I have a question about no and niu response in INCLUGH. It looks like in 1999 and 2009, for everyone who worked in the last year (WORKLY=2), yes responses are the same for these two variables. However, all niu responses in INCLUGH became no responses in GRPOWNLY. So, a) for 2019 data, can I recode the niu responses in GRPOWNLY as no responses? b) in years prior to 1999, is it safe to assume that niu responses to INCLUGH can be treated as no responses?
Thanks for posting. We are looking into your question at IPUMS HQ and will report back soon!
INCLUGH and GRPOWNLY measure the same thing, namely that an individual was the policyholder for a private, employer-based health plan last year. They will have the same YES answers. The only difference is in the universe.
Post-1995, the universe for INCLUGH is individuals who had private health insurance in their own name (i.e. were policyholders) last year. INCLUGH then identifies which of these people were in an employer-based plan (as opposed to individually purchased insurance). The universe page for INCLUGH currently has an error for this time period, and this will be fixed in the next data release. The correct universe for 1995+ is the same as 1988-1994.
In contrast, for GRPOWNLY, the universe for 1996-2018 is “All persons age 15+”. This identifies for everyone older than 15 whether they were the policyholder for an employer-based health plan. This implies that anyone who is 15+ and who was NOT the policyholder for a private insurance plan will have GRPOWNLY=NO and INCLUGH=NIU. So to answer your question, yes you can assume all NIU responses for INCLUGH do NOT have employer-based insurance policies in their own name, either because they were not covered by a private health plan at all, because they were not the policyholder of that plan, or because they were under 15 so presumably cannot be a policyholder. However if this is what you’re interested in, you should just use GRPOWNLY (simply recoding NIUs to ‘No’ responses conflates those who are not policyholders of employment-based plans with those who are not eligible to be policyholders of such plans). Note that this is also true of GRPOWNLY in 2019 - the NIU category includes those who are not covered by private insurance, who are not the policyholder for their private insurance, or who are too young to be policyholders.
Other useful variables for identifying the proper universes are COVERPI, COVERGH, and PHIOWN.