Effects of Response Formats of Stated Preference Analysis for Travel Mode Switching Behaviour - Bivariate Probit and Interval Data Models for One-and-one-half Bound Format
N Sanko, Kobe University, JP; T Morikawa, Nagoya University, JP
This preliminary study applies SP data from a family of double-bounded formats to transport research. The study calculates and examines several models, explores suitable models for the formats, and investigates biases related to the formats.
Widely used for analysing travel behaviour, SP (Stated Preference) data analysis examines respondent preferences under hypothetical conditions. However, no consensus has been achieved concerning the appropriate approaches for SP data analysis and researchers continue to use the process of trial and error. Two of the issues examined in this study are closely related to each other: the response formats of the SP experiment design and the related modelling framework.
Regarding the first issue, although the choice format is currently the most frequently used for transport research, attempts have been made to apply the iterative choice format that is commonly used in CVM (Contingent Valuation Method). One of the iterative choice formats most frequently used, although not for transport analysis, applies a family of double-bounded (DB) formats.
Regarding the second issue, the modelling framework must suit the design of the experiment (including the response format), since unsuitable mathematical models produce incorrect estimations. Suitable models of a family of DB formats for transport behaviour analysis are rarely explored.
In this preliminary study, a family of DB formats, particularly the one-and-one-half bound (1.5B) format, is applied to an analysis of commuter travel mode switching behaviour in the Kyoto-Osaka-Kobe metropolitan area in Japan. For mode switching behaviour between mass transit and auto, the 1.5B format is applied in the following way: In the first bound, mass transit users are asked, would he or she change transport mode if the level of service for mass transit became worse or if that of the auto became better? Only if the respondent does not change modes in the first bound is the second bound applied: Would the respondent change transport mode if the level of service for mass transit became much worse or if that of the auto became much better? Not only does the cost level change in the survey, but other level-of-service variables also change, which is very different from CVM. Of course, auto users also can be respondents.
The aims of this study are as follows.
1) Regarding the first issue, the study will compare models using the first-bound response and models using both the first- and second-bound responses. The authors will determine whether the second-bound question provides useful information.
2) Regarding the second issue, the study will develop and compare several models using both the first- and second-bound responses. Models include the interval data model, where the first- and second-bound responses are assumed to come from an identical utility function, and the bivariate probit model, where the first- and second-bound responses can come from different functions. The authors will offer suggestions regarding model selection for the 1.5B format.
3) Regarding both issues, the study will investigate the reliability of the 1.5B data. Although some studies insist that the DB formatfs greater statistical efficiency is advantageous, other studies argue that the biases caused by the second-bound question make the DB format disadvantageous. The authors examine the RP (Revealed Preference) models and statistically test and compare the parameter equality of the RP and SP models (SP models using first-bound response only and both first- and second-bound responses). The authors will determine whether the SP models are as reliable as the RP models. The authors will also analyse the bias of the 1.5B format in the transport survey settings.
Association for European Transport