118 – Issues, Challenges, and Solutions in Modernizing NHTSA's NASS
Composite Measure of Size Evaluation and Primary Sampling Unit Formation for NHTSA's New National Automotive Sampling System
William Cecere
Westat
Rui Jiao
Westat
Martha Rozsi
Westat
Jacqueline Severynse
Westat
Sharon Lohr
Westat
James Green
Westat
One of the primary objectives of the new National Automotive Sampling System (NASS) sample design is to update and improve the previous NASS' primary sampling unit (PSU) sample. For probability proportional to size (PPS) sampling, finding a composite measure of size (MOS) that is closely related to the multiple outcome variables of interest will reduce the variability of the estimates. In order to achieve this, external information from multiple sources was considered for both the new NASS' General Estimates System (GES) and the Follow-on Passenger Vehicle (FOPV) modules. The external information was used to develop and evaluate multiple composite measures of size against key outcome variables for each module. A MOS was then selected on the basis of correlation with outcome variables and the anticipated variance. The MOS for the secondary sampling units (SSUs) based on obtained crash counts is also presented. The selected MOS was also used to define a minimum MOS for GES PSU formation, while a different minimum PSU MOS variable was used for the FOPV module.