Attribute Framing in Choice Experiments: How Do Attribute Level Descriptions Affect Value Estimates?

Choice Experiments present survey respondents with alternative options that are described by a number of attributes. Respondents are assumed to evaluate each option based on the levels of the attributes, which vary across alternatives and choice sets. The way in which attributes are described to respondents is likely to affect their choices. In this study, the impacts of two attribute level descriptions are assessed: describing non-market attributes as absolute levels or in relative terms; and using positive versus negative contextual descriptions of attribute levels. These tests were performed using data from a choice experiment on catchment management in Tasmania, Australia. Contrary to a priori expectations, including explicit information cues about relative attribute levels in the choice sets is not found to affect stated preferences. The data do reveal significant differences in value estimates when attribute levels are described as a 'loss', compared to a 'presence'.