Institut Français des Sciences et Technologies des Transports, de l'Aménagement et des Réseaux

Type de document

RAPPORT DE RECHERCHE

Langue

anglais

Auteur

ARMOOGUM ; MADRE ; HAN ; POLAK ; - ; - ; - ; - ; - ; -

Résumé / Abstract

From a survey, it is not always possible to directly obtain complete data. There is usually something missing: the household might refuse to answer, a part of the questionnaire might not be returned, some trips omitted, or an item is missing or inconsistent (for instance a too high trip distance or a too short trip duration inducing an unreasonable speed). Two main categories of methods are used to cope with this variety of non-responses: weighting and imputation. After a presentation of the main useful concepts in this field, we will consider these two methods, both theoretically and practically, with examples drawn from a methodological survey (vats) and from the French national personal transportation survey (NPTS). This report presents original methodologies developed for the MEST project and is a synthesis of many papers, which are either presented in the appendices or mentioned in the references. RAPPORT DE CONTRAT