The current geographic variables do not allow for analysis of an MSA unless it aligns with PUMA or METAREA. (I’m specifically interested in the Atlanta MSA, where METAREA reflects the 2000 MSA boundary but not 2003-current MSA boundary). Will new data releases include an option for MSA level analysis?
As mentioned in the METAREA Comparability Statment, since the Census Bureau did not release information associating the PUMA boundaries with 2005-onward samples, IPUMS-USA maintained the 2000 MSA boundaries. The base geographic level for PUMS data is the PUMA, so IPUMS-USA must base other other geographic objects on PUMAs. This is way many METAREAs and CITYs are unidentifiable. This will continue to be the case with IPUMS data. The recently released 2012 ACS files use a new set of PUMAs based on the 2010 Census. The IPUMS generated greographic varaibles (listed in the release note for the 2012 ACS IPUMS release) are currently being created, but, comparing the PUMA and Metropolitan Area layers on with the TIGER/web interactive map, it does not appeare that the Atlanta MSA will be identifiable with the new PUMAs either.