In studies where validation data is not available, effective interviewers are identified on the basis of various measures of interviewing performance. Yet little is known about the validity of these measures. In the study reported, sample addresses were selected on a probability basis from the records of financial institutions and the holdings reported to the interviewer were compared with institution records for the day of the interview. The frequency with which an interviewer obtained the validated account(s) forms the basis for the criteria of interviewer effectiveness. This paper suggests that many frequently used measures of interviewer performance are poor indicators of interviewer effectiveness. Hypotheses for future testing are also presented on how the identification procedure might be improved.