Eli S. Marks

Influence of the Interviewer on the Accuracy of Survey Results

Abstract This paper reports results of a large scale study of the effect of interviewers on survey results. Where significant effects of the interviewer upon the results are found, the important factors appear to be (1) interviewer “resistance” to a given question—i.e., a tendency to omit or alter the question and/or to assume the answer; (2) relatively high ambiguity, “subjectivity,” or complexity in the concept or wording of the inquiry; (3) the degree to which additional questioning (“probing”) tends to alter initial respondent replies. The study also investigated the relationship of interviewer performance to interviewer characteristics as given by (1) scores on a number of tests; (2) by such personal characteristics as age, sex, occupation, education; and (3) by attitudes toward, and expectations of respondent reactions. At least one measure of poor quality of enumeration (number of improperly omitted Census entries) seems to show substantial correlation with age, with some test scores, and with expe...

Response Errors in Census Research

Abstract Some efforts of the Bureau of the Census to measure and evaluate response errors are described. Particular attention is devoted to the October, 1948, pretest of Census procedures and the measurement of response errors of the various procedures. The procedures were: (1) Self-enumeration, using a household schedule. The schedule was a single sheet of paper 18″ × 24Prime; with 10 columns of questions for the various members of the household—two columns of questions for each individual repeated five times—on one side, and a letter to the household and questions on housing on the other side. (2) Self-enumeration using a 4 page booklet, the page size being 9Prime; × 11Prime;, with Individual Enumeration Forms (9Prime; × 11Prime;) inserted for each member of the household. The questions were printed front and back in two columns on the Individual Enumeration Form. (3) Obtaining information from respondents by direct interview. (4) Obtaining basic information by direct interview and leaving schedule with...