Questions about CPS county code

#1

I was trying to use the county code from CPS data, and found two questions.

  1. Based on the whole CPS data I downloaded from the IPUMS webstie, there are about 280 counties that have county code, i.e. could be identified, after year 2008, however, based on the information provided here: https://cps.ipums.org/cps/codes/count…, there should be about 549 counties that could be identified after 2005. I do not know why I could only identify 280 counties from the original data. I checked some cases, like county code== 6013, which should be identified based on the above list, but I could not find the county code.

  2. For the counties that could be identified, I was wondering if this kind of case exists in the CPS that within the same county, some observations miss county codes while another observations do not. Because I need to aggegate the CPS to the county level and I was worried about the selection biased. If there is no selection in the counties that could be identified, i.e. all the observations in that county have the county code, then the aggregation process should be fine, otherwise, it might have a problem.

Thank you in advance.

#2

The counties listed on the County Codes: August, 2005 Onward page are a list of all counties identifiable in all of the IPUMS CPS combined samples from August 2005 and onward. Therefore, there is no guarantee that each individual sample during this time frame will be able to identify each of these counties. As is stated on the description tab of the COUNTY variable, to preserve respondent confidentiality, not all counties are identified. In recent samples, only about 45% of counties can be identified in each sample. More details about “missing” US counties can be found in this blog post.

Since it sounds like you are interested in data aggregated at the county level, you may find the data available via IPUMS NHGIS to be helpful. Since this project provides aggregated tables, the data are not subject to the same confidentiality restrictions as public use microdata. Therefore, the information available in IPUMS NHGIS is available at the county level for the full set of US counties. Give this project a look and feel free to email ipums@umn.edu if you have any specific questions.