Psychometric properties of the Beck Depression Inventory-II in nonclinical adolescent samples

• Published in: Journal of clinical psychology 2008-01, Vol.64 (1), p.83-102
• Main Authors: Osman, Augustine, Barrios, Francisco X., Gutierrez, Peter M., Williams, John E., Bailey, Jennifer
• Format: Article
• Language: English
• Subjects: Adolescent; Biological and medical sciences; Cognition & reasoning; Correlation analysis; Depression - diagnosis; depression severity; Discriminant analysis; Estimates; factor analyses; Female; Humans; Male; Medical sciences; Mental depression; Midwestern United States; nonclinical adolescents; Psychology. Psychoanalysis. Psychiatry; Psychometrics; Psychometrics. Diagnostic aid systems; Psychopathology. Psychiatry; Studies; Surveys and Questionnaires; Techniques and methods; Teenagers
• ISSN: 0021-9762; 1097-4679
• DOI: 10.1002/jclp.20433

This study examined the factor structure and psychometric properties of the Beck Depression Inventory‐II (BDI‐II; A. T. Beck, Steer, & Brown, 1996) in samples of high‐school adolescents (N=414; 210 boys and 204 girls, ages 14–18 years). Confirmatory factor analyses provided satisfactory fit estimates for the two‐ and three‐factor oblique solutions reported frequently in the extant literature. The solution to a general factor with domain‐specific somatic and cognitive‐affective factors was examined as an alternative model to previously established models for the current high‐school sample data. Results provided stronger support for the general factor model. Estimates of internal consistency for scores on this instrument were high (coefficient α=.92, average interitem correlation=.35). The mean BDI‐II total score for the nonclinical samples (M=12.50, SD=10.50) was compared with the mean scores reported for various adolescent normative samples in the extant literature. The BDI‐II total score correlated moderately and significantly with scores on self‐report measures of hopelessness (r=.63), anxiety (r=.53), and suicide‐related behaviors (r=.57), providing support of adequate correlates for the BDI‐II. Estimates of known‐groups validity were evaluated using data from a small sample of 167 clinical adolescent inpatients. Specific study findings, limitations, and recommendations are discussed. © 2007 Wiley Periodicals, Inc. J Clin Psychol 64: 83–102, 2008.