Metropolitan area code (IPUMS vs. Census Bureau)



I am going to use 2011 ACS IPUMS data for multi-level modeling.

I want to restrict my sample to people in metropolitan areas.

For the purpose of this study, I will use some variables at the metropolitan level from the census bureau factfinder.

When I checked the METAREA variable in IPUMS, it was based on 2000 census boundary. In the census bureau, they use the recent definition of metropolitan areas.

Is there any solution how to match them?

Thank you.



At this time, we know of no good way to match the 2000 boundaries to the more recent boundaries, unfortunately. Some researchers choose to use metro-level variables created with recent boundaries along with microdata using the 2000 boundaries. Of course, though, this isn’t a perfect solution.



The newest version of MABLE/Geocorr12 can help you out. The matching is sometimes only probabilistic for PUMAs to Metro Areas.