How Much of Interviewer Variance Is Really Nonresponse Error Variance?

Kish’s (1962) classical intra-interviewer correlation (ρint) provides survey researchers with an estimate of the effect of interviewers on variation in measurements of a survey variable of interest. This correlation is an undesirable product of the data collection process that can arise when answers from respondents interviewed by the same interviewer are more similar to each other than answers from other respondents, decreasing the precision of survey estimates. Estimation of this parameter, however, uses only respondent data. The potential contribution of variance in nonresponse errors between interviewers to the estimation of ρint has been largely ignored. Responses within interviewers may appear correlated because the interviewers successfully obtain cooperation from different pools of respondents, not because of systematic response deviations. This study takes a first step in filling this gap in the literature on interviewer effects by analyzing a unique survey data set, collected using computerassisted telephone interviewing (CATI) from a sample of divorce records. This data set, which includes both true values and reported values for respondents and a CATI sample assignment that approximates interpenBrady T. West is a PhD candidate in the Michigan Program in Survey Methodology (MPSM), University of Michigan, Ann Arbor, Michigan, USA. Kristen Olson is an Assistant Professor in the Survey Research and Methodology Program and the Department of Sociology, University of Nebraska–Lincoln, Lincoln, Nebraska, USA. The authors express sincere thanks to Paul Biemer, Bob Groves, Mick Couper, Frauke Kreuter, Jim Lepkowski, and four anonymous reviewers for helpful comments and guidance on earlier drafts, and to Vaughn Call for providing access to these data. The U.S. National Science Foundation generously provided support [SES-0620228 to K.O.]. The Wisconsin Divorce Study was supported by the National Institute of Child Health and Human Development, the National Institutes of Health [HD-31035 and HD32180–03 to Vaughn Call and Larry Bumpass]. The study was designed and carried out at the Center for Demography and Ecology at the University of Wisconsin–Madison and Brigham Young University under the direction of Vaughn Call and Larry Bumpass. h o w muc h i n te r v i e w er v ar i an c e i s n o n r es p o n s e e r r o r v ar i an c e? 1005 etrated assignment of subsamples to interviewers, enables the decomposition of interviewer variance in means of respondent reports into nonresponse error variance and measurement error variance across interviewers. We show that in cases where there is substantial interviewer variance in reported values, the interviewer variance may arise from nonresponse error variance across interviewers.