Abstract Several recent basic investigations have pointed out that the choice, specific format and content of the survey method employed for an empirical study already constitute an important determinant for the type and scale of answers to be expected. In the context of transportation research this implies, for example, that the type and level of measured mobility also depends on the method by which these mobilities were determined. Such basic investigations provide important indications of the effects to be expected, but generally they are not able to quantify the true level of such influences. For that purpose studies are necessary that actually investigate the same behavioral phenomenon by means of different survey methods. Such a comparison of the trip purpose “vacation travel” was possible for the Federal Republic of Germany on the basis of three different surveys. Surveys (1) and (2) were similar in their design, namely in the “classical” form of the personal interview. By contrast, the third survey employed mail-back questionnaires. In addition, in survey (3) an attempt was made to include factors other than the commonly used sociodemographic weightings (applied in all three surveys) that bias the results. These other factors are especially the non-response effect, the order of magnitude for non-reported trips and other influences that can be attributed to the survey design. Substantial differences between the three surveys became evident when the key measures of vacation travel were subjected to a quantitative analysis. In general, it could be observed that the corresponding values for surveys (1) and (2) were substantially (up to 30%) below those generated by survey (3). However, these deviations could be attributed to the survey method selected, both in terms of their direction and in their order of magnitude. This means that the reported variances of the random sampling errors do not permit statements about the validity of the data, since the systematic bias due to design is at least ten times the size of the random error. It was shown also that the reliability of the survey data is affected by the survey design. The paper concludes with a number of suggestions derived from these insights. They should be taken into account in future surveys of vacation travel in order to improve the accuracy of the reported results.
Influence of survey methods on the results of representative travel surveys
Transportation Research Part A: General ; 17 , 2 ; 149-156
01.10.1982
8 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
How Representative of Bicycling Populations Are Smartphone Application Surveys of Travel Behavior?
Transportation Research Record | 2016
|Representative Projection of Sample Surveys
British Library Online Contents | 2004
Factoring Household Travel Surveys
Online Contents | 1993
|Nonresponse and Travel Surveys
British Library Conference Proceedings | 2003
|