Choosing the appropriate weight variable for NHIS 2015 data analysis

Hello, I need some guidance on selecting the appropriate weight variable for my analysis using the 2015 IPUMS NHIS data. My analytic sample is limited to individuals aged 18 and above. There are three different weight variables associated with the variables in my project: PERWEIGHT, SAMPWEIGHT, and SUPP1WT. Could you please advise if SAMPWEIGHT would be the most suitable choice for all variables? Thank you! :slight_smile:

These are three different weights that should be used in different circumstances when working with NHIS data from 2015. Which weight you should use depends on which variables you are analyzing. I encourage you to read the descriptions of each weight and read this user note about weights in the NHIS to determine which weight is appropriate to use when analyzing NHIS data. If you are performing analysis that uses variables from multiple categories (i.e., a variable that is only available for sample adults and a variable that is available for all persons), you should use the more restrictive weight. The more restrictive weight is the one that is applicable to the smaller number of people. In that example, the more restrictive weight would be the one that is used for variables available only for sample persons (SAMPWEIGHT). Using the most restrictive weight ensures you are weighting the analysis based on the probability of having a response to all of the variables used in the analysis.

The weight SAMPWEIGHT should be used with person-level analyses of variables that are only available for sample adults and/or sample children. This weight represents the inverse probability of selection into a sample adult/child supplement.

The weight PERWEIGHT should be used with person-level analyses of variables in which information was collected on all persons.

The weight SUPP1WT should be used with variables from the Functioning and Disability Supplement.

1 Like

Thank you, Isabel! I really appreciate you sharing the additional resources. :slight_smile: