Estimating stated preferences with rated-pair data: Environmental, health, and employment effects of energy programs

Respondents' stated preferences for attributes related to various electricity-generation scenarios are analyzed using a series of pairwise ratings. Multiple observations for each respondent facilitate estimating individual scale parameters. Scale estimates can identify uninformative rating patterns and make it possible to delete such observations or adjust WTP calculations. Cross-section/time-series analysis of the data indicates that nonprice attributes increase in salience during the course of the experiment. Thus later responses may be better indicators of preferences than earlier responses. Comparisons of polychotomous with dichotomous models indicate that most respondents' relative-preference ratings are not simply dichotomous, but discriminate systematically along the rating scale.

Duke Scholars

Author F. Reed Johnson Population Health Sciences