Suchergebnisse

During the past 20 years, survey methodology has undergone a paradigm shift. The old paradigm was based on a statistical model that focused on the effects of survey errors on the estimates derived from survey data. The new paradigm is based on a social scientific model that focuses on the causes of survey errors. Several developments have helped bring about this shift -- the application of methods & concepts from cognitive psychology to the reduction of survey measurement error, the development of new computerized methods of data collection, & the increase in concern about measurement & nonresponse as sources of error in survey estimates. The new paradigm has little to say about the topics, such as sampling error, that were central to the old one; similarly, the old paradigm had little to say about how to reduce or prevent errors, a major concern for the new one. Thus, the two paradigms do not clash so much as complement each other. 17 References. Adapted from the source document.

Introduction / Uwe Engel -- Motivated misreporting : shaping answers to reduce survey burden / Roger Tourangeau, Frauke Kreuter, and Stephanie Eckman -- Audio-recording of open-ended survey questions : a solution to the problems of interviewer transcription? / Patrick Sturgis and Rebekah Luff -- Framing effects / Uwe Engel and Britta Köster -- Estimating and comparing the quality of different scales of an online survey using an MTMM approach / Melanie Revilla and Willem E. Saris -- Collecting MTMM data on satisfaction with life / Laura Burmeister and Uwe Engel -- On the quality of web panels / Jelke Bethlehem -- Online surveys and the burden of mobile responding / Marika de Bruijne and Marije Oudejans -- Well-being, survey attitudes, and readiness to report on everyday life events in an experience sampling study / Laura Burmeister, Uwe Engel, and Björn Oliver Schmidt -- Nonresponse, measurement error, and estimates of change : lessons from the German PPSM panel / Suat Can and Uwe Engel -- Handling of missing data in statistical analyses / Daniel Salfrán and Martin Spiess -- Multiple imputation of overdispersed multilevel count data / Kristian Kleinke and Jost Reinecke

Wissenschaftliche Umfragen können keine aussagekräftigen Ergebnisse liefern, wenn ihre Datenqualität durch fehlende oder verfälschte Antworten beeinträchtigt wird. Eine Herausforderung der Sozialforschung besteht darin, solche Fehlerquellen zu erkennen und zu kontrollieren. Der Band präsentiert Erkenntnisse und Methoden zur Behandlung von Unit Nonresponse, Missing Data und verschiedene Arten von Messfehlern im Kontext von Web und Mixed-Mode Panel, Mobile Web und Faceto-Face-Befragungen. Uwe Engel ist Professor für Soziologie mit dem Schwerpunkt Statistik und empirische Sozialforschung an der Universität Bremen.

High-quality data from large-scale surveys provide a solid basis for outstanding research in the social sciences. Because of the unique demands of survey measurement in terms of the resources and skills required, it should be viewed as a specific sector of the research data infrastructure. In Germany, large-scale surveys have been established both within and outside academia, and major new projects are underway. Clearly, the sector is expanding. There is a need to discuss future challenges, not only with a focus on individual large projects, but with a view to the sector of large-scale survey measurement in general. One aspect is the segmentation of large-scale survey measurement in Germany along institutional lines (statistical offices, ministerial or government agency research (Ressortforschung), public research institutions, and the academic community). Here, we recommend that an overall framework be developed covering all sub-sectors. A second aspect is the infrastructure required for largescale, high-quality data collection. In Germany (outside the sector of Statistical Offices), this infrastructure is provided by private survey organisations. We argue that these should be recognised as relevant actors within the research data infrastructure. They have to invest in technological and human resources in order to provide the professional services required, and they need conditions and forms of cooperation that encourage this investment.

"High-quality data from large-scale surveys provide a solid basis for outstanding research in the social sciences. Because of the unique demands of survey measurement in terms of the resources and skills required, it should be viewed as a specific sector of the research data infrastructure. In Germany, large-scale surveys have been established both within and outside academia, and major new projects are underway. Clearly, the sector is expanding. There is a need to discuss future challenges, not only with a focus on individual large projects, but with a view to the sector of large-scale survey measurement in general. One aspect is the segmentation of large-scale survey measurement in Germany along institutional lines (statistical offices, ministerial or government agency research (Ressortforschung), public research institutions, and the academic community). Here, we recommend that an overall framework be developed covering all sub-sectors. A second aspect is the infrastructure required for largescale, high-quality data collection. In Germany (outside the sector of Statistical Offices), this infrastructure is provided by private survey organisations. We argue that these should be recognised as relevant actors within the research data infrastructure. They have to invest in technological and human resources in order to provide the professional services required, and they need conditions and forms of cooperation that encourage this investment." [author's abstract]