Development of an Analytical Method to Discriminate Respondent Attribute Processing Strategies: Application in a Choice Experiment

Abstract

In the choice experiment framework, it is assumed that respondents consider all the trade-offs among attributes when making their choices. However, there is evidence that respondents may not consider all of them. In the choice-experiment survey to value public goods reported here, a methodology has been developed for analytically determining the value of the coefficient of variation of the conditional distributions to discriminate respondents attending/ignoring one attribute. Results show that (at the respondent level) the analytically determined critical values exhibit a better concordance with the stated preferences than if the rule of thumb with an arbitrary value of two (assumed in previous research) is used.