Lexicographic discontinuous choices Lexicographic choices Respondents base their

Lexicographic / discontinuous choices

Lexicographic choices § Respondents base their choice on a subset of the presented attributes § Continuity axiom is violated = no tradeoff between the attributes § Biased welfare estimates

Sources § Choice heuristics complexity) § Protest responses pay vehicle) § Design issues levels) § True preferences (choice set (paying for env, (causal att,

Some empirical work § Hensher, Rose and Greene, 2005 § Campbell, Hutchinson and Scarpa, 2007 § Carlsson, Kataria and Lampi, 2008

Hensher et al. (2005) “The implications on willingness to pay of respondents ignoring specific attributes” § Study = car commuters in Sydney § Model = Mixed logit model with individual specific coefficients § Coefficients are restricted to zero if the respondent ignored that attribute § Results = restricted model leads to 18 -62% higher value of time

Campbell et al. (2007) “Incorporating Discontinuous Preferences into the Analysis of Discrete Choice Experiments” § Study = rural environmental landscapes in Ireland § Model = ECM allowing for differences in scale and error variance between subsets of respondents. § Also weighting of attributes § Findings: • Error variance in discontinuous subset is significantly higher

Carlsson et al. (2008) “Ignoring attributes in choice experiments” § Study = three environmental objectives in Sweden § Model = RPL with parameters restricted to 0 if ignored § Three assumptions in estimating WTP: • All respondents positive WTP • Ignoring env attribute zero WTP • Ignoring costs and env attribute zero WTP § Findings: • No systematic differences between restr and unrestr model

CM challenges § Survey design? • Complexity • Follow-up questions § Reasons for choice behaviour? • Choice heuristics • True zero WTP § Analysis? • Econometric models • Respondents who ignore cost attribute