Developing Sampling Weights for Statistical Analysis of Parent-Child Pair Data From the National Health Interview...

Area of Science:

• Public Health
• Biostatistics
• Survey Methodology

Background:

• The National Health Interview Survey (NHIS) is a primary data source for U.S. health trends.
• Existing NHIS weights are designed for individual adult or child analyses.
• There is increasing demand for analyzing parent-child dyad data within NHIS.

Purpose of the Study:

• To develop and describe methodology for creating analytic weights for parent-child pairs in the NHIS.
• To enable researchers to conduct robust analyses of dyadic data from the NHIS.

Main Methods:

• Adult-child pair-level sampling weights were constructed using conditional selection probabilities and household weights from 2019 NHIS data.
• Weights were adjusted for pair-level nonresponse.
• Large sampling weights were trimmed at the 99th percentile.

Main Results:

• Dyad weights were successfully created for analyzing parent-child pairs within the NHIS.
• The report provides examples of domain estimation methods for dyadic analysis.

Conclusions:

• The National Center for Health Statistics has developed usable dyad weights for NHIS parent-child pair studies.
• This methodology may be adaptable to other surveys with similar designs and analytical requirements.