I would like to get a good estimate on the percentage of people living in the US territory earning within a certain INCTOT range, for example INCTOT between 80k and 100k. Should I use the CPS or the USA data?
Thanks in advance!
Either data collection is generally appropriate for the broad use case you have described. When I compare the estimates for persons earning between $80,000 and $99,999 in 2019 using INCTOT in IPUMS USA and IPUMS CPS the estimates are pretty close to one another. You might consider the following criteria for choosing between the datasets:
- Are you interested in other information alongside the INCTOT variable? Is this available in the dataset?
- The two surveys have slightly different target populations; CPS collects data on the civilian, non-institutionalized population. In contrast, the ACS and decennial census data in IPUMS USA are more expansive and would be preferable if you are interested in capturing persons in institutions (e.g., nursing homes, prisons) or the military.
- Note that the sample sizes in IPUMS USA are larger; if you are interested in smaller subpopulations (e.g., narrower income bands, specific geographic locations, certain demographic groups) these data may be preferable.
- Note that ASEC data are used for the official poverty measurement by the Census Bureau.
- ACS is mainly self reported (on a mail-in form) as opposed to the ASEC, where questions are asked by an enumerator.
You may be interested in this Census Bureau comparison of the two datasets.
Thank you very much for your detailed response Kari! 
1 Like