BENEFIT ESTIMATES FOR LANDSCAPE IMPROVEMENTS: SEQUENTIAL BAYESIAN DESIGN AND RESPONDENTS’ RATIONALITY IN A CHOICE EXPERIMENT STUDY

A multi-attribute stated preference approach is used to value low and high impact actions on four major landscape components addressed by the Rural Environment Protection (REP) Scheme in Ireland. Several methodological issues are addressed: the use of prior beliefs on the relative magnitudes of parameters, standardized description of different levels of landscape improvements via image manipulation software, adoption of efficiencyincreasing sequential experimental design, and sensitivity of benefit estimates to inclusion of responses from “irrational” respondents. Amongst other things, our findings indicate that Bayesian design updating can deliver significant efficiency gains, and that estimates may be up-ward biased when irrational respondents are not excluded.