COMBINING CHOICE EXPERIMENTS WITH CONTINGENT VALUATION AND THE FRISCH ELICITATION METHOD

This paper presents a survey that involved three different stated preference methods: choice experiments, contingent valuation, and Frisch elicitation. The survey was designed to assess the valuation of various quality aspects of public transport trips. Although choice experiments is generally the most common and preferred stated preference method, there are noted shortcomings related to respondent response, e.g. lexicographic answering. These shortcomings warrant more testing. To this end, this paper compared the three different methods in the context of parameter estimate convergence. We show that for this sample, the three methods produced similar parameter values. Moreover, our study examined the potential for easing the response task through the simultaneous application of methods. We used contingent valuation results to "balance" the attribute levels in the choice experiments. The effects of balancing were measured with respect to (1) the share of lexicographic answering, (2) the share of respondents expressing that the experimental choices were difficult to answer, and (3) the estimates of attribute values. it is shown that contingent valuation to balance attribute levels in the choice experiments significantly decreased the share of respondents answering lexicographically.