|1.||1) A very business-like interviewer. Asks the most personal questions in a brusque professional manner and gets accurate information in the shortest possible time. Although not very conciliating, her respondents seem to like her. A conscientious, willing worker. Quite popular with the younger set, but her blunt, offhand manner does not seem to be appreciated by her more mature teammates.|
|2.||2) A pretty good (i.e., not very good) interviewer, but a mature personality. Establishes excellent rapport, probes a bit and is somewhat personal. Has quite a conciliatory manner which subdues the most difficult respondent though she does not seem to be a very easy person to get along with in everyday life.|
|3.||3) A conscientious worker who never slacks on the job. Follows the questionnaire but is inclined toward intimate conversation, a very good person for case work, I should imagine.|
The supervisor is struggling to distinguish the different traits which are important for an interviewer. The first interviewer is described as being oustanding in the conduct of the interview itself, but as having personality traits which do not seem likely to help in rapport or in relations on the job. The last of the three interviewers is described in the opposite way: she has the traits which win the confidence of respondents and the admiration of co-workers, but she tends to depart from the interviewing process. In the second description only ability to build rapport is stressed. The definition of the several traits which are important in the interviewer's job is not clear and consistent from one rating to the next. As part of the research program we used measures of different traits during different phases of the study. In this way we attempted to gain greater understanding of the relevant qualities throughout the interviewer performance.
The program of systematic study of interviewer performance falls by necessity into two parts. The interviewers can be tested, their progress in training can be measured, and their performance can be evaluated while the survey is still progressing. On the other hand, the product of their work, the quality of the interviews themselves can only be studied in detail when the survey is analyzed, usually some time after the completion of the field work. We shall call the first part, the measurement of interviewer traits, and the second part the performance criteria. The procedure for collecting data for the measurement of interviewer traits accompanied the selection and training procedure and part of the field work and is described below. The procedure for obtaining the performance criteria was built into the questionnaire. The derivation of certain aspects of performance from the interviews is discussed in a later section.
As we have noted before (pp. 14-15), the screening and selection procedure involved a great deal of collection of personal data and situational tests on which reliable ratings could be based. Throughout the training period, especially in the first days, the interviewers completed additional tests and questionnaires. They were told frankly that, because of the novelty of the whole research design, it was necessary to study both the data collection methods themselves and the people who had the primary responsibility for obtaining them, the interviewers. Completing the tests accurately and truthfully was thus described as an important part of the job.
The progress of the trainees was evaluated as part of the interviewers' training. For this purpose tests, covering different aspects of the subject matter, were administered, and the interviewers rated each other, both in terms of performance and of personal preferences. These ratings were useful in making assignments, in addition to being valuable measures of interviewer traits.
After the interviewers started their field work, preliminary evaluations could be made of their performance by counting gross mistakes in their interviews and by rating their work. These measures are useful for making administrative decisions, e.g., an interviewer can be discharged or retrained on the basis of these ratings. They fit well into the progress of the survey, apart from their use as research measures.
As can be seen, the research program had a practical value during the study: it helped in selection, training, supervision and later in interpretation of the results. However, this connection with practical purposes detracted some from the purity of methodological research. For instance, the screening program was really used to select interviewers and thus it is impossible to determine how the rejected applicants would have turned out. Further, insofar as the screening program was effective, it resulted in a group of interviewers which was quite homogeneous on relevant characteristics. Consequently correlations of these characteristics with performance are necessarily attenuated. Similar influences affected the