Weights in ACS, ATUS, CPS ASEC

I have a question about the weights in ACS, ATUS, and CPS ASEC for married couples. I am not sure, but I thought that the weights for both partners in a married family are the same. In other words the weight of one partner equals the weight of the other partner. I think from looking at the data it is true for CPS ASEC and ACS but not for ATUS. Is that true? And if so why are the weights the same for CPS ASEC and ACS but not for ATUS?

At a high level, different datasets have different sample designs and target populations, as such, they include different procedures around sampling weights.

You can find confirmation that weights are adjusted to give married partners (and unmarried partners) the same weight on page 23 of CPS Technical Paper 77 (it is page 32 of the PDF). I did not find explicit confirmation of this being the case in ACS data, but see this in the data. The CPS and ACS are household surveys; they cover the entire household population (though certain measure or supplements in the CPS may be only available for a specific subset). In contrast, the ATUS reports the time use for a single person in the household; as such, the ATUS respondent’s spouse should not have the same weight as they are not included in the time diary data. You can read more about the weighting procedure in the ATUS in the user guide (please note that the variable names will differ between the original Bureau of Labor Statistics data and IPUMS versions).

1 Like

Thanks Kari! That helps a lot!

1 Like