Biases in response over time in a seven-day travel diary

Abstract

Data from multi-day travel or activity diaries might be biased if recording inaccuracies and tendencies for respondents to skip certain types of trips or activities increases (or decreases) from day-to-day over the diary period. One objective of the research reported here is to test for such temporal biases in a seven-day travel diary. A second objective is to calculate correction factors which can be applied to the data in the case that biases are found. The analyses were conducted using regression and analysis-of-variance techniques. The variables investigated included total trips per day, total travel time per day, and trips per day by various modes (such as walking, car driver and car passenger). Results showed that most biases per capita statistics are due to increases over time in the percentage of respondents reporting no travel at all for an entire day. However, even after accounting for this bias by measuring statistics in terms of "per mobile person", there remains a decrease over time of about 3.5 percent per day in the reporting of walking trips. This appears to be the main factor in the overall bias of about one percent per day in total trips per mobile person per day. No significant differences were found among population segments in terms of the levels of their biases. Objectives and ScopeTravel diaries, in which respondents record salient facts about the trips they make over a fixed period of time, provide the fundamental data for most studies of travel behaviour. Usually, the time period for such diaries is one or two days. It is quite possible that this emphasis on only one or two days of travel and related activities has severely restricted the subject matter and the methodologies of travel-behaviour studies. Indeed, several authors have pointed out the limitations imposed by diaries of one or two days (e.g., Scheuch, 1972;Goodwin, 1979).Extending the recording periods for travel and activity diaries, perhaps to seven days or more, permits the study of day-to-day variations in travel behaviour. Focus can be placed more effectively on activity patterns which typically occur in cycles of multi-day duration (e.g., weekly). Also, time and money expenditure patterns (so-called budgets) can be more effectively analysed, because one or two days is in general too short a time for observing overall expenditures. Examples of studies addressing these topics using extended-period data are provided by Bentley et al.

Panel analysis in transportation planning: An overview

“…As in Golob and Meurs (1986), they found trip reporting decline gradually within each diary period.t Using the average number of trips reported on the first diary day as a reference, Meurs et al, estimate that trip under-reporting amounted to an average of 2.27 trips per week per person in the first survey that a respondent participated. This increased to 4.44 trips in a second survey, and up to 8.35 trips by the time a respondent stayed in the panel to participate in a seventh survey .…”

“…In the case of attrition, a probabilistic model of attrition, constructed using measurement from the earlier survey wave(s), can be used to formulate more efficient weight that takes advantage of mobility and other available information (Kitamura and Bovy, 1987). For discussions on panel conditioning and reporting accuracy, see Golob (1990), Golob and Meurs (1986), Meurs et al (1989), andvan de Pol (1987).…”

A dynamic model system of household car ownership, trip generation, and modal split: model development and simulation experiment

“…Additional discussions of the aim, format and characteristics of this panel data set can be found in Golob et al (1986a), Golob and Meurs (1986), Kitamura and Bovy (1987), Meurs et al (1987a), and Meurs et al (1987b). A sample of analyses using the Dutch Panel data set includes Golob (1986Golob ( , 1987, Meurs (1987, 1988), Golob et al (1986b), Goodwin (1987), Kitamura (1987aKitamura ( -c, 1988), Kitamura and van der Hoorn (1987), Recker et al (1987), and van der Hoorn and .…”

Impact of telecommuting on spatial and temporal patterns of household travel

“…Also, trip attributes such as origin, destination, and duration, were imputed wherever logically possible using information available in the reported trips. This effort was undertaken to account as much as possible for the potential effects of trip reporting errors, which are common in multi-day panel travel diary surveys of this type [Golob and Meurs, 1986;Meurs, et al, 1989;and Pas, 1986].…”