In the November CPS Supplement, the last variable (PUSCK4 in the original and VOTERESP in IPUMS) records if the responses are self-reported or proxy-reported. However, there is also a not-in-universe (NIU) value which is about 13% of the 2020 adult citizens in the November CPS. Most of these are missing answers for other questions in the supplement, but not all are. Who are these people, if not self or proxy reported? At first, I suspected there are people that have left the home, died, etc between survey months, but 13% seems like a lot.
The NIU cases you highlighted are present in the original data from the Census Bureau as well; the team looked for potential reasons for this, but did not find a systematic explanation for why these cases are NIU. The IPUMS CPS team will document this in the VOTERESP variable description. I wish I could have delivered a more definitive answer, but I hope that it’s helpful nonetheless.
Thanks! Do you have advice on who to reach out to at Census to answer this question? You can email me privately, if you wish, via google email with my username dougrhess.
Sure thing! I recommend reaching out to Census at firstname.lastname@example.org. They respond to errors in CPS source microdata and can provide more information.