MetArea, County, and Place codes

USA
1

I’m looking to combine some HUD data, which appears to be identifiable at the county level, with ACS data for 2000-2011. I’m currently having some issues where there appear to be a number of counties without any representation in the ACS data; for example, Miami-Dade county has no observations for county FIPS at all. I think that this is because people living in Miami and other cities would normally be identified by a place code instead, but this isn’t available from IPUMS. The data I’m trying to merge with also has some MSAs, but I realize these can be unreliable when there are large numbers of observations outside the MSA, and the ACS has MetArea instead of MSA.

From what I have been able to find online, my options appear to be:

  • Try combining MSA with MetArea codes (not sure if I can do this)
  • Attempt to aggregate to PUMA level from county (not sure if this would be successful)
  • Try assigning people a county based on their MetArea (probably unreliable)

Am I likely to be more successful in trying to aggregate to Met2013 for the MSA codes, or would something like PUMA make more sense?

2

Due to confidentiality reasons, not all counties are identifiable in post-1950 samples in the individual-level microdata from IPUMS USA, so you are correct that there are a sizeable number of counties without representation in the ACS data (you can read more about it and find a spreadsheet outlining which counties are available in this blogpost). The sub-state geography variables in IPUMS USA, including METAREA, MET2013, and COUNTYFIP, are derived from PUMAs, which can be related to MSAs (read more about that on the description tab for MET2013). IPUMS NHGIS offers aggregate/summary statistics for smaller areas using ACS Summary File data; these data may be preferable depending on your research needs.

If you can tell me a bit more about your HUD data and the information you would like from the ACS, I can suggest resources and/or share considerations for each of the options you presented. For instance, are the HUD data you are working with individual-level or aggregated at the MSA- or county-level? Are you interested in summary statistics for these geographies or do you require individual-level ACS data? If you provide more information about the research application of joining this data, I can provide more targeted information.