The NHIS documentation highlights a change of sampling design methodology in 2016 and emphasizes that merging data across years requires treating sample design periods as independent. I am analyzing 2007-2016 of the NHIS child data. It looks like IPUMS has already taken the design changes into account when creating the PSU & Strata & weight variables, but I want to make sure I don’t need to do any additional transformation of these variables in order to compare across years? I am doing both point estimates and some regression trend lines.
If I am merging multiple NHIS years, including 2016, do I need to do anything different to survey design vars?
The integrated variables STRATA and PSU in the IPUMS NHIS database have been adjusted from the original NHIS design variables to account for these issues. Thus, you can simply select the STRATA and PSU variables to use for analysis of one year or for many years of IPUMS NHIS data.
However, when pooling multiple years of data, sample weights will need to be adjusted. This documentation provides some information sampling weights and pooling data. Additional weighting information, which includes information about combining sampling weights when a variable is located in different files across years, can be found here.