How the Weather Can Influence Your Data Collection?

How the Weather Can Influence Your Data Collection?


Christian, CEREMA, Hurez, CEREMA, Rabaud, CEREMA


The aim of this paper is to explain how the weather may affect the level of mobility when the day of interview is not randomly selected in travel surveys.


In France, for local & regional household travel surveys, The Centre For Studies and Expertise on Risks, Environment, Mobility, and Urban and Country planning (CEREMA) proposes several methods of surveys adapted to the size of the towns and the various types of region. The purpose of household travel surveys (HTS) is to collect data on the travel habits of an urban population. They are different from surveys which look at only one means of transport, such as surveys on public transport or surveys carried out on motorists by the side of the road. They count all journeys made by the people questioned, whatever the mode of transport used, including walking, regarded as a mode in its own right, without consideration of distance nor of minimum duration. Understanding these global travel practices provides essential techniques for drawing up and assessing transport policies in towns. But, data quality is a real challenge, mainly because response rates are declining and interviewees are more and more reluctant to respond to burdensome questionnaires. Until know, to keep a high response rate, in France, for each selected household, we do not choose at random the day of description of the mobility, but let the possibility to the interviewees to set up the rendez-vous and this may cause some measurement errors.
The Rhône-Alpes regional council has decided to develop an operational mobility observation and prediction method at a regional level in order to facilitate the choice of transport policy, notably for regional rail and road passenger transport services (TRV). The Rhône-Alpes Regional Travel Survey (referred to as EDR-RA) is one of the results of this decision. As the survey aims to observe the mobility of alternatives modes to the car (such as the train). In order to provide a reliable basis for the study, it must include at least 1,200 trips by train, which implies interviewing between 30,000 and 40,000 people (3-4 trips per person and modal share of 1%). Thus, the total survey sample was expected to include about 37,000 people in 77 sectors. We therefore interviewed 480 people in each sector, in order to have a sufficient statistical threshold at this scale. We choose the option to interview one third of the people (one third of the target in each of the 77 sectors). This choice provides a vision of regional mobility right from the first year of survey, and a comparison between the different survey years. Data collection will take place by phone over three consecutive winters (2012-2013 / 2013-2014 / 2014-2015), from October to March, outside school holiday periods.
When analyzing each of the three waves separately, we found statistically very different levels of trips frequency. The aim of this paper is to explain these differences, is it due to the effect of interviewers (many of them are the same from one year to the next) or due to external factor such as weather and so on.


Association for European Transport