Conference Agenda

This paper tests whether the inclusion of an environmental attitudinal battery is a neutral survey component or a consequential design choice that can causally alter welfare estimates in environmental choice experiments. We embed a randomized split-sample experiment in a choice experiment on climate-mitigation measures implemented by accommodation establishments in Mallorca (Spain). Tourists were randomly assigned to a questionnaire in which an environmental attitudinal battery was administered either before or after the choice tasks. Preferences are estimated with panel random-parameters logit models allowing for heteroskedastic scale, and welfare is summarized through distributions of marginal willingness to pay (WTP). Making attitudes salient before the choice experiment produces a large downward shift in welfare estimates: mean WTP for all mitigation attributes in the “attitudes-before” condition is 75–90% lower than in the “attitudes-after” condition. A set of auxiliary checks indicates that the effect is not driven by sample composition, differential task exposure, or respondent fatigue, and is instead consistent with a measurement-induced framing/priming mechanism that changes how respondents construct valuations. The findings imply that attitudinal batteries are not innocuous add-ons: question order can materially affect WTP and compromise comparability across studies and policy appraisals when attitudinal constructs are integrated into CE designs. We therefore recommend that researchers explicitly report and justify attitudinal-block placement and, where feasible, treat question order as a design sensitivity in welfare reporting.