IPUMS-USA: Official documentation on the representativeness of data for small geographical samples?

Mr. Moreland,

I am a PhD candidate in sociology and am using IPUMS USA data for a research project (1970, 1980, 1990, 2000 1 or 5% Census waves, 2007-2011 ACS waves). I am computing measures of inequality in local geographical areas using IPUMS wage and household income data.

A reviewer of my manuscript claimed that the IPUMS-USA data is not representative of small areas (such as PUMAs). However, based on your answer to a forum question (Does IPUMS ACS provide PUMA level person weights), I believe that the reviewer is incorrect in his/her assertion. It seems that the data can be considered representative of sub-national geographic areas, as long as the data is correctly weighted.

My question is the following: does the IPUMS have any official statement / documentation noting that the weighted data is representative of small units? I would like to provide some form of authoritative documentation to the reviewer. I have looked through much of the IPUMS documentation but haven’t yet found an exact answer to this question.

Thank you for your time and assistance. Also, thanks to your whole team for the wonderful service you provide along with the IPUMS data. I am constantly amazed with the high quality documentation and help you all provide. It makes the researcher’s job much easier!!

If you need any additional information, please let me know. Thank you, and I hope this message finds you well.


Tom VanHeuvelen
PhD Candidate, Sociology
Indiana University

You are correct that the IPUMS-USA data can be considered representative of sub-national geographies, when weighted properly. We make note of this in our FAQ section under the question “You say the IPUMS is nationally representative; is it also representative of smaller areas?” Since the Census Bureau collected the source data and created the weight values, it may be useful to contact them for a second source of authoritative documentation on the matter.

Hope this helps!