Hi. I am using the CPS from 1976-1990 and am trying to remove all individuals who were in the armed force. I am using the monthly CPS data. I have looked at different variables (empstat, occ1990, classwkr) to try and identify this group, but it looks like this data does not exist before 1984. Is there anyway I can identify individuals in the armed forced from 1976-1983?
The variable POPSTAT reports if a person is an adult civilian, in the Armed Forces, or a child. It is available for ASEC samples only in 1962-1983, and both basic monthly and ASEC samples in 1984-forward. Let me know if you are looking for basic monthly coverage of this variable before 1984 and I can look further.
Thanks for sharing this information. Having this information for the ASEC months is helpful, but does not cover my whole sample since I am using the CPS basic monthly data for each year. It would be helpful if you could look further to find out if this variable exists before 1984 for the basic monthly survey.
I am following up to report that I haven’t found any obvious candidate variables for finding this information in these earlier years. I have a request out to the IPUMS CPS team to see if they have ideas, but suspect if this information was available, we would have harmonized it as part of POPSTAT. For a subset of cases, you could leverage the panel component of CPS to link their ASEC response to the BMS data and append the POPSTAT value there; however there are linking limitations in 1976-1978, and you won’t be able to cover the full BMS sample (i.e., only a subset of the ASEC respondents can be linked to other months of data) or all months of data (i.e., because of the CPS rotation pattern, you will not have any cases in July-November).