NONRESPONSE ISSUES IN HOUSEHOLD TRAVEL SURVEYS

Sample surveys in transport are invariably associated with some level of nonresponse. The issue of nonresponse is fundamentally connected with the questions of reducing survey bias and increasing the accuracy of sample estimates. This is because nonrespondents are from segments of the population having characteristics significantly different from those of respondents. The differences are in sociodemographic characteristics and, more important, in travel behavior characteristics. Some of the background concepts of transport surveys within which nonresponse should be considered are described. In particular, the distinction between sampling error and survey bias and the trade-offs between quantity, quality, and cost of survey data are described. Ways of reducing nonresponse and the impacts of nonresponse are discussed. The need to consider all sources of nonresponse in different types of survey and the importance of using consistent methods of calculating nonresponse are stressed. The use of population expansion factors, nonreporting weights, and nonresponse weights as ways of allowing for the remaining extent of nonresponse is described. The use of these methods is demonstrated by reference to a mailback questionnaire survey.

Language

  • English

Media Info

  • Features: Figures; References; Tables;
  • Pagination: p. 79-114
  • Monograph Title: CONFERENCE ON HOUSEHOLD TRAVEL SURVEYS: NEW CONCEPTS AND RESEARCH NEEDS, IRVINE, CALIFORNIA, MARCH 12-15, 1995
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 00722471
  • Record Type: Publication
  • ISBN: 0309062101
  • Files: TRIS, TRB, ATRI
  • Created Date: Jun 4 1996 12:00AM