Published online by Cambridge University Press: 01 January 2025
We compare different implementations of the Stochastic Becker–DeGroot–Marschak (SBDM) belief elicitation mechanism, which is theoretically elegant but challenging to implement. In a first experiment, we compare three common formats of the mechanism in terms of speed and data quality. We find that all formats yield reports with similar levels of accuracy and precision, but that the instructions and reporting format adapted from Hao and Houser (J Risk Uncertain 44(2):161–180 2012) is significantly faster to implement. We use this format in a second experiment in which we vary the delivery method and quiz procedure. Dropping the pre-experiment quiz significantly compromises the accuracy of subject’s reports and leads to a dramatic spike in boundary reports. However, switching between electronic and paper-based instructions and quizzes does not affect the accuracy or precision of subjects’ reports.
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s40881-018-0046-5) contains supplementary material, which is available to authorized users.
We thank Amy Corman, Laboratory Manager at the University of Melbourne’s Experimental Economics Lab. We gratefully acknowledge the financial support of the Australian Research Council through the Discovery Early Career Research Award DE140101014 as well as the Faculty of Business and Economics at the University of Melbourne.