Development of Stated Preference Interviewing Using Micro Computers

Research Project

Project/Area Number	04555130
Research Category	Grant-in-Aid for Developmental Scientific Research (B)
Allocation Type	Single-year Grants
Research Field	交通工学・国土計画
Research Institution	HIROSHIMA UNIVERSITY
Principal Investigator	SUGIE Yoriyasu Hiroshima Univ, Fac of Eng, Professor, 工学部, 教授 (70034410)
Co-Investigator(Kenkyū-buntansha)	FUJIWARA Akimasa Hiroshima Univ, Fac of Eng, Research Assistant, 工学部, 助手 (50181409) IMADA Hirofumi Hiroshima Univ, Fac of Eng, Research Assistant, 工学部, 助手 (80093730)
Project Period (FY)	1992 – 1993
Project Status	Completed (Fiscal Year 1993)
Budget Amount *help	¥3,200,000 (Direct Cost: ¥3,200,000) Fiscal Year 1993: ¥700,000 (Direct Cost: ¥700,000) Fiscal Year 1992: ¥2,500,000 (Direct Cost: ¥2,500,000)
Keywords	Stated preference / Computer-based survey / Rank logit model / Multinomial logit models / New Transit System / 交通需要分析 / マイクロコンピュータ / ロジットモデル
Research Abstract	This paper is aimed at identifying the bias included in Stated Preference data which does not coincide with the Revealed Preference data using the mode choice SP data obtained in Hiroshima and also developing a practical computer program for interviewing which minimize the bias as much as possible. The computer program was improved specifically for on-board the bus survey and SP survey was carried out in 1992 for bus passengers who commute from the suburban residential area located along the New Transit System line which is due to be open in 1994. Twenty-eight persons on bus were finally interviewed using portable color computes. Four different SP questions were done for one respondent, so we could obtain 112 samples as a whole. After the survey, oral questions were done in order to evaluate the computer-based SP survey. It was found that more respondents took a great interest in the computer-based interviewing and they prefer the new survey method to the traditional paper-based questionnaire survey. This indicates the computer-based survey will become a useful tool in transport survey, especially on-board surveys. Ranked data of three travel modes including NTS, bus and car were used to develop Rank Logit models for mode choice. It was found that good models were obtained by checking the significance of model parameters and goodness-of-fit indices. Besides, Multinomial Logit models were also developed by defining the best mode as a chosen one. This also presented a good degree of fit. These results indicate that the data obtained by on-board the bus survey were highly reliable and the computer interview software used in this survey was fully operational.