Abstract
The volume of information collected on long distance travel in the N.T.S. conducted in 1993-94 has been maximised. It implies oversampling high mobility groups (multi-car households and the most mobile person in the household) and enlarging the survey period (a retrospective interview + a self-administrated questionnaire, both for 3 months). Consequently, complex weighting procedures are needed because of memory effects and of the low response rate (60%) for the self-administered questionnaire. Comparisons are made between different survey instruments and traffic counts. Confidence intervals are calculated. (A)