Sample optimality in the design of stated choice experiments

Abstract

Stated choice (SC) experiments represent the dominant data paradigm in the study of behavioral responses of individuals, households as well as other organizations, yet little is known about the sample size requirements for models estimated from such data. Current sampling theory does not adequately address the issue and hence researchers have had to resort to simple rules of thumb or ignore the issue and collect samples of arbitrary size, hoping that the sample is sufficiently large enough to produce reliable parameter estimates. In this paper, we demonstrate how to generate efficient designs (based on D-efficiency and a newly proposed sample size S-efficiency measure) using prior parameter values to estimate multinomial logit models containing both generic and alternative-specific parameters. Sample size requirements for such designs in SC studies are investigated. Using a numerical case study, we show that using S-efficiency can substantially reduce the sample size required of SC studies.