An investigation of taste and reference dependence heterogeneity
Access status:
Open Access
Type
Working PaperAbstract
This paper considers reference dependence and loss/gain asymmetry in a stated choice (SC) study in which the attribute levels pivot around those of a recent trip. A latent class (LC) model is presented that allows for heterogeneity both in the tastes, and the nature and magnitude ...
See moreThis paper considers reference dependence and loss/gain asymmetry in a stated choice (SC) study in which the attribute levels pivot around those of a recent trip. A latent class (LC) model is presented that allows for heterogeneity both in the tastes, and the nature and magnitude of any reference dependence. In addition to the expected taste heterogeneity, differences in reference dependence across the sample are observed. For both attributes for which reference dependence is possible in this study, running cost and trip time, a lack of reference dependence is most common, followed by loss aversion, then, for a small but significant minority, gain seeking. Two classes have no reference dependence to either attribute, and are differentiated by the magnitude of the values of time. Two classes present mixed evidence about the reference dependence across the attributes. One class demonstrates loss aversion to both attributes, leading to a clear willingness to pay (WTP)/willingness to accept (WTA) asymmetry, where the mean value represents a near average value of time. Most intriguing is a class, representing about 9 percent of the sample, which exhibits gain seeking to both attributes, and a high value of time. The gain seeking thus pushes the high WTP higher. Such a finding could have implications when evaluating infrastructure investments with respect to highly time sensitive travellers.
See less
See moreThis paper considers reference dependence and loss/gain asymmetry in a stated choice (SC) study in which the attribute levels pivot around those of a recent trip. A latent class (LC) model is presented that allows for heterogeneity both in the tastes, and the nature and magnitude of any reference dependence. In addition to the expected taste heterogeneity, differences in reference dependence across the sample are observed. For both attributes for which reference dependence is possible in this study, running cost and trip time, a lack of reference dependence is most common, followed by loss aversion, then, for a small but significant minority, gain seeking. Two classes have no reference dependence to either attribute, and are differentiated by the magnitude of the values of time. Two classes present mixed evidence about the reference dependence across the attributes. One class demonstrates loss aversion to both attributes, leading to a clear willingness to pay (WTP)/willingness to accept (WTA) asymmetry, where the mean value represents a near average value of time. Most intriguing is a class, representing about 9 percent of the sample, which exhibits gain seeking to both attributes, and a high value of time. The gain seeking thus pushes the high WTP higher. Such a finding could have implications when evaluating infrastructure investments with respect to highly time sensitive travellers.
See less
Date
2014-07-01Department, Discipline or Centre
ITLSShare