On determining priors for the generation of efficient stated choice experimental designs

Bayesian priors are required in order to generate efficient and robust experimental designs for stated choice surveys. Although such priors are commonly obtained through a pilot study, in this paper we provide a simple alternative in which the analyst depends only on their own expert judgement and possibly on parameter estimates obtained from the literature. The process consists of ranking attribute levels, balancing choice tasks to obtain trade-offs, and setting probabilities in sample choice tasks to establish scale. (C) 2016 Elsevier Ltd. All rights reserved.