Attribute processing in environmental choice analysis: implications for willingness to pay

Data from a discrete choice experiment is used to investigate the implications of failing to account for attribute processing strategies (APSs). The research was designed to elicit the economic benefits associated with landscape restoration activities that were intended to remediate environmental damage caused by illegal dumping activities. In this paper we accommodate APSs using an equality constrained latent class model. By retrieving the conditional class membership probabilities we recover estimates of the weights that each respondent assigned to each attribute, which we subsequently use ensure unnecessary weight is not allocated to attributes not attended to by respondents. Results from the analysis provide strong evidence that significant gains in models fit as well as more defensible and reliable willingness to pay estimates can be achieved using when the APSs are accounted for.