An Investigation of Two Different Methods for Evaluation of Interviewer Job Performance¹

Abstract

BY means of concealed wire recorders and the use of planted respondents in what was ostensibly a routine opinion survey, rather unequivocal data were obtained concerning the cheating and error behavior of fifteen public opinion interviewers during thirty-three recorded interviews. In a surprising number of instances, serious discrepancies were noted between the respondent's actual statements and the replies with which the interviewer had credited him.Schedules for 158 additional unrecorded interviews by the same interviewers were also available for analysis. When the NaMni Stewart has been director of research for The Personnel Laboratory since 1960.

Reliability of retrospective event histories within the German Generations and Gender Survey

“…Therefore interviewers will choose the behaviour from which they expect the highest returns and the lowest costs. Several factors may affect these benefits and costs: 1) questionnaire characteristics, such as length, complexity, and intelligibility of questions, 2) personal questions in the questionnaire which interviewers might be reluctant to ask people they do not know, 3) organisational factors, such as inadequate remuneration, commitment and training of the interviewers, as well as 4) external factors, such as bad weather or a neighbourhood, in which interviewers do not feel comfortable (Crespi 1945;Stewart and Flowerman 1951;Köhne-Finster and Güllner 2009). These factors may lead to "data fabrication" as an attractive option (Crespi 1945;Schräpler and Wagner 2003).…”