Comparison of Web and Face-to-face Household Travel Survey- Application to Lyon Case

Comparison of Web and Face-to-face Household Travel Survey- Application to Lyon Case


C Bayart, P Bonnel, Laboratoire d'Economie des Transports, FR


The paper presents the results of the data comparison between a face to face household travel survey and a web survey, the latter being proposed only to those who refused to respond or are not joinable for the face to face survey


Household travel surveys response rates are decreasing. Efforts are made to increase response rate for traditional survey by improving the questionnaire, reducing respondent burden, increasing reminders? Even if results are generally positive, it is in most cases not sufficient. Weighting aims at reducing the impact of non response, but it is always necessary to postulate that people with some socio-demographic characteristics who do not respond to a survey have the same behaviour than people with the same socio-demographic characteristics who respond. But evidence seems to indicate that it is not always the case for travel (Richardson, Ampt, 1993; Richardson, 2000, Ampt 1997, Murakami, 2004). To reduce this bias of non-response, we initiated a project of a web survey in parallel of the household travel survey conducted in face to face in Lyon. The idea is to propose to households who refuse to respond in face to face or are not reachable after a certain number of attempts to respond by the web. This new and interactive mode of data collection offers to the respondents the possibility to choose a more appropriate moment to complete the questionnaire, and does not require to set an appointment with the interviewer. However, Internet penetration rate is still low, and users capabilities and equipment vary a lot. If Web surveys allow to reduce the non-response rate, the generalization of the results to the whole population remains an issue (Myles & Tibert, 1998). Moreover, the implementation of a Web survey raises specific problems, in terms of design and administration of the questionnaire. Last but not least, if the launch of a web survey makes it possible to study behaviours little represented up to now (hyper-mobiles households, with shifted schedules...), the question of data comparability remains entire (Stopher & Jones, 2003). ). Although no similar experiment (web/face-to-face travel survey) was reported in the literature, we will return on mixed surveys examples (postal/Web) to try to understand the differences noticed.

Around 370 full web questionnaires have been stored in our database. That represents a global response rate of 8.5%, remaining that some households are not able to connect to Internet at home or at their working place and that the survey was proposed only to a portion of households who have refused or who were not reachable after 8 attempts. In our context web combined with face to face seems therefore more promising.
As the methodology used allows us to study net surfers behaviour, when they fill out the questionnaire, we present some details on response time, number of connections, abandonment causes and non-response items.
But the main objective of the data analysis is to assess the potential of Web for households travel surveys. First, we analyse socio-demographics of web respondents compared to face to face respondents and to the whole population of the surveyed area. The second step is dedicated to travel behaviour analysis. We compare web data with face to face data, to show how the media could influence response behavior. Results indicate that short trips are often omitted in web survey. They could be explained in part by the absence of interviewer (Bonnel, 2003; Bonnel, Le Nir, 1998). The comparison is therefore not limited to the number of trips, but we also consider the number of tours (succession of trips and activities between the exit of home and the next way back) in order to assess if differences are mainly due to short trip omissions or if some tours are also omitted. The comparison follow with trip/tour characteristics like transport mode, purpose, distance, duration, time of day. To take into account socio-demographics difference between populations which might generate different behaviours, we focus then on the working population, which constitutes 70% of the web sample.

This paper initially discusses web potential for households travel surveys, especially in a mixed modes framework (section 1). Then, some thoughts on Lyon on-line questionnaire and the choices operated compared to its paper version are provided (section 2). Finally, we will present the results of the Lyon web travel survey compared to the face-to-face survey, and give some perspectives for future households travel surveys (section 3).


Association for European Transport