Can the counts of individual IPUMS USA data be summed and weighted to be representative at the PUMA level?


#1

I need PUMA level data for 2002, 2005, 2008 in NYC to pair with a NYC HVS analysis that I’m doing. ACS data doesn’t have it for 2002 or 2005. So, I was thinking of using IPUMS USA data. Could I sum the counts and then weight them to be representative of the PUMAs in NYC?

If so, any help would be appreacited.


#2

Yes, US Census and ACS data is, generally speaking, representative at the PUMA level. Of course, if you are counting relatively small sub-groups within PUMAs then your estimates may have a relatively large margin of error. In general, however, it sounds like what you want to do will be valid.


#3

Ok…good to confirm. Another question, then, how com the US Census ACS does not have data at the PUMA level for 2001-2005, but IPUMS ACS does? I was under the impression that the ACS started in 2006. If so, how is there data for IPUMS for 2001-2005?