Estimating willingness to accept using paired comparison choice experiments: tests of robustness/ created by David Kingsley and Thomas C. Brown

Paired comparison (PC) choice experiments offer researchers and policy-makers an alternative nonmarket valuation method particularly apt when a ranking of the public's priorities across policy alternatives is paramount. Similar to contingent valuation, PC choice experiments estimate the total value associated with a specific environmental good or service. Similar to choice experiments, the questions posed to respondents are choices between alternatives. In contrast to both methods, respondents in PC choice experiments make choices between pairs of dissimilar alternatives including private goods, public goods, and monetary amounts. The alternatives may include competing policy alternatives, thus providing a ranking of the public's priorities among those alternatives. We investigate the robustness of estimated welfare measures to econometric modelling and choice set composition across two PC choice experiments. Results suggest that accounting for repeated observations increases the efficiency of welfare estimates but also reveals, contrary to previous research, sensitivity to choice set composition. Thus, while PC choice experiments may be advantageous in certain situations the results presented here suggest that further research is needed to better understand the sensitivities of the resulting welfare estimates.