Problems in the Collection of Opinion-Research Data

A central problem in the collection of data is the effect of interviewer on the quality of the results. This paper describes research findings of the National Opinion Research Center's project on the isolation, measurement, and control of interviewer effect. Variations in results derive from interviewer fallibility, unreliability of respondents, and, finally, interactional processes. Disparities in the group memberships of interviewer and respondent affect the results; beliefs the interviewer has about the respondent produce expectations which in turn affect results. Experiments on the role of situational factors in mediating interviewer effects and on the validity of interviewer data are describred.