Identifying the Influence of Stated Choice Design Dimensionality on Willingness to Pay for Travel Time Savings
Access status:
Open Access
Type
Working PaperAuthor/s
Hensher, David A.Abstract
This paper explores the influence of the dimensions of stated choice (SC) designs on the value of travel time savings. Utilising principles of experimental design, 16 choice designs are embedded within a global design in which we vary the number of choice sets, the number of ...
See moreThis paper explores the influence of the dimensions of stated choice (SC) designs on the value of travel time savings. Utilising principles of experimental design, 16 choice designs are embedded within a global design in which we vary the number of choice sets, the number of alternatives in each choice set, the number of attributes per alternative, the number of levels of each attribute and the range of attribute levels. A mixed logit model is estimated in which design dimensions are interacted with the attribute parameters to explore the influence of these dimensions on willingness to pay (WTP) for travel time savings. The evidence in the context of a sample of respondents in Sydney choosing amongst trip attribute bundles for their car commuting trip suggests that design dimensionality does influence variations in WTP; with higher overall mean values of travel time savings associated with more complex designs (in terms of the number of items to process). Given evidence external to the SC design that most individuals tend to add up the components of travel time in processing bundles of attributes, we interacted aggregate travel time with a choice complexity index as a more general test of the directional impact on VTTS due to the net effect of all positive and negative design influences. The prior finding was confirmed. Importantly, the evidence should not be taken to suggest that simpler designs (i.e., less items) are preferred, but rather that individuals process the information in SC designs in ways that appear to support at least two hypotheses, one related to coping strategies and the other related to behavioural relevance. We recommend, however, that the selection of the design specifications (e.g., relevant attributes and their range) should be guided by, at least, the specific application setting.
See less
See moreThis paper explores the influence of the dimensions of stated choice (SC) designs on the value of travel time savings. Utilising principles of experimental design, 16 choice designs are embedded within a global design in which we vary the number of choice sets, the number of alternatives in each choice set, the number of attributes per alternative, the number of levels of each attribute and the range of attribute levels. A mixed logit model is estimated in which design dimensions are interacted with the attribute parameters to explore the influence of these dimensions on willingness to pay (WTP) for travel time savings. The evidence in the context of a sample of respondents in Sydney choosing amongst trip attribute bundles for their car commuting trip suggests that design dimensionality does influence variations in WTP; with higher overall mean values of travel time savings associated with more complex designs (in terms of the number of items to process). Given evidence external to the SC design that most individuals tend to add up the components of travel time in processing bundles of attributes, we interacted aggregate travel time with a choice complexity index as a more general test of the directional impact on VTTS due to the net effect of all positive and negative design influences. The prior finding was confirmed. Importantly, the evidence should not be taken to suggest that simpler designs (i.e., less items) are preferred, but rather that individuals process the information in SC designs in ways that appear to support at least two hypotheses, one related to coping strategies and the other related to behavioural relevance. We recommend, however, that the selection of the design specifications (e.g., relevant attributes and their range) should be guided by, at least, the specific application setting.
See less
Date
2004-04-01Department, Discipline or Centre
ITLSShare