490 – Advances in Statistical Software
Expanding the Number of Primary Sampling Units for the National Health Interview Survey
Chris Moriarity
National Center for Health Statistics
Van Parsons
National Center for Health Statistics
The National Health Interview Survey (NHIS) is a continuous survey that has collected health data using personal interviews since 1957. The NHIS sample design is complex, with multiple stages of sampling, beginning with primary sampling units (PSU) and then additional stages of sampling within each PSU to obtain a sample of addresses. The current NHIS sample design period (2006-2015) has a state-level stratification, but only a limited number of states have adequate sample to support reliable estimation. Recently, additional funding became available for sample expansion in select states. The NHIS has a time-static PSU selection procedure that keeps the first-stage sample areas fixed over the length of a survey sample design period. To increase sampling efficiency, we decided to put resources into expanding the number of sampled PSUs rather than only expanding the within-PSU samples. The preferred method was to extend the probability proportionate to size (PPS) procedure that occurred at the beginning of the current design so that the expansion probabilities also remained PPS. In this paper we discuss issues that arise in such an expansion.