February 03, 2015, 13:45 h
Mannheim, Conference room B2,8
Dipl.-Soz. Johannes Bauer
Random route samples are widely used in German, European and international surveys. In a preceding study random route instructions were tested for their theoretical property to select respondents with equal probability. This survey reproduces these findings for the Noelle, TNS Infratest and the Gesellschaft für Konsumforschung (GFK) routes and extends the analysis to deviations in variables. Registration office data are applied to verify the negative impact of biased household selections on survey results. All three routes lead to biased selection probabilities and biased expected values in multiple variables. The strongest errors were found in variables that were related to the spatial location of a household. Furthermore, some errors were not only related to random route instructions, but also to the selection of starting points for these routes. Data analysts can use these results to gain an impression about the degree of bias, while researchers that apply random route samples should view these findings as inducement to develop alternatives or improved random route samples. The presentation closes with an alternative sampling procedure that can serve as substitute for problematic random route instructions.
About the speaker
Johannes Bauer is a research assistant at the University of Munich (LMU). He teaches statistical analysis and survey methods. His doctoral thesis focuses on problems in random route samples and is partly published in Sociological Methods & Research. His research interest focuses on sampling methodology, statistical modeling with incomplete data, partial identification and simulations models.