Choice Modelling: Subjective Logic Format

This alternative format can be viewed as an "opinion" in  subjective logic format. See the Wikipedia link below

Daniele implemented this approach along with the standard dichotomous choice framework with and without and opt-out.  Estimation from this approach can proceed by turning the opinion into a probability or log-odds or alternatively viewing it as a Beta distribution which results from a fixed number of binomial trials (along with a uniform prior) if we also assumed that those expressing no uncertainty had a tiny amount of residual uncertainty.  The results in the form a slide are below. We found that people did not find this format any more difficult than the standard format, and because it allows people to express a wider range of beliefs that while the mean WTPs were simiilar, the individual WTPs using this approach (using a Hierachical Bayesian approach) were substantially more diffuse than for the discrete choice approach. This can be seen in the slides below.  We are exploring this approach further in a number of projects. Further results (along with STAN code to estimate the models will follow)