nep-dcm New Economics Papers
on Discrete Choice Models
Issue of 2010‒05‒22
ten papers chosen by
Philip Yu
Hong Kong University

  1. Nonparametric vs Parametric Binary Choice Models: An Empirical Investigation By Bontemps, Christophe; Racine, Jeffrey S.; Simioni, Michel
  2. Testing heterogeneous anchoring and shift effect in double-bounded models: The case of recreational fishing in Tasmania By Jennings, Sarah; Rust, Steven; Yamazaki, Satoshi; Lyle, Jeremy
  3. Framing for incentive compatibility in choice modelling By Mazur, Kasia; Bennett, Jeff
  4. Strategic response to a sequence of discrete choice questions By McNair, Ben; Bennett, Jeff; Hensher, David A.
  5. The Impacts of Attribute Level Framing and Changing Cost Levels on Choice Experiments Value Estimates By Kragt, M.E; Bennett, J.W
  6. Comparing responses from web and paper-based collection modes in a choice modelling experiment By Windle, Jill; Rolfe, John
  7. Valuing protection of the Great Barrier Reef with choice modelling by management policy options By Rolfe, John; Windle, Jill
  8. Understanding values associated with stormwater remediation options in marine coastal ecosystems: A case study from Auckland, New Zealand By Bastone, Chris; Stewart-Carbines, Megan; Kerr, Geoff; Sharp, Basil; Meister, Anton
  9. Climate change and Australiaâs comparative advantage in wheat By Sanderson, Todd; Ahmadi-Esfahani, Fredoun Z.
  10. Effects of alternative elicitation formats in discrete choice experiments By Scheufele, Gabriela; Bennett, Jeff

  1. By: Bontemps, Christophe; Racine, Jeffrey S.; Simioni, Michel
    Abstract: The estimation of conditional probability distribution functions (PDFs) in a kernel nonparametric framework has recently received attention. As emphasized by Hall, Racine & Li (2004), these conditional PDFs are extremely useful for a range of tasks including modelling and predicting consumer choice. The aim of this paper is threefold. First, we implement nonparametric kernel estimation of PDF with a binary choice variable and both continuous and discrete explanatory variables. Second, we address the issue of the performances of this nonparametric estimator when compared to a classic on-the-shelf parametric estimator, namely a probit. We propose to evaluate these estimators in terms of their predictive performances, in the line of the recent ”revealed performance” test proposed by Racine & Parmeter (2009). Third, we provide a detailed discussion of the results focusing on environmental insights provided by the two estimators, revealing some patterns that can only be detected using the nonparametric estimator.
    Keywords: binary choice models, nonparametric estimation, specification tests
    Date: 2009–12–16
  2. By: Jennings, Sarah; Rust, Steven; Yamazaki, Satoshi; Lyle, Jeremy
    Abstract: This paper explores the extent and nature of anchoring and shift effects in a double-bounded contingent valuation of recreational fishing in Tasmaniaâs inshore saltwater fishery. In particular we model the situation where respondents, when answering the second valuation question, evaluate the bid amount partly with reference to the size of the first bid amount. The estimates of the coefficients and mean WTP for a day of fishing are compared across different contingent valuation models, including a single-bounded model, a conventional double-bounded model and models that control anchoring and exogenous shift effects in both homogeneous and heterogeneous forms. Overall we find consistent evidence of anchoring, but mixed evidence of a shift effect. Results show that both males and females anchor in the same way, but that respondents who have a mainstream view of what recreational fishing represents anchor more strongly than those whose view of fishing is not mainstream. The estimated mean WTP for a dayâs recreational fishing is consistently higher in all models which account for bias in responses than in either the single-bounded or double-bounded models. We indicate the possibility that anchoring behaviour may be more complex than is captured in our models and suggest that this needs to be addressed if the results of contingent valuations are to reliably inform resource allocation decisions and recreational fishing management.
