Économie et Statistique n° 437 - 2010 Employees over 50 and Information Technology - Impact of Thresholds of 10, 20, and 50 Employees on the Size of French Firms - Impact of Survey Method on Measurement of Mobility Behaviour
Impact of Survey Method on Measurement of Mobility Behaviour
Mobility surveys are essential both for understanding, analyzing, and modelling travel behaviour and for helping to define and assess transport policies implemented. Yet response rates to the “Household Travel Surveys” (Enquêtes Ménages Déplacements: EMDs) have been declining steadily. In recent years, data-collection methods have evolved in order to provide inputs for ever more complex models and incorporate new technologies (such as the Web and GPS) into survey protocols. In particular, the combination of different media has proved its worth as a cost-effective means of improving the quality of the data produced by raising the total response rate. However, while the fact of offering a choice of media has boosted the total response rate, data comparisons remain a difficult exercise. Respondents' socioeconomic characteristics vary with the survey method, and these differences may account for specific behaviours. In our article, we show that it is possible to distinguish between the effect of the survey method and the effect of socioeconomic differences observed between respondent samples. We can also quantify the survey method's impact on the measurement of reported mobility. We have borrowed the econometric model for this analysis from the field of qualitative variables. More specifically, we use a sample-selection model and estimate its parameters by means of the two-stage procedure developed by Heckman in the late 1970s. We illustrate our analysis with the results of the 2006 EMD survey conducted in Lyon face-to-face and online. Our goal is to quantify the impact of the survey method on the measurement of mobility behaviour for each respondent.