Question on CPSIDP


I have two questions:

  1. If I want to use CPS as a panel dataset, is cpsidp sufficient to be the unique person ID throughout the whole dataset, or I should combine it with some other variables to generate a real unique person ID that can be used for linking purpose?
  2. When I am looking at 1977m1 and 1977m2, there’s no duplicates for cpsidp within those two month, which means no cpsidp was observed consistently for two months (they all show up once only in two months). Is that possible?


I’ll address each of your questions one at a time.

(1) Yes, CPSIDP uniquely identifies individuals across CPS samples. This variable allows linking of respondents, based on the 4-8-4 rotation pattern. CPSIDP will only ever appear a maximum of 8 times, which is the number of times a household may be observed in the CPS survey. In some cases, individuals will appear fewer than 8 times due to migration, mortality, non-response, and recording errors. Extensive documentation about the creation of CPSIDP is available here.

(2) As you point out here, there are several time intervals where linking individuals across samples is not possible. As is noted on page 126 of this paper:

Data from 1962 to 1978 present the most serious linkage challenges. Each housing unit was assigned a unique identifier during most (but not all) years in this period, but person-level identifiers do not reliably identify the same individual in multiple samples. […] Furthermore, because of changes in the numbering scheme for housing units, household-level identifiers cannot be used to link housing units between 1962 and 1963, 1971 and 1972, 1972 and 1973, and 1976 and 1977.



Hi Jeff,
Thank you very much for the quick response!
A follow-up question to Q2: in the paper it says linking between 1976 and 1977 is not possible. does it mean within 1977, the linking for month over month (i.e., 1977m2 and 1977m3) is also not possible?



Yes, unfortunately the 1977 samples do not perform well to linking with giving linking keys. More detail about linking records with CPS data can be found on our 2018 summer workshop page.