Data quality

Interviewer-administered surveys are susceptible to intentional deviant behavior by interview- ers and this type of behavior is a potential source of survey error. One example of such deviant behaviour is the falsification of entire interviews, which can negatively impact data quality if such cases are not identified. Therefore, the development and application of falsification de- tection methods is important to ensure high quality data. However, methods of detecting falsi- fication are usually evaluated using simulated or laboratory data instead of actual falsified data. This report examines the effectiveness of statistical identification methods for detecting falsifi- ers in the IAB-BAMF-SOEP Survey of Refugees in Germany, in which there was an actual case of interviewer fraud.