NHIS: paid sick leave variable

I’m confused by who is considered NIU for the PAIDSICK variable. The description of the variable says that the universe should be those who worked last week, but there are YES responses in all the EMPSTAT categories. Below is a cross-tab of EMPSTAT and PAIDSICK from the 2018 dataset for Region 1.

The universe for PAIDSICK is as follows:

  • 1997-2010: Sample adults age 18+ who worked during the previous week.
  • 2011-2018: Sample adults aged 18+ years who were working last week, or who were with a job or business but not at work, or who were working, but not for pay, at a family-owned job or business, or who have ever worked.

The most important detail here is that the PAIDSICK variable is only applicable to “sample adults.” A number of questions in the NHIS are only asked of a randomly selected adult within households. These randomly selected adults, noted as “sample adults,” are asked a longer set of questions—such as PAIDSICK. So, in addition to the employment restrictions on the universe the PAIDSICK variable is also defined by ASTAFLG, the sample adult identifying variables.