longitudinal identifiers after 2000


can I match individuals in IPUM-CPS OR IPUM-ACS across waves after 2000? Or maybe across waves of IPUMS-USA (ACS)? NBER only does it (for CPS) till 1998


It is not possible to track people across time using the ACS. Since the ACS was not designed to track households across time the way the CPS was, there are no identifiers created for this purpose. Furthermore, once a household is included in an ACS sample it is not allowed to be sampled again for five years.

The methods NBER uses for linking March Samples across years can be applied to beyond 1998 since the survey has not undergone a major restructuring since 1994. See this answer for more information on using this method with IPUMS-CPS data.

I hope this helps.

But in the NBER data extracts for years 2000-2007 (the ones I tred to use so far) the variable MIS, which is crucial for the NBER matching algorithm, is not included… what are the right datasets to download? there are so many sources, it is pretty confusing

In the NBER created Stata .do and .dct files, the variable MIS (Month In Sample) is named h_mis.

I hope this helps.