Tackling non response in household travel surveys: a case study

Abstract

Ever since car ownership and car use started to increase in Western Europe and the US, transportation planners attempted to model people's travel behaviour. A series of models was developed, each successor ameliorating the previous one (Jovicic, 2001), however, the question why people undertake trips was completely neglected. This is where the activity-based models came into play, which have set the standard for transportation and travel behaviour modelling during the last decade (McNally, 2000). The main idea behind these activity-based models is that travel demand is derived from the activities that individuals and households need or wish to perform (Jones, et al., 1983). In order to receive the necessary information to build these models, data need to be collected. Methodological research in the time use and transportation community is accumulating, o®ering a good basis to decide on the most appropriate format for data collecting. A questionnaire, asking people for their typical travel behaviour on an average day has long been the dominant form of data collection in transportation research. It has been argued, however, that such a questionnaire may result in under-reporting of trips, especially of o®-peak, non home-based trips of short duration (Koppelman, 1981; Dijst, 1993). Stimulated by these ¯ndings, Stopher (1992) and Clarke and co-authors (1981) reported that an activity diary outerforms a travel survey with regard to under-reporting of trips, even by 13-16%. Similar ¯gures have been reported in time use literature as well (Nieme, 1993). Nevertheless, diaries are not perfect either. Collecting diary data is quite demanding for the respondents and therefore, although this is not unique for diary data, there may be evidence of lower response rates and of di®erential non response by certain socio-demographic groups. For example, Sen et al. (1995) reported that larger households had lower response rates when compared to single member households. Regardless of the potential source of bias introduced in this way, an activity-based model requires detailed information, so a diary seems a good instrument to collect it. In this paper, a possible solution will be presented on how non response in household travel surveys can be tackled. Section 2 outlines the data collection process and tries to provide a solution on how to go about non response, while Section 3 provides the results and gives pro¯les of the responders of the study. The paper concludes with a discussion of the major ¯ndings and possible avenues for future research.