Interviewer-administered survey modes are often considered superior data collection methods in terms of data quality. However, interviewers can also negatively affect data quality and, in the worst case, fabricate data. Despite a growing body of literature, interviewer falsification is still an understudied topic; most contributions present one particular identification method, neglecting evaluations of multiple methods. This makes it difficult to design targeted quality controls. This dissertation provides a practical guide to improve quality controls using statistical identification methods to detect various forms of interviewer falsification. The dissertation includes a multitude of falsification indicators, cluster analysis, methods focusing on duplication, and innovative machine learning methods. Analyses are based on different datasets, including real-world survey data and experimental data. Beside complete falsifications, this dissertation also examines the performance of detection methods in the context of partial falsification. With this, this dissertation supports survey practitioners in selecting appropriate tools for their quality controls.
Dieser Eintrag ist Teil der Universitätsbibliographie.
Das Dokument wird vom Publikationsserver der Universitätsbibliothek Mannheim bereitgestellt.
![[img]](https://madoc.bib.uni-mannheim.de/style/images/fileicons/application_pdf.png)
Search Authors in