Status:Closed    Asked:Apr 13, 2015 - 01:22 PM

Creating unique IDs for PUMAs

I'm using PUMAs as my geographic unit of analysis for a project using the ACS2012 5 year file. I know that the PUMA codes are state dependent, so I've created a unique puma id by combining them with the statefips variable, using the following r code:

fam$pumaid = paste(as.numeric(fam$statefip), as.numeric(as.character(fam$puma)), sep="-").

This gives me an ID code with the following format: "[statefip code] - [puma code]". My problem is that this returns 3344 unique IDs, whereas there are only 2378 PUMAs. Any idea what I'm doing wrong?

Thanks so much for your help!

Do you have the same question? Follow this Question

Staff Answer




In the 2012 ACS 5-year file, the PUMA variable for respondents from 2008-2011 is coded based on the 2000 Census boundaries, while PUMA is coded based on the 2010 Census boundaries for 2012 respondents. As a result, the total number of PUMAs in the file will be greater than the number of PUMAs in any one of the coding schemes alone.

Hope this helps.


Apr 13, 2015 - 02:57 PM

Report it


Since pumas are unique only within state, could there be a state-puma variabl...
Correspondence between Puma 1990 and Puma 2000
1990 1% PUMA boundaries file when using 1990 1% unweighted IPUMS USA data?
PUMA codes assigned to all states [ACS 2012 and 2013 (1% sample)]
Login   |   Register

Recently Active Members

View More »

Share |