INFERENCES REGARDING THE BACS OF NON-RESPONDENTS IN THE CANADIAN ROADSIDE SURVEY

This paper addresses the problem of non-respondent bias in roadside surveys and develops a mathematical paradigm relating the absolute and conditional blood alcohol content (BAC) distributions (with response as an argument of conditionality). The implications of this model are evaluated on the basis of various assumptions regarding the probability of response at defined BAC levels, using data from the Canadian National Roadside Survey (CNRS). Accordingly, the objectives of the investigation are twofold: (1) To examine the theoretical relationship between the (obviously known) conditional BAC distribution of BAC-tested drivers (respondents), and the absolute (unknown) BAC distribution of drivers in the population at large (respondents plus non-respondents); (2) To evaluate that theoretical framework on the basis of various assumptions regarding the probability of response at defined BAC levels, using concrete summary data from the CNRS.

Media Info

  • Features: Figures; References; Tables;
  • Pagination: p. 68-91

Subject/Index Terms

Filing Info

  • Accession Number: 00396583
  • Record Type: Publication
  • Source Agency: National Highway Traffic Safety Administration
  • Report/Paper Numbers: HS-037 586
  • Files: HSL, USDOT
  • Created Date: Aug 31 1985 12:00AM