Abstract:
|
The National Health Interview Survey (NHIS) is a cross-sectional survey conducted by the National Center for Health Statistics (NCHS). The sampling plan of NHIS follows a multistage probability design that permits representative sampling of the civilian, noninstitutionalized U.S. population. The NHIS selects one adult and one child (when applicable) randomly within a family, and sampling weights for sample adults and sample children have been developed by NCHS. In recent years, NCHS has received inquiries on analyzing parent-child pair data using NHIS. However, sampling weights for the selected pairs, in particular, parent-child pairs, have not been developed. In this research, we propose to derive the sampling weights for the adult-child pair data. We then identify parent-child pairs from the sampled adult-child pairs and conduct domain analyses for inferences about parent-child pairs. The proposed procedure is useful for studies using the NHIS parent-child pair data. It could also be adapted to other national surveys with similar statistical needs, in which one or more people within a family having a specified relationship have been sampled independently of such relationships.
|