Bias in einem komplexen Surveydesign

Abstract

"This report contains an analysis of nonresponse bias in the IAB survey ALWA ('Working and Learning in a Changing World'). Besides the usual processes of attrition in standardized surveys, which result from persons who cannot be reached and persons who do not cooperate, two additional sources of nonresponse were relevant here: first, persons in the register based address sample for whom no telephone number could be located, and second, additional dropouts due to the complex two-step design of the study. Our results show that the ALWA telephone survey is biased in similar ways as other population surveys. For the largest part, this selectivity can be traced back to non-location because of missing telephone numbers. Positive experiences were made with collecting reading and mathematical competence data in a face-to-face interview separately from the rest of the survey. Even though only 40 percent of the CATI respondents participated in these tests, the two-step design caused only a slight amount of additional selectivity." (Author's abstract, IAB-Doku) ((en))

Cite article

Kleinert, C., Ruland, M. & Trahms, A. (2013): Bias in einem komplexen Surveydesign. Ausfallprozesse und Selektivität in der IAB-Befragung ALWA. (FDZ-Methodenreport 02/2013 (de)), Nürnberg, 30 p.