Comparing contingent valuation and choice modeling using field and eye-tracking lab data

Abstract

We compare contingent valuation and choice modeling with field and eye-tracker data. Contrary to previous research, results yield significantly different structural models. These divergences remain for modified formats that minimize visual and cognitive differences between formats. We also find divergent results concerning information processing. In choice modeling, respondents devote more time to attributes, including the bid, although total answering time does not vary. Presenting several questions with varying attribute levels works for choice modeling but not for contingent valuation. Using the attribute-stimulus format does not change contingent valuation results. Dominated alternatives increase the probability of paying in choice modeling