    Keywords: Contingent valuation, anchoring bias, shift effect, heterogeneity, recreational fishing, Environmental Economics and Policy, C35, Q26,
    Date: 2010
  3. By: Mazur, Kasia; Bennett, Jeff
    Abstract: The incentives that motivate respondents to reveal their preferences truthfully have been a long-standing area of research in the non-market valuation literature. A number of studies have been undertaken to investigate incentive compatibility in nonmarket valuation. Most of these used laboratory environments rather than field surveys (e.g. Carson and Burton, 2008, Harrison, 2007, Lusk and Schroeder, 2004, Racevskis and Lupi, 2008). Only a few studies investigating incentive compatibility have considered multi-attribute public goods with an explicit provision rule in a choice experiment (Carson and Groves, 2007, Collins and Vossler, 2009, Carson and Burton, 2008). The design of a choice modelling study that avoids strategic behaviour has proven particularly difficult because of multiple choices and difficulties in developing a majority voting provision rule. This study investigates the impact of the inclusion of a framing statement for incentive compatibility in a field survey choice modelling study. An incentive compatible statement (provision rule) that sets out to respondents the rule relating to when the good under consideration will be provided was employed. The impact of a provision rule across three alternative choice modelling multiple choice questionnaires was tested by comparing results between split samples with and without a provision rule. Four split samples were used to test the impact of a provision rule on preferences across different communities including local/rural residents and distant/urban residents. A choice modelling analysis that involved a conditional logit model and a random parameter model was used to elicit household willingness to pay for improvements in environmental quality in the Hawkesbury-Nepean catchment. The results of the study show that the inclusion of a provision rule had an effect on preferences in the distant/urban communities. However, the impact of a provision rule in the local/rural community sub-samples was negligible. This study suggests that the impact of a provision rule should be analysed in the context of different community characteristics.
    Keywords: Choice modelling, Incentive comparability, Provision rule, Non-market valuation, Environment, Environmental Economics and Policy,
    Date: 2010
  4. By: McNair, Ben; Bennett, Jeff; Hensher, David A.
    Abstract: According to neoclassical economic theory, the only stated preference elicitation format that can feasibly be employed in field studies to which truthful response can be the dominant strategy for all respondents is a single binary choice between the status quo and one alternative. In studies where the objective is estimation of preferences for multiple attributes of a good, it is preferred (and, in some cases, necessary) based on econometric considerations, to present respondents with a sequence of choice tasks. Economic theory predicts that utility-maximising respondents may find it optimal to misrepresent their preferences in this elicitation format. In this paper, the effect on stated preferences of expanding the number of choice tasks per respondent from one to four is tested using a split sample treatment in an attribute-based survey relating to the undergrounding of overhead electricity and telecommunications wires in the Australian Capital Territory. We find evidence to suggest that presenting multiple choice tasks per respondent decreases estimates of total willingness to pay and that this effect is related to the ordering of cost levels presented over the sequence of choice tasks. Two behavioural explanations can be advanced - a weak cost minimisation strategy, which implies divergence between stated and true preferences, and a âgood deal / bad dealâ heuristic, in which stated preferences reflect true preferences that change over the course of the sequence of choice tasks. Preferences stated in the first of a sequence of choice tasks are not significantly different from those stated in the incentive compatible single binary choice task. A key objective of future research will be to establish whether this effect becomes less prevalent as the number of attributes and alternatives per choice task are increased.
    Keywords: Choice experiments, willingness to pay, incentive compatibility, strategic behaviour, order effects, underground electricity, Research Methods/ Statistical Methods,
    Date: 2010
  5. By: Kragt, M.E; Bennett, J.W
    Abstract: Choice Experiments (CE) are increasingly used to estimate the values of environmental goods and services. CE questionnaires represent the environmental good under valuation by varying levels of non-market attributes. Inclusion of a cost attribute enables the estimation of monetary values for changes in the non-market attributes presented. The ways in which the levels of the attributes are described in the survey - the âattribute frameâ - may affect respondentsâ choices. Furthermore, varying levels of the cost attribute may impact CE value estimates. The challenge for CE practitioners is to identify the âappropriateâ attribute frames and cost levels. In this paper, the impacts of changing cost levels and the impacts of describing non-market attributes as absolute levels or in relative terms are assessed. These tests were performed using data from a CE on catchment management in Tasmania, Australia. Contrary to a priori expectations, including explicit information cues about relative attribute levels in the choice sets is found not to affect stated preferences. However, comparisons between different split samples provide evidence that respondentsâ preferences are impacted by changing the range in cost attribute levels, with higher levels leading to significantly higher estimates of WTP for one of the three environmental attributes.
