Statistical identification of fraudulent interviews in surveys: Improving interviewer controls

Interviewer effects from a total survey error perspective
,
pp. 91-105
,
New York
,
Chapman and Hall/CRC
,
2020,

Kristen Olson, Jolene Smyth, Jennifer Dykema, Allyson Holbrook, Brady T. West,
Schwanhäuser, Silvia, Joseph W. Sakshaug, Yulia Kosyakova, Frauke Kreuter
ISBN: 978-0-367-89631-7 (print), 978-1-00302021-9 (online)
DOI: 
10.1201/9781003020219-10
Abstract

Survey data are important for establishing new insights in many disciplines such as sociology, economics, and others. This chapter provides a broad overview of statistical methods for identifying interviewer falsification and demonstrating promising statistical identification strategies using data from a large-scale refugee survey in Germany that includes confirmed falsifications. In practice, non-statistical strategies for detecting falsifications are usually part of standard quality control methods. Some approaches are used during the field period, while others are used after all interviews have been conducted. Interview monitoring, mainly applied in telephone surveys, is another commonly used procedure for identifying falsifiers, which also serves as a deterrent to interviewers. To illustrate the statistical identification tools, the following indicators are considered: acquiescent response style, extreme response style, interview duration, middle response style, recency effects, semi-open responses, and stereotyping. Statistical identification methods have been demonstrated to be effective and are becoming increasingly popular tools for identifying falsified interviews in surveys.