This is a follow-up to a couple questions regarding federal workers. In the IPUMS ACS sample (and CPS sample) there are MANY more federal workers than reported in ACS-IPUMS (empstatd== 10, classwkrd == 25) than .
Part of the reason is that “federal workers” in the ACS (and CPS) includes “employees of foreign governments, the United Nations, and other international organizations.” But this cannot account for the difference between the 3.6 million (CPS) and 2.6 million (OPM).
Any other suggestions on why these numbers could differ so dramatically?
I believe you are addressing the discrepancy between the CLASSWKR variable in IPUMS CPS and CLASSWKR in IPUMS USA. If I’ve misunderstood, please let me know. It seems this discrepancy can be attributed to how each project handles members of the armed forces. As is stated in the Comparability Tab: “Members of the armed forces are subsumed into the general category of federal government employees in the census, beginning in 1960, and in the ACS”. This is not the case in comparable samples within IPUMS CPS.
I had a follow-up question to this. Since the ACS’ CLASSWKR variable subsumes armed forces members into the general category of government employees, is there any other variable that I could use to help me distinguish between non-armed forces federal government employees and armed forces federal government employees? Should I use the EMPSTAT variable along with CLASSWKR to identify active members of the armed forces (empstatd ==14 and Classwkr == 25)?
Yes, you can use EMPSTATD==14 | ==15 to identify individuals employed within the armed forces. This will allow you to distinguish between federal government employees and members of the armed forces in CLASSWKRD.