crosswalk between PWPUMAs (called POWPUMAs in the Census Bureau documentation) and PUMAs in the 2012 1-year code list.
I hope this helps.