Conferences

With regard to the role of interviewers in the data collection process, some ambiguity can be observed. On the one hand they are considered as essential agents. Their task is complex and their responsibilities wide-ranging. Consequently, they can play an important role in the improvement of data quality. On the other hand, interviewers are also a potential source of systematic and/or variable error. This means that interviewers can facilitate good data quality or can cause error. This difference in point of view is also reflected in two different approaches to survey data quality: Total Quality Management (TQM) and Total Survey Error (TSE). It will be demonstrated that both approaches encourage different types of research regarding the impact of interviewer performance on data quality.

The first data quality approach is ‘Total Quality Management’ (TQM). TQM is a comprehensive approach with strong emphasis on organizational performance and process quality being prerequisites for optimum quality of output. It will be shown that, in the context of survey interviews and interviewer performance, this approach considers interviewers as an ‘instrument’ to improve data quality. TQM stimulates research into interviewing techniques, the effects of interviewer training and interviewer behavior during interviews.

The Total Survey Error (TSE) is the second and, in survey methodology, more dominant data quality approach. In this approach a negative definition of quality is used. It is the absence of error. All the different components, actors and activities necessary for survey research are considered as a source of error. Interviewer error is one component of the Total Survey Error. Evaluation of interviewer effects by means of analysis of interviewer variance perfectly fits within the TSE approach. A response and an interviewer oriented evaluation of interviewer effects will be discussed.
An integration of both quality approaches in a Total Survey Data Quality Matrix will be argued.