General social surveys are usually conducted face-to-face and have long time series to track public opinion trends. To ensure consistency, these surveys typically keep their design changes minimal over time. However, face-to-face surveys have faced declining response rates and increasing costs in recent decades. Consequently, self-administered mixed-mode surveys (mail and web) are becoming more popular since they can overcome these issues. Nonetheless, switching data collection modes can influence trends due to differences in who participates (selection) and how respondents answer (measurement) in the different modes. This dissertation explores the transition of a general social survey from a face-to-face to a self-administered mixed-mode data collection. The first two studies examine design decisions made prior to data collection. The first study analyzes the effect of different response formats on measurement error in web and mail mode. The second study examines how the mode choice sequence in a self-administered mixed-mode design affects nonresponse bias and whether tailoring the mode choice sequence can reduce nonresponse bias. The third study focuses on the outcomes of a transition by comparing nonresponse and measurement bias between a face-to-face and a self-administered mixed-mode design. The findings of this dissertation provide valuable insights into the consideration of both nonresponse and measurement bias before and after the transition, along with practical implications for survey practitioners transitioning from face-to-face to self-administered surveys.
Dieser Eintrag ist Teil der Universitätsbibliographie.
Das Dokument wird vom Publikationsserver der Universitätsbibliothek Mannheim bereitgestellt.