Abstract
Evidence is now beginning to accumulate that shows that interviewer attitudes, personality, and behavior are predictive of success in achieving contact and cooperation with sampled households. A less frequently explored possibility, however, is that these same characteristics might also be the source of variability in the extent to which interviewers follow best practices in the implementation of standardized interviewing. That is to say, there may be a correlation between interviewer-induced nonresponse bias and measurement error. In this article, we provide the first empirical investigation of the direction and magnitude of the relationship between interviewer skill in obtaining contact and cooperation and correlated interviewer error. Drawing on face-to-face interview data from a large, multistage probability sample of the British population, we use cross-classified multilevel models with a complex error structure to examine how the interviewer variance component varies as a function of historical measures of interviewer skill in obtaining contact and cooperation. Our results show that, across a broad range of variables, interviewers with a history of obtaining poor rates of contact and cooperation exhibit higher levels of correlated interviewer error than their better-performing colleagues. For cooperation, we find some evidence of a U-shaped relationship, with the least and the most successful interviewers having the largest interviewer variance component.