Beck Depression Inventory (BDI)

The BDI measures attitudes and symptoms that are characteristic of depression including mood, pessimism, sense of failure, lack of satisfaction, guilt, sense of punishment, self-hate, self-accusations, self punitive wishes, crying spells, irritability, social withdrawal, indecisiveness, body image, work inhibition, sleep disturbance, fatigability, loss of appetite, weight loss, somatic preoccupation, and loss of libido.

Free (original version)

Approx. 5-10 minutes (Ref: 2-3)

Original version in 1961; Revised 1971; Copyright 1978

Fifth to sixth grade reading level. (Ref: 2-3) Items have simple vocabulary and structure. The most difficult words are "repulsive" and "pessimistic."

21 items with four to five response categories ordered by severity.

Originally designed for administration by trained interviewers (clinical psychologist or sociologist), however the BDI is most commonly self-administered.

Total scores are sum of 21 item scores. Scores range from low of 0 to high of 63, with most severe levels of depression indicated by scores in the 40-50 range. Clinically depressed or maladaptive non-clinical patients have been reported by researchers in two studies to score in the 10-30 score range. These researchers interpreted scores of (1) 5-9 as indicating no or minimal depression, (2) 10-18 as indicating mild to moderate depression, (3) 19-29 as indicating moderate to severe depression, (4) 30-63 as indicating severe depression, (5) 0-4 as indicating possible denial of depression, (6) 40-63 as indicating possible exaggeration of depression or histrionic or borderline personality disorder. (Ref: 2,4) (See Validity Evidence below) In one study, mean BDI scores were reported as 10.9 (SD=8.1) for minimal, 18 (SD=10.2) for mild, 25.4 (SD=9.6) for moderate, and 30.0 (SD=10.4) for severe depression. (Ref: 2) No evidence was presented in any of these studies to support the validity of these score interpretations.

The original BDI was revised in 1971 to refine wording and remove double negatives in response choices. Other forms include a card form, computerized forms, the 13-item short form, and the most recent form, the BDI-II. (Ref: 3) Spanish (Kay M and Portillo C. Health Care Women Int. 1989;10:273-93) and German (Kammer D. Diagnostica. 1983;29:48-60) translations have been created.

The original BDI was developed by Aaron T. Beck, CT Ward, M Mendelson, J Mock & J Erbaugh in 1961.

Department of Psychiatry
Research Laboratories
Hospital of the University of Pennsylvania
3400 Spruce Street, Philadelphia, PA

No information found.

mail.med.upenn.edu/~abeck/

top

Some factor analytic studies have reported a primary general depressive factor and others have reported the existence of additional factors. (Ref: 2-3) In a review of 13 factor analytic studies, the number of factors ranged from 3 to 7. Three more recent studies employing latent structure analysis found that one general factor accounts for the majority of variance in BDI items. An additional latent structure analysis found that three highly related factors comprise the one general factor, with those three factors reflecting negative attitudes toward self, performance impairment, and somatic disturbance. (Ref: 2)

No normative information found.

The BDI and a clinical estimate of depth of depression were administered to 38 patients twice with a time interval of 2-6 weeks. The changes on the two measures of depression were found to be similar. These results were presented as evidence of stability of the instrument. (Ref: 1) Test-retest reliability estimates ranged from 0.48-0.86 in five studies with psychiatric patients and time intervals ranging from 5 days to 1 month. (Ref: 2) Test-retest reliability estimates ranged from 0.60-0.90 in five studies with non-psychiatric patients and time intervals ranging from 1 hour to 7 weeks. (Ref: 2)

Scores (number not reported) obtained from 3 interviewers were compared to clinical ratings of depth of depression. Authors reported that a high degree of consistency was found among the mean scores for the interviewers at each depression level (Ref: 1)

The split-half reliability for 97 subjects in the developmental sample was assessed. The Pearson Correlation was 0.86. When the Spearman-Brown correction for reduced length was applied, the correlation was 0.93. (Ref: 1) A meta-analysis of 25 studies reported a mean internal consistency estimate of 0.86, with estimates ranging from 0.73-0.92. (Ref: 2) Internal consistency of White and Mexican-American gerontic subjects (mean age was 70.2 yrs.) was 0.80. (Ref: 4)

Correlations between short and long forms ranged from 0.89-0.97 in three studies. (Ref: 2-3) Correlations between original and revised BDI was 0.94 with 204 undergraduates. (Ref: 2)

The BDI was shown to discriminate between psychiatric and non-psychiatric patients, between patients with major depressive disorders and those with dysthymic disorders, and between adjustment levels of seventh graders. BDI was able to discriminate between patients with and with out self-reported loneliness, stress, anxiety, and general distress. BDI was unable to discriminate between endogenous, involutional, and psychogenic depression.(Ref: 2-3) Some studies have reported higher BDI scores in women than men, others have not. One study reported higher BDI scores in older psychiatric patients than younger ones. Three studies reported higher BDI scores in adolescents than adults. Three studies reported that BDI scores were inversely related to educational level. Some studies have found that blacks have higher BDI scores than whites. The effect sizes of these comparisons between demographic groups were not large. (Ref: 2)

The BDI was compared to four ordered clinical ratings (none, mild, moderate, and severe) of depth of depression. Results indicated that with each increasing category of the criterion measure, BDI scores significantly increased. The Kruskal-Wallis One-way ANOVA by Ranks provided evidence of the association of the BDI to the clinical ratings, p < .001 for both the original sample and replication sample. The Mann-Whitney U test assessing the power of the BDI for discriminating between the clinical ratings of depression was also significant, p < .001 for all categories except differences between moderate and severe. Pearson biserial correlation for evaluating the association between BDI scores and clinical ratings (2 groups including none and mild versus moderate and severe) was 0.65 for original sample and 0.67 for replication sample. (Ref: 1) Correlations between concurrent measures of depression and BDI in 35 studies were reported. In summary, correlations between BDI and clinical ratings ranged from 0.55 to 0.96, correlations between the BDI and the Hamilton Psychiatric Rating Scale for Depression ranged from 0.61 to 0.86, correlations between the BDI and Zung Self Reported Depression Scale ranged from 0.57 to 0.83, and correlations between the BDI and the Minnesota Multiphasic Personality Inventory Depression Scale ranged from 0.56 to 0.75. (Ref: 2-3)

Development of initial instrument based on recorded observations of psychiatrist and consensus from clinicians regarding symptoms of depression. Items were chosen to reflect overt behavioral manifestations of depression. Items were not chosen based on a theory of the etiology of depression. (Ref 1) Six of the DSM-III categories of depression diagnoses are included. (Ref: 2-3)

BDI and clinical rating score changes over 2-5 weeks for 38 patients were examined. Changes in the BDI were greater than changes in the clinical rating scores, as predicted. (Ref: 1)

Yes (Ref: 5)

Yes (Ref: 1-6)