Evaluation of the Quality Assurance Falsification Interview used in the Census 2000 Dress Rehearsal

In order to reduce errors in surveys, quality assurance (QA) programs are often instituted. The purpose of the falsification interview is to evaluate interviewer's work and identify those who are either falsifying data or are not following correct procedures. One way interviewers are targeted for the falsification interview is by examining the original interview characteristics (such as the length of the interview), respondent characteristics (such as age), or interviewer characteristics (such as being a new interviewer). Tolerance levels are used to determine if an interviewer exceeds acceptable limits. There are three important issues relating to falsification interviews. First, most interviewers do not falsify data, thus finding the interviewer who does is like searching for the "needle in the haystack." Second, because resources are being used to conduct the falsification interviews, the interview must be effective compared to other QA programs. There is also additional respondent burden. To address these issues the following questions are answered: