Hi,
I’m working with the 2024 ASEC and trying to merge to the March-June basic monthly files.
I’m interested in the EITC and have a handful of cases I’m trying to work through where results seem a bit unusual to me based on the family relationships and income values.
One of them is CPSID 20240304044600, CPSIDP 0, RELATE 101 as the Head/householder, ASECOVERH = 0, and ASECOVERP = 1. The other three members of the household are ASECOVERH = 0 and ASECOVERP = 0.
The two children in the household show up with RELATE codes of 1260 in the ASEC and March basic.
However, when the same CPSID reappears in months 5 and 6, the two children (same ages) show up with RELATE codes of 301. The head of household has CPSIDP 20240304044604 for months 5 and 6.
So my question is whether this is a data issue or if there’s some scenario in which a person would be part of the ASEC oversample but not other members of their family/household? And if so, what is that scenario? Or is there another possibility I’m missing?
Thanks so much.
Thanks for your question and a specific household to illustrate the issue. I can confirm that I also see the householder listed as being part of the ASEC oversample in their 2024 ASEC as per their ASECOVERP value despite their ASECOVERH value indicating they are present in the March Basic. In general, you should expect a person’s oversample status to match that of their household; while a household may be selected into the oversample based on the characteristics of a single person (e.g., child under 18; person of Hispanic origin), the entire household is selected to participate. In 2024, we were unable to link a small number (<200) of records between the March BMS and ASEC samples because of an issue in the original CPS data; the example you provide matches the criteria of these (currently) unlinkable records (householders with month-in-sample values of 5). Note that in some of the earlier years of the CPS (1976-1988) we are also unable to link all records between the BMS and ASEC and there are cases in these years where ASECOVERP and ASECOVERH values do not align.
You asked a secondary question about the relationship to the householder of the two household children – categorized as 1260: Other relative in the first observation for this household but as 301: Child in subsequent BMS observations. Note that the household also includes an unmarried partner of the householder. My best guess is that the household children are the biological children of the unmarried partner. It seems feasible that either the in-person nature of the first interview captured this detail which was obscured in subsequent interviews (the first and, in some cases, fifth interview in the CPS are conducted in-person while the remainder are on the phone). Alternatively, it is possible that this is capturing a change in how the household respondent thought of the children in relation to the householder. While I cannot say definitively that this is the situation, it seems reasonable to me.