Recently, I tried to merge the IPUMS CPS data with another edition of CPS data and found that HRHHID2 was different between the datasets. When I subsequently checked the original Census releases, it matched the non-IPUMS dataset.
Does anyone know why HRHHID2 is different in IPUMS, and how I can convert it back to match the original Census release?
I suspect that this is a formatting difference rather than an actual difference in HRHHID2 values. The IPUMS version of the variable is a right-justified long data type with 12 columns (in Stata, the format in IPUMS is
long %12.0g). I suggest checking the format between the two datasets and bringing them into alignment. If this doesn’t work, it would be helpful if you could tell me which samples you are using so I can look into the issue more closely.