As the website suggested, the value 3 of STEPPOP (explicitly specifying a step-parent information) is not available in the ACS samples. However, there is indeed a small portion of observations coded as stepmom/stepdad == 3 in 2023. I wonder how the coding process worked.
Thanks for your question. STEPMOM and STEPPOP were inadvertently released for the 2023 ACS and PRCS PUMS samples; we do not generally offer these variables in the modern samples. We are in the process of removing these variables from IPUMS USA 2023 samples. Users who included STEPMOM and STEPPOP in their extracts should NOT use these variables.
I will provide a bit of background on why these variables are not offered in the modern samples from IPUMS USA. STEPMOM and STEPPOP flag cases in which a parent and child are likely to be related through a non-biological relationship. While the variables do a good job of flagging certain combinations of variables that suggest that a parent is a step-parent (e.g., improbable age differences between a parent and a child or certain relationships combinations in a household), there are many other cases in which step-parents cannot be flagged by STEPMOM and STEPPOP. For example, in many cases in which a step-parent and step-child have an age difference that a biological parent and child could also have, the parent will not be flagged as a step-parent. Because of the potentially substantial share of step-parents who are not identified in STEPMOM and STEPPOP, IPUMS has decided to ultimately not re-design or re-release these variables for samples from 1970-onward. We want to ensure that researchers are not under the impression that all step-parents can be identified using these variables, or that these variables could be used to identify biological parents or infer fertility.