Selection, training, and quality control of Type A interviewers in a prospective study of young adults

This paper describes Type A/B interviewer selection, training, and quality control results in a prospective study of coronary artery risk development in young adults (CARDIA). Interviewer behaviors from 152 CARDIA structured interviews were audited and compared with 747 Western Collaborative Group Study (WCGS) interviews and 577 Multiple Risk Factor Intervention Trial (MRFIT) interviews. The results show success in modeling the CARDIA interviewer behaviors on those of the WCGS. CARDIA interviews were very similar to WCGS interviews for interview length, number of questions asked, and speed of speaking; they were similar to MRFIT interviews in latency of asking questions. CARDIA interviewer behaviors remained fairly consistent over the four time periods. Comparing the clinics, there were regional differences in latency of asking and speed of speaking, with the Southern clinic having a longer asking latency and speaking more slowly. There were differences between individual interviewers in most characteristics, particularly those that were more free to vary. The study provides procedures and guidelines designed to maintain quality control of the structured interview process.