Multivariate Sample Design Optimization for NHTSA's New National Automotive Sampling System

• Activity Number: 118
• Type: Topic Contributed
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Yumiko Sugawara*+ and Barnali Das and Rui Jiao and James Green
• Companies: Westat and Westat and Westat and Westat
• Keywords: Multivariate sample design ; optimization

Abstract

NHTSA's new National Automotive Sampling System (NASS) is expected to address a large number of research questions and analytic objectives, and therefore requires a multi-purpose study and sample design. In addition, the new NASS consists of multiple modules, with future funding levels and precision requirements unknown and subject to change for any given module. A multivariate sample design optimization system was designed and built for NHTSA to address these design requirements, parameters and constraints. The system offers two options: A) Minimize cost subject to variance constraints; B) Minimize a weighted sum of variances subject to cost constraints. This paper presents the development and architecture of the sample design optimization system, a description of its outputs, the iterative process of reviewing those outputs, and the ability to modify design parameters and constraints.