    Keywords: Choice Experiments, Environmental Valuation, Bias, Tasmania, Environmental Economics and Policy,
    Date: 2010
  6. By: Windle, Jill; Rolfe, John
    Abstract: Comparing responses from web and paper-based collection modes in a choice modelling experiment
    Keywords: web-based surveys, internet surveys, paper-based surveys, stated preference, collection mode, choice experiments, Environmental Economics and Policy,
    Date: 2010
  7. By: Rolfe, John; Windle, Jill
    Abstract: Valuing protection of the Great Barrier Reef with choice modelling by management policy options
    Keywords: Resource /Energy Economics and Policy,
    Date: 2010
  8. By: Bastone, Chris; Stewart-Carbines, Megan; Kerr, Geoff; Sharp, Basil; Meister, Anton
    Abstract: This paper describes the design and implementation of a choice experiment to understand Aucklandersâ preferences for environmental qualities associated with the effects of urban run-off on marine coastal environments. Aucklandâs coastal environments are affected by a range of ecological and human factors. While much research has been undertaken in the area of ecology, little is understood of human preferences for coastal environments and their management. An unlabelled choice experiment was developed with three environmental quality attributes specified at three broad coastal categories. The environmental qualities are ecological health, water clarity, and underfoot conditions. Willingness to pay estimates for these attributes indicates that respondents show a strong preference for improved environmental quality at outer coastal beach locations over middle and upper harbour locations. Water quality leads ecological health, then underfoot conditions in importance at beach locations. An application is discussed in which a hypothetical project consisting of policy and engineering components delivers changes in water quality and underfoot conditions in the Auckland upper harbour areas. A 95% confidence estimate of the money value of that change ranges from $ 783 m. to $ 1,122 b. The key outcome is demonstration of the choice experiment as a statistically robust and flexible approach to making sense of Aucklandersâ complex preferences for coastal ecosystem management.
    Keywords: Environmental Economics and Policy, Research Methods/ Statistical Methods,
    Date: 2010
  9. By: Sanderson, Todd; Ahmadi-Esfahani, Fredoun Z.
    Abstract: Australia has long been a major exporter of wheat, a commodity well suited to the economic and climatic conditions of Australia. According to the conventional wisdom, Australia holds a comparative advantage in wheat. However, the future validity of this proposition is sensitive to the proposed impacts of climate change. This paper develops a framework with which to examine the future patterns of comparative advantage in wheat given the projections of several global climate models. We find support for the conventional wisdom, and identify the presence of substantial resilience in Australiaâs comparative advantage to adverse yield change.
    Keywords: International Relations/Trade,
    Date: 2010
  10. By: Scheufele, Gabriela; Bennett, Jeff
    Abstract: An elicitation format prevalently applied in DCE is to offer each respondent a sequence of choice tasks containing more than two choice options. However, empirical evidence indicates that repeated choice tasks influence choice behavior through institutional learning, fatigue, value learning, and strategic response. The study reported in this paper employs a split sample approach based on field surveys using a single binary elicitation format with a majority vote implementation as the baseline to expand the research on effects of sequential binary DCE formats. We provide evidence for effects caused by institutional learning and either strategic behavior or value learning after respondents answered repeated choice questions. However, we did not find any indications for strategic behavior caused by awareness of having multiple choices. The choice between a sequential and a single elicitation format may thus imply a trade-off between decreased choice accuracy and potentially increased strategic behavior due to an incentive incompatible mechanism. Further research is needed to explore strategic behavior induced by incentive incompatible elicitation formats using alternative approaches that are not compromised by a confounded baseline, that facilitate the differentiation between value learning and strategic behavior, and that allow the use of less restrictive model specifications. Such research should also investigate the effects of varying incentives induced by the order in which choice questions are presented to respondents.
    Keywords: discrete choice experiments, split sample approach, elicitation format, incentive compatibility, strategic behavior, learning effects, panel mixed logit models, Environmental Economics and Policy,
    Date: 2010

This nep-dcm issue is ©2010 by Philip Yu. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.