Which geographical variable should I use to hold urban geographical regions constant across decades? Is there anything I should know about implementing this? Thank you.
The most relevant IPUMS USA variables are METAREA and MET2013.
METAREA is available for 1850 through 2011 samples, but it does not hold geographic regions constant. It uses contemporary delineations of metro areas, which is appropriate insofar as changes in metro area boundaries reflect “real” growth in metro regions.
MET2013 uses constant metro area delineations (those of 2013), but is available only for samples back to 2000.
See the METAREA Comparability page for info on how its basis changes over time.
See the MET2013 Comparability page for info on how MET2013 differs from METAREA.
Thank you very much!
I have a quick follow-up. Does METAREA hold the individual units of geography constant? This is really my main concern. Census Tracts (and other delineations) change over time and do not work for my research. Would METAREA work for this purpose and, if not, where might I find a variable type that would? This, I believe, is one thing that IPUMS excels at. I’m sorry if I was not clearer in my original question.
And I’m sorry I didn’t state this more clearly in my original response: IPUMS USA has no variable that identifies geographically static urban regions across all samples, and I don’t know of another source of one.
(I appreciate your faith in IPUMS’s integration prowess, but I’m afraid there are some integration challenges that even IPUMS has not yet conquered!)
Two other variables that might be useful for you: CONSPUMA identifies “consistent PUMAs” for all samples from 1980-2011, and CPUMA0010 identifies consistent PUMAs for 2000 through the most recent ACS sample.
I know some research applications have constructed long time series of metro area data. If you can used pre-defined census summary tables (rather than microdata) for your research application, you could construct such time series by obtaining county-level data from IPUMS NHGIS and aggregating it up to metro areas, using the Census Bureau’s delineation files to identify associations between counties and metro areas. You’d still have to do some extra work to deal with some county definition changes across time.
Thank you very much, I really appreciate all your help.
Best,
Aleex