Abstract
This article models negative impact on the environment as one of the attributes of transport mode. By this modelling, we are able to examine whether individual environmental consciousness of this impact plays a significant role in his/her choice of transport mode. A survey data from Saito and Onohara Area in Northern Osaka of Japan is used to estimate the model with the Heteroscedastic Extreme Value specification. Both of the estimated and simulated results imply that individual environmental consciousness does influence his/her decision on transport mode choice in the sample. Furthermore, the likelihood ratio tests indicate that both the utility and scale parameters are not equivalent across sub-samples of university commuters, research-facility commuters and residents. The results of the comparison across sub-samples suggest that sometimes we may learn more from sub-dividing a whole sample into several sub-samples if we could distinguish them by their characteristics.
Acknowledgements
The authors would like to thank one anonymous referee for his/her useful suggestions, which resulted in an improved revision of the article. The authors would also like to appreciate the financial support from Japanese Ministry of Education, Culture, Sports, Science and Technology (Grand-in-Aid for Scientific Research A(2) 16203020). All the views expressed in this article and any errors are the sole responsibility of the authors.
Notes
1 For the issue on additive structure of the utility function, see a recent study by Selvanathan and Selvanathan (Citation2005).
2 Posting is a similar method to mail survey, which can be used in some simple and familiar products’ surveys. It is executed by delivering questionnaires into the surveyed respondents’ posts and asking respondents to mail the completed questionnaires back.
3 The scale parameter of bus is set to one for identification.
4 The results for listed pooled sample in Tables and are for university commuters + research-facility commuters + residents. To conserve space, we do not list the results for other pooled samples. Those results are available upon request.
5 We have also tried interacting days (average number of days in a week travelling in the surveyed section) with ASC of monorail. However, this interaction term is not significant in any sub-sample. Therefore, we drop it in our final estimation.