Decomposing the Interviewer Variance Introduced by Standardized and Conversational Interviewing

• Activity Number: 426
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Brady T. West* and Frederick G. Conrad and Frauke Kreuter and Felicitas Mittereder and Jen Durow
• Companies: University of Michigan Institute for Social Research and Institute for Social Research and Joint Program in Survey Methodology and Institute for Social Research and Institute for Social Research
• Keywords: interviewer effects ; interviewer variance ; total survey error ; conversational interviewing ; standardized interviewing ; multilevel modeling

Abstract

Recent research has attempted to decompose interviewer variance into its potential sources, using the Total Survey Error framework. These sources include sampling error variance, or a breakdown of interpenetrated assignment of sampled cases to interviewers; nonresponse error variance, or variance among interviewers in the types of cases recruited; and measurement error variance, or variance among interviewers in mean response errors. No research to date has attempted to compare this decomposition for different interviewing techniques. In this study, we will perform a descriptive decomposition of the interviewer variance associated with conversational (CI) and standardized (SI) interviewing. We will present results from an experimental study mounted in Germany, where a national sample of employed individuals was measured on a variety of complex employment items by interviewers randomly assigned to use either CI or SI. This study featured an interpenetrated area sample design, enabling estimation of interviewer variance, and the presence of administrative information on the sampling frame enabled estimation of interviewer variance in the various sources of error for each technique.