The measurement of expressed emotion in relationships between staff and service users: The use of short speech samples

Objectives. Research on expressed emotion (EE) has demonstrated a remarkable consistency across cultures and over time; the psychosocial climate in relationships is important in determining the course of problems in mental health. The rating of EE might be described as the least accessible aspect of this literature to those who have not undertaken a training course. The purpose of this study was twofold: first, to obtain EE ratings for staff‐patient relationships via interview and speech sample methods, in order to estimate the validity of the shorter method (the Five‐Minute‐Speech‐Sample, FMSS); second, to examine the generalizability of the FMSS rating method to raters who were not previously trained to rate EE. Design and Method. Staff (N = 15) working in a day hospital service for people with enduring mental health problems were interviewed about their work with at least one patient (N = 32), and also asked to provide an FMSS on each relationship. Ratings of FMSS‐EE were subsequently compared with the Camberwell Family Interview‐EE ratings. Following an hour‐long training period, the FMSS‐EE ratings of five postgraduate students were then compared with those of a criterion rater. Design and Method. Staff (N = 15) working in a day hospital service for people with enduring mental health problems were interviewed about their work with at least one patient (N = 32), and also asked to provide an FMSS on each relationship. Ratings of FMSS‐EE were subsequently compared with the Camberwell Family Interview‐EE ratings. Following an hour‐long training period, the FMSS‐EE ratings of five postgraduate students were then compared with those of a criterion rater. Results. Correspondence between the two measures of EE was found to be good, with overall agreement achieved in 89.7 of cases. Raters with very limited training in the concept and rating of EE were accurate in identifying the overall rating of the relationship in question, but less accurate in identifying specific critical comments. Conclusion. The FMSS technique can be used reliably to identify negative relationships even by raters given very limited training. Clinical and research applications are suggested.