Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

‘The usefulness of Bayesian optimal designs for discrete choice experiments’ by Roselinde Kessels, Bradley Jones, Peter Goos and Martina Vandebroek

‘The usefulness of Bayesian optimal designs for discrete choice experiments’ by Roselinde... 1 Introduction The article by Kessels et al. provides a very important contribution to optimal design for discrete choice models. By means of a case study the authors investigate the performance of Bayesian designs for this statistical method. Optimal designs enable precise parameter estimates for discrete choice models. Because the underlying multinomial logit model is nonlinear, no globally optimal designs can be developed. Mostly, locally optimal designs based on a zero vector for the regression weights are used because these designs correspond to well‐known designs for linear models. A severe drawback of such designs for discrete choice models is the very unrealistic assumption that all alternatives in a choice set have the same utility. Such designs are called utility‐neutral designs. To circumvent this problem, designs using prior information for the unknown parameters may be developed. These designs are usually called optimum or efficient Bayesian designs even if they do not correspond to a full Bayesian approach. Such Bayesian designs are considered by Kessels et al. and compared with utility‐neutral designs. There is some literature about comparing Bayesian designs and utility‐neutral designs for discrete choice models. A first influential article was published by Sándor and Wedel . Kessels et http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Applied Stochastic Models in Business and Industry Wiley

‘The usefulness of Bayesian optimal designs for discrete choice experiments’ by Roselinde Kessels, Bradley Jones, Peter Goos and Martina Vandebroek

Download PDF
4 pages

Loading next page...
 
/lp/wiley/the-usefulness-of-bayesian-optimal-designs-for-discrete-choice-wf3bZw3054

References (4)

Publisher
Wiley
Copyright
Copyright © 2011 John Wiley & Sons, Ltd.
ISSN
1524-1904
eISSN
1526-4025
DOI
10.1002/asmb.904
Publisher site
See Article on Publisher Site

Abstract

1 Introduction The article by Kessels et al. provides a very important contribution to optimal design for discrete choice models. By means of a case study the authors investigate the performance of Bayesian designs for this statistical method. Optimal designs enable precise parameter estimates for discrete choice models. Because the underlying multinomial logit model is nonlinear, no globally optimal designs can be developed. Mostly, locally optimal designs based on a zero vector for the regression weights are used because these designs correspond to well‐known designs for linear models. A severe drawback of such designs for discrete choice models is the very unrealistic assumption that all alternatives in a choice set have the same utility. Such designs are called utility‐neutral designs. To circumvent this problem, designs using prior information for the unknown parameters may be developed. These designs are usually called optimum or efficient Bayesian designs even if they do not correspond to a full Bayesian approach. Such Bayesian designs are considered by Kessels et al. and compared with utility‐neutral designs. There is some literature about comparing Bayesian designs and utility‐neutral designs for discrete choice models. A first influential article was published by Sándor and Wedel . Kessels et

Journal

Applied Stochastic Models in Business and Industry – Wiley

Published: May 1, 2011

There are no references for this article.