CROSS-VALIDATION OF LONDON HOUSEHOLD SURVEY TRAVEL DATA WITH DATA FROM IN-MODE SURVEYS: THE LATS EXPERIENCE

This paper describes work, performed by the MVA Consultancy for the Department of Transport (DTp) and the London Research Centre, during the Database project for the London Area Transport Survey (LATS). It reports the results of comparisons between the household survey and in-mode surveys, including roadside interviews and rail surveys. These surveys were conducted as part of the process of developing LATS' proposed integrated comprehensive database of travel in London and of the related socio-demographic characteristics of residents and, as far as possible, visitors. In particular, definitive estimates were required for the totality of trips on a typical weekday in the London area. An essential feature of the surveys' designs was that results could be cross-validated, so that the results from different core surveys could be compared for identifiable categories of trip sampled in more than one of them. Several obstacles had to be overcome, to achieve cross-validation. The following projects are discussed in detail, with tables summarising their results: (1) comparisons between the household survey and the roadside survey; (2) cross-validation of rail trips; and (3) cross-validation of bus trips. For the covering abstract see IRRD 869495.