Does using the INCLUGH variable over time requires correcting for the VERIFY=1 cases?


I have a question about using the ASEC variable INCLUGH over time.

Except for the universe, INCLUGH is virtually identical to GROUPOWN (since 1996, which is the first year GROUPOWN is available).^1 Clicking on the questionnaire text link for each variable does indeed point to the same question in the ASEC questionnaire.

Now, the documentation for GROUPOWN says that, if one wants to use GROUPOWN for years before and after 2001, one should adjust the values since 2001 for the fact that in 2001 the VERIFY variable was introduced (this variable captures cases of health insurance previously undetected).

So, if I want to use INCLUGH since, say, 1993 to 2005 then, given that INCLUGH is the same as GROUPOWN (since 1996 and except for the universe) and given that using GROUPOWN over time requires adjusting for the verification question, should not I also adjust the values of INCLUGH since 2001 for the verification question? If so, then you may want to add this fact to the documentation for INCLUGH.

Thanks as usual for your help.

^1 More precisely, my tabulations show that INCLUGH is a “Yes” if and only if GROUPOWN is a “Yes” while the No and the NIU values may differ because of the difference in universe.


Yes, I think you are correct. Since (as you point out) GROUPOWN technically defines the universe for INCLUGH, the same details about comparability over time that apply to GROUPOWN also apply to INCLUGH. I’ll send a note to the IPUMS CPS team about clarifying this in the documentation.