Reliability of the Comprehensive Scoring System with the Early Memories Procedure

Abstract

This paper reports reliability statistics for the Comprehensive Early Memory Scoring System-Revised (CEMSS-R), a system for analyzing Early Memory Procedure (EMP) data, and provides a brief overview on using early memories in clinical practice. Whereas previous CEMSS-R studies utilized an orally administrated form of the EMP, the current study utilized the written procedure commonly used in clinical settings. The target stimuli were 35 memories from six participants in a treatment program for incarcerated female substance abusers. The protocols were selected to sample a range of psychopathology severity (2 low, 2 middle, 2 high) based on psychological testing and clinical interview material. Raters were three male and two female clinicians having various degrees of familiarity with the EMP and the CEMSS-R. Significant rater sex differences were found on clinician ratings for two continuous measures of Object Relations: Distinctiveness and Interpersonal Contact. Adjusting continuous measures for significant rater sex differences resulted in average intra-class correlations of .78 for two randomly selected raters and .98 for a set of 5 raters. Average Kappa coefficients for two categorical items were .63 and .66 respectively. The authors concluded that the rating system is an efficient, unobtrusive, reliable, and valid means of extracting quantitative data from written memories collected in assessment and therapy.