A new paper from the Marie Skłodowska-Curie project Econ-ML has been published in Transportation journal. The paper is titled: “Analyzing the reporting error of public transport trips in the Danish national travel survey using smart card data” and co-authored with Filipe Rodrigues, Maya Abou-Zeid, and Francisco Camara Pereira. This blog post provides an overview of the study, including its background, key findings, and overall impact. You can click here for the full paper.
Background and Motivation
Travel surveys have long been an essential tool for understanding travel behavior of individuals and households. However, these surveys often suffer from recall bias, where participants might struggle to accurately report their travel details, such as the time or number of trips. With the rise of digital systems like smart cards, which automatically record travel data, we now have an opportunity to examine and quantify these reporting errors.
The Study: Using Smart Card Data to Assess Reporting Accuracy in Travel Surveys
The paper investigates the accuracy of self-reported public transport trips in the Danish National Travel Survey (TU) by comparing this data with the Danish Smart Card system (Rejsekort). The study focuses on identifying and quantifying time reporting errors by matching data from the TU survey with Rejsekort data over a five-year period, from 2018 to 2022. The time reporting error corresponds to the absolute time difference between the reported time in the TU data and the tap-in time in the Rejsekort data at the first stop of a trip.
Key Findings
In our analysis, we successfully matched approximately 70% of TU respondents with corresponding Rejsekort travel cards by aligning the reported travel behaviors in the survey with actual recorded trips in the Rejsekort system. The study categorizes trips into three types: train, bus, and mixed modes, and analyzes the reporting accuracy within these categories.
The findings reveal a median time reporting error of 11.34 minutes, with an interquartile range of 28.14 minutes. The study identifies several factors that influence reporting accuracy. For instance, it was found that females generally report their travel times more accurately than males. Individuals with fixed schedules, such as students and employees, tend to be more accurate in reporting their trips than those with flexible schedules. Additionally, trips reported on weekdays are more accurately recalled compared to those reported on weekends and holidays. The mode of survey also plays a role, with internet-based surveys leading to higher accuracy than telephone-based surveys. Geographically, trips in Jutland were reported with greater accuracy compared to those in Zealand and Funen. A more detailed statistical analysis can be found in the paper.
Impact
Understanding and quantifying reporting errors in travel surveys can greatly benefit policymakers and researchers in two key areas: improving data collection and correcting data post-collection. By identifying groups that are more prone to reporting inaccuracies, data collectors can enhance their survey methods with additional checks or probing questions. Furthermore, using smart card data to correct self-reported survey data allows for a more accurate and reliable dataset without discarding valuable survey information.
Such approach could significantly improve the accuracy of activity-based models, especially in predicting travel times and peak periods. Inaccurate reporting of travel times can lead to errors in estimating travel demand, but by correcting these discrepancies, we can make better-informed decisions for transport planning and policy.
Leave a Reply