Pumas in data is less than the total number of puma in IPUMS


I’ve downloaded the IPUMS USA data for 2012-2019, but when I summarize the geoid(I’ve already combined the statafip code and puma code to generate a unique geoid for each puma) in STATA, there are only about 1,800 puma. But I learn that the total number of puma included in IPUMS USA survey are more than 2,400. And every puma is included in annual survey. Where are the lost ones? Could you help me with that? Thank you for the reply.

There are a total of 2,378 2010-based PUMAs (see this Census Bureau documentation on PUMAs). Of these 2,378 PUMAs, 2,351 are in U.S. states, and 27 are in Puerto Rico. The 2010-based PUMA definitions are used in ACS samples from 2012-2021 and the 2010 decennial census.

I am not able to reproduce your estimate of 1,800 PUMAs in an extract from IPUMS USA that includes all 1-year ACS samples from 2012 to 2019. Are you restricting your data in some way, for example, looking only at identifiable counties, or only specific states, or a subpopulation? Any restriction on your sample may result in omission of some of the PUMAs.

Thank you Isabel!
I realize there is something wrong when I generate the unique geoid with statefip and puma code. Now I get 2,351 pumas. Thank you very much:)