Merge/match consecutive years of ASEC datasets (longitudinal design)


Hello everyone,

I am facing a problem while using ASEC datasets for year 1968-2015. Ideally, I would need to merge/match consecutive ASECs by household and person. I would like to know if this is possible for the years I need to analyse.

I am currently using IPUMS-CPS-ASEC.

I don’t get if it is possible to match two consecutive ASEC of not. It seems to me that there is the possibility of matching IPUMS-ASEC with IPUMS-MarchCPS. Some other combinations by using NBER and IPUMS as sources.

Those are some link I read:

Q&A, Q&A1, Q&A2, and so on.

Thanks in advance



The short answer is, yes it is possible to make a panel out of the ASEC CPS samples. You first need to link each ASEC sample to its corresponding March Basic sample using MARBASECIDP (for persons) and MARBASECIDH (for households). Next you can link basic monthly samples using CPSIDP (for persons) and CPSID (for households). Much more information about how to perform this linking can be found in this working paper.

The more complicated answer is, these variables are only available for ASEC samples from 1989-2015. The IPUMS-CPS Team is working on making these variables available for all years in the CPS, but it will be several months before these variables are available. For pre-1989 samples you can try to replicate the procedure described in the working paper (linked above). Based on some preliminary work, however, we know that you’ll need to prune the ASECs for duplicates prior to merging across years and that you’ll probably need to include AGE, SEX, and RACE as linking keys.
I hope this is helpful.



Thank you very much Joef! Let’s hope the Ipums project team will deliver this new variable to merge ASEC before 1989 very soon.

I switched to NBER CPS (March). I might ask you about a problem I am facing when creating the Stata file. By using the .dat, the .dct and the .do provided on NBER website, I get a strange warning message as labels are already defined. It never happened to me before.

Do you have a solution for this?

I can be more specific if needed.



I’m not entirely certain if you are trying to merge NBER data with IPUMS data or if you are just trying to download NBER data. If you are trying to merge NBER with IPUMS try using the “nolab” option on the merge command. If you are having issues with NBER data, then I’d suggest that you contact NBER for a more detailed answer.