CPS metarea data only going until 2004, while IPUMS-USA is going until 2011. Reasoning?


On the state level, IPUMS-CPS gives me householder population data until 2013 whereas IPUMS-USA is giving me information until 2012.

Then, when filtering by householder status [101 for CPS, 1 for USA] and metarea(1920,1921) for DFW, TX, I can only get until 2011 in IPUMS-USA, and 2004 in CPS. Any reasoning behind this?

Also–the population numbers in each different data base are quite different. I’ve emailed the team for clarification on this, but any other information would be helpful as well!

If it helps


row: year

column: age(c: 5,15)

selection filter(s): relate(101) metarea(1920,1921)

Weight: wtsupp


row: year

column: age(c: 5,15)

selection filter(s): relate(1) metarea(1920,1921)

weight: Household Weight



In IPUMS-CPS, the METAREA definitions changed in the May 2004 sample. For your specific analysis, the new code you would want to use for March 2005 onward would be METAREA==1922 “Dallas-Fort Worth-Arlington, TX”. Similarly, the primary geographic variable for modern IPUMS-USA samples, PUMA, was redefined by the Census Bureau based on the 2010 Decennial Census. The 2012 ACS/PRCS file was the first to use these new definitions.

IPUMS-CPS and IPUMS-USA figures are expected to be different as they represent separate estimates of a single total population figure. Because the total population figures can sometimes appear very different, frequencies of similar variables between the two projects can also appear very different. However, if you look at the percentages rather than the counts and apply standard error estimates the two projects agree. I hope this helps.