RIEB Discussion Paper Series No.2019-14

Title

An Empirical Comparison Between Discrete Choice Experiment and Best-worst Scaling: A Case Study of Mobile Payment Choice

Abstract

As an alternative method to discrete choice experiments, best-worst scaling provides additional information about consumers, slightly lessens the burden of mental process, and shows better quality. However, its advantages were ambiguous in previous literature, since each case of the best-worst scaling contained distinct information, and results from comparisons with discrete choice experiment varied with different data. In this study, we applied a goodness of fit statistic named count R square in evaluating the best-worst scaling profile case, the discrete choice experiment, and the best-worst scaling multi-profile case by using data from a survey of preference for mobile payment. The results suggest that the best-worst multi-profile case surpasses other methods. We also compared the mixed logit model and the latent class model using three non-nested tests. The results indicate that the mixed logit model is superior to the latent class model in all three tests.

Keywords

Discrete choice experiment, Best-worst scaling, Goodness of fit, Latent class model, Mixed logit model

Inquiries

Qinxin GUO
Graduate School of Economics, Kobe University
2-1 Rokkodai, Nada,
Kobe, 657-8501, Japan

Junyi SHEN
Research Institute for Economics and Business Administration, Kobe University
2-1, Rokkodai, Nada,
Kobe, 657-8501, Japan
Phone: +81-78-803-7036
FAX: +81-78-803-7059
School of Economics, Shanghai University