The study examined three methodological aspects of expressed emotion (EE) as assessed in the course of PACE (Psychosocial Assessment of Childhood Experiences) interviews with a parent. In a sample of 87 children, aged 6 13 years, enrolled in a prospective study examining the role of stress on the course of asthma, EE was assessed at three time points, 9 months apart. A high degree of agreement was found among the three concurrent measures of negative and positive EE (kappas from 0.74 to 0.97, and from 0.45 to 0.88, respectively; p < or = 0.0001 in all instances). The temporal stability of all measures was lower, although statistically significant in all but 2 instances (kappas from 0.19 to 0.59, and from 0.11 to 0.39, respectively). The temporal stability across measures, as well as across interviewers and over time, was broadly similar (kappas from 0.21 to 0.56 for negative EE, and from 0.09 to 0.38 for positive EE, with all but three of the 36 statistically significant). The findings provide support for the underlying assumptions of the PACE-EE and show the utility of measures based on just very brief periods of non-directive interviewing, making them practical in a wide range of studies with EE just one of a larger set of measures.