Interviewer Bias Once More Revisited

In 1955, and again in 1964, we conducted an extensive search of the literature dealing with interviewer bias and summarized our findings in articles in the Journal of Marketing and in the Journal of Marketing Research [5 and 6]. The major conclusion of both investigations was that interviewers were a major source of error in marketing field studies and that little was known about the magnitude of such error under varying conditions or ways to minimize it. The 1964 study revealed progress in reducing errors in listing dwelling units, controlling the size and patterns of not-at-homes, and understanding the effects on data collection generated by race, caste, class, deference, age, and sex. Despite such developments, it seemed clear in 1964 that little had been accomplished in dealing effectively with the overall problem. What can we say about this subject today? Has any progress been made in reducing this major source of error? Has research been reported that provides us with important insights concerning the errors which occur because of the interface between respondents and interviewers under a variety of conditions? Do we know any more about how to select, train, and control interviewers so as to reduce bias? This article describes the progress made since. 1964 in providing answers to these and related questions.