Children's type A interview: interrater, test-retest reliability, and interviewer effect

The Miami Structured Interview--1 was developed to assess Type A behavior in preadolescents and adolescents in the United States and Greece. This report describes the interrater and test-retest reliability of the Miami Structured Interview--1, and the effect of different interviewers on the assessment process. Interrater agreements on the A vs not A scale were 73% for Greek American (k = 0.49, p less than 0.01, n = 88), 80% for Mixed American (k = 0.61, p less than 0.01, n = 69), and 88% for Native Greek interviews (k = 0.76, p less than 0.01, n = 65). Exact agreements on a five-point scale were 48% (weighted k = 0.73, p less than 0.01), 42% (weighted k = 0.70, p less than 0.01), and 48% (weighted k = 0.80, p less than 0.01), respectively. Although levels of agreement were all greater than chance, rater bias was observed. Test-retest administrations within 10 min showed 91% agreement (k = 0.82, p less than 0.01) on the A vs not A scale. Analyses suggest interviewers can affect the assessment of Type A behavior. These findings indicate the Miami Structured Interview--1 can (1) reach acceptable levels of interrater reliability, (2) is susceptible to rater bias, (3) has sufficient test-retest reliability for use in longitudinal studies, and (4) must be presented in a standardized way before ratings from interviews conducted by different interviewers can be combined or compared. The strengths and limitations of the Miami Structured Interview-1 are discussed.