Codes and Frequencies
Most IPUMS data transformations are performed using variable harmonization tables that specify how each value in the source data is recoded. Some variables also require programming logic in addition to the harmonization table. The harmonization documents for this variable are:
GEO2ALT_US identifies the most detailed areas that can be consistently delineated from the geographic codes available in United States PUMS (Public Use Microdata Series) files from 1980 to 2010. In 1980 PUMS files, the smallest identified units are “county groups,” and in later PUMS files, the smallest identified units are Public Use Microdata Areas (PUMAs). All of these units nest within states, but the definitions of PUMAs differ from county groups, and the 1990 PUMAs differ from the 2000 PUMAs (which are used for 2000, 2005 and 2010 samples). GEO2ALT_US is spatially harmonized to account for these boundary changes. Each GEO2ALT_US code corresponds to a distinct set of 1980 county groups and 1990 and 2000 PUMAs that collectively cover a consistent geographic area across time. More details on construction of PUMAs can be found in IPUMS USA website.
Some detail is lost in harmonization; see the comparability discussion. A GIS map (in shapefile format), corresponding to GEO2_US can be downloaded from the GIS Boundary files page in the IPUMS International web site. The GIS boundary files for GEO2_US are based on the U.S. Census Bureau’s 2000 TIGER/Line files.
The full set of geography variables for the United States can be found in the IPUMS International Geography variables list. For cross-national geographic analysis on the first and second major administrative level refer to GEOLEV1, and GEOLEV2. More information on IPUMS-International geography can be found here.
Comparability — General
PUMA and county group boundaries were visually inspected within a geographic information system; where boundaries were coincident, or nearly coincident, across time, they were aggregated to the smallest possible number of units to retain the highest possible level of geographic detail. The nation was split into 543 areas that can be consistently identified across samples. A small degree of mismatch was allowed along some boundaries in order to avoid over-aggregation and thereby maintain a high level of geographic detail. Although these boundary mismatches all appeared to be insignificant in size, it is possible that some mismatches involved a substantial population and some portion of the change in population characteristics across samples may be due to changes in the boundaries of the associated county groups and PUMAs. More information on consistent PUMA boundary construction can be found in CONSPUMA. Labels in GEO2ALT_US match CONSPUMA.
Due to significant population losses in the New Orleans area following Hurricane Katrina, the Census merged 3 PUMAs in Louisiana (2201801 and 2201802 merged into PUMA 840022172; and PUMA 2201905 merged into PUMA 840022173). For the 2000 and 2005 samples, PUMAs 2201801 and 2201802 are allocated to GEO2ALT_US unit 172 and PUMA 2201905 is allocated to GEO2ALT_US unit 173. In the 2010 sample, the combined PUMA 2277777 is allocated wholly to GEO2_US unit 172. In effect, the entire area of PUMA 2201905 switches from one GEO2_US unit to another between 2005 and 2010.
Any analysis that uses GEO2ALT_US for both 2010 and earlier samples should recode GEO2ALT_US codes 172 and 173 to a single code in order to achieve comparability.
- All households and group quarters; not available for Puerto Rico.
- United States: 1980, 1990, 2000, 2005, 2010