Modelling heterogeneity in response behaviour towards a sequence of discrete choice questions: a latent class approach

There is a growing body of evidence in the non-market valuation literaturesuggesting that responses to a sequence of discrete choice questions tend to violate theassumptions typically made by analysts regarding independence of responses andstability of preferences. Heuristics such as value learning and strategicmisrepresentation have been offered as explanations for these results. While a fewstudies have tested these heuristics as competing hypotheses, none have investigatedthe possibility that each explains the response behaviour of a subgroup of thepopulation. In this paper, we make a contribution towards addressing this research gapby presenting an equality-constrained latent class model designed to estimate theproportion of respondents employing each of the proposed heuristics. We demonstratethe model on binary and multinomial choice data sources and find three distinct typesof response behaviour. The results suggest that accounting for heterogeneity inresponse behaviour may be a better way forward than attempting to identify a singleheuristic to explain the behaviour of all respondents.