Journal of Obesity & Metabolic Syndrome

Search

Article

J Korean Soc Study Obes 2004; 13(1): 42-52

Published online January 1, 2004

Copyright © Korean Society for the Study of Obesity.

Reliability and Factor Analysis of The Korean Version of Eating Disorder Examination

Si-Young Heo M.D., Min-Kyu Rhee1 Ph.D., Young-Min Choi2 M.D.,
Chang-Ho Sohn3 M.D., Hye-Kyung Lee M.A., Young-Ho Lee M.D.

Nanum Obesity and Eating Disorders Clinic, Seoul, Korea, Department of Psychology, Gyeongsang National University, Jinju, Korea1, Department of Neuropsychiatry, Sanggye Paik Hospital, Inje University, Seoul, Korea2, Department of Neuropsychiatry, Seoul Paik Hospital, Inje University, Seoul, Korea3

Abstract

Background: This study was attempted to investigate the reliability, factor analysis and norms of the Korean Version of Eating Disorder Examination (KEDE).
Methods: The KEDE was administered to 106 female patients with eating disorders and 101 normal female college students by two raters who were proven to have a high inter-rater reliability. Construct validity was tested by an exploratory factor analysis. Inter-rater reliability was tested by Pearson's correlation coefficiency and internal consistency was tested by Cronbach's alpha. The author also tried to make T score norms of the KEDE using T score 70 as a cutoff point.
Results: By a factor analysis, 4 factors were extracted. Factor 1 was ‘eating concern' which was consisted of 5 items with 17.86% of variance, factor 2 was ‘weight concern' which was consisted of 8 items with 12.24% of variance, factor 3 was ‘restraint’ which was consisted of 5 items with 12.23% of variance, and factor 4 was ‘shape concern’ which was consisted of 4 items with 11.55% of variance. Inter-rater consistency was r=1.0 for ‘eating concern', r=.98 for ‘weight concern', r=.99 for ‘restraint’, and r=.99 for ‘shape concern’. Cronbach's internal consistency was .78 for ‘restraint’, .89 for ‘eating concern’, .72 for ‘shape concern’, and .89 for ‘weight concern’. T score norms as a cutoff score were 16 for ‘restraint’, 7 for ‘eating concern’, 21 for ‘shape concern’, and 35 for ‘weight concern’.
Conclusion: These results support that the KEDE is a reliable and valid tool for evaluating the patients with eating disorder and related eating problems for the various purposes. The authors also demonstrate T score norms of the KEDE as a cutoff score of each subscale.

Keywords: Eating disorders, KEDE, Reliability, Factor analysis, Norms