Beginning with the 1995 ASEC, all individuals with an EMPSTAT value of 32 (NILF, unable to work) are coded as “not wanting a job” for the WANTJOB variable. However, this is not the case for prior ASEC years.
For the 1995 and later ASEC files, do these “No” values represent actual responses to WANTJOB or does they reflect assignment by CPS and/or IPUMS CPS? In the IPUMS and CPS documentation I have seen, there does not seem to be clear indication that these values were assigned, but given consistency from 1995 onward and the discrepancy with earlier years, this seems plausible.
Note that EMPSTAT code==32 “NILF, unable to work” is only available starting in 1995. In years prior, the NILF codes were: 30 “NILF”, 34 “NILF, other”, and 36 “NILF, unpaid, less than 15 hours”. Therefore, from 1995 onward it is reasonable to expect that those who reported that they were not able to work and therefore not in the labor force also report not wanting a regular job. Note that in all years (e.g. both before and after 1995) WANTJOB==1 “No” corresponds to each of the available NILF codes available via EMPSTAT. There is a data quality flag for WANTJOB (see QWANTJOB) but it is only available from 1990-1993. Therefore, these values do represent actual responses by respondents, rather than imputations.
EMPSTAT code==32 “NILF, unable to work” is also available from ASEC 1962 through ASEC 1987 (it is unavailable for ASEC 1988 through ASEC 1994). For ASEC 1976 through ASEC 1987 (years for which WANTJOB and EMPSTAT code == 32 are both available), a number of individuals with EMPSTAT code == 32 have WANTJOB codes == 2 “Yes” or 3 “Maybe, it depends”.
While it does seem plausible that individuals reporting that they were unable to work also do not want a job, it seems odd that there was a marked change in the prevalence of “Yes” or “Maybe” WANTJOB responses, which ranged from 15 to 30 cases per year prior to 1988 to 0 cases for “Yes” ("Maybe is not available) after 1994.
Ah, good catch! I hadn’t selected samples far enough back. It seems that the “NILF, unable to work” category isn’t perfectly comparable between when it was first available and its re-introduction. Looking back at the coding schemes from the 1980s it looks like those who are involuntarily “not in the labor force” are included. In 1995 and onward, however those who are not in the labor force, but want to work are not included in the “NILF” categories. The IPUMS CPS team will add a note about his on the Comparability Tab of EMPSTAT.
We like to reward users, who report inconsistencies or errors in the data such as this, with an IPUMS coffee mug. If you email your mailing address to ipums@umn.edu I’ll get you mug in the mail.