Assignment: Evaluating Assessment Instruments

Beck Depression Inventory [1993 Revised]

 

Authors: Beck, Aaron T.; Steer, Robert A.

 

Publication Date: 1961-1993.

 

Publisher Information: The Psychological Corporation.

 

Purpose: Designed to assess the severity of depression in adolescents and adults.’

 

Test Category: Personality.

 

Population: Adolescents and adults.

 

Scores: Total score only; item score ranges..

 

Administration: Group or individual.

 

Time: (5-15) minutes.

 

Price Data: 1994: $46 per complete kit including 25 record forms and manual (’93, 24 pages); $25.50 per 25 record forms; $22.50 per manual. Assignment: Evaluating Assessment Instruments

 

Cross References: See T4:268 (660 references);  for reviews by Collie W. Conoley and Norman D. Sundberg of a previous edition, see 11:31 (286 references). Assignment: Evaluating Assessment Instruments

 

Reviewers:Carlson, Janet F.; Waller, Niels G..

 

Special Editions: Also available in Spanish..

 

Yearbook Volume:13.

 

Yearbook Reference:J. C. Impara & B. S. Plake (Eds.), The thirteenth mental measurements yearbook. 1998.

 

:

Beck Depression Inventory [1993 Revised]. Purpose: ‘Designed to assess the severity of depression in adolescents and adults.’ Population: Adolescents and adults. Publication Dates: 1961-1993. Acronym: BDI. Scores: Total score only; item score ranges. Administration: Group or individual. Price Data, 1994: $46 per complete kit including 25 record forms and manual (’93, 24 pages); $25.50 per 25 record forms; $22.50 per manual. Foreign Language Edition: Also available in Spanish. Time: (5-15) minutes. Authors: Aaron T. Beck and Robert A. Steer. Publisher: The Psychological Corporation. Cross References: See T4:268 (660 references); for reviews by Collie W. Conoley and Norman D. Sundberg of a previous edition, see 11:31 (286 references). Assignment: Evaluating Assessment Instruments

Accession Number:

test.391

Beck Depression Inventory [1993 Revised]

Full Text

Listen

Review of the Beck Depression Inventory by JANET F. CARLSON, Associate Professor, Counseling and Psychological Services Department, State University of New York at Oswego, Oswego, NY: The Beck Depression Inventory (BDI) is a well-known and widely used self-report inventory that taps overall severity of depression in adolescents and adults. The original BDI was developed by Beck and his associates in 1961 (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961) and revised in 1971, at which time it was introduced at the Center for Cognitive Therapy (CCT) of the University of Pennsylvania Medical School where a large portion of research bearing on the BDI has been conducted. The current edition consists of 21 symptoms and attitudes, which the subject rates on a 4-point scale of severity. Test takers are asked to rate the items for the past week, including the day on which the test is taken. The items cover cognitive, affective, somatic, and vegetative dimensions of depression, although the inventory itself was developed atheoretically. Assignment: Evaluating Assessment Instruments

Collectively, the items correspond reasonably well to the symptoms of depression listed in the Diagnostic and Statistical Manual of Mental Disorders (4th ed.) (DSM-IV, American Psychiatric Association, 1994). APPLICATIONS. The use of the BDI has expanded well beyond its original intended application with psychiatric populations. In addition to its continued use among this population, it is accepted and commonly used by clinicians as a screening instrument among normal populations. The test authors are careful to avoid endorsing the use of the BDI for purposes other than those for which it was developed, and simply acknowledge that many practitioners have found the inventory useful in these other contexts. The test authors indicate the vast amount of research and attending literature has supported the use of the BDI in myriad applications, from research, to screening, to assessment of therapeutic outcomes. Despite the extensive body of research bearing on numerous aspects of the BDI, the test manual itself contains only “an overview of the published information on the revision” despite noting that “specific characteristics of the revised BDI have not previously been described in detail” (p. 1). ADMINISTRATION, SCORING, AND INTERPRETATION. The BDI may be administered individually or in group format, in written or oral form. Instructions to the test taker were modified slightly from the previous edition, as the current version directs test takers to described themselves or their feelings over the past week, including the present day. Thus, the 1993 edition taps more trait aspects of depression whereas the earlier version appears to have measured state aspects of depression. Assignment: Evaluating Assessment Instruments

The test manual authors indicate that total administration time is no more than 15 minutes, irrespective of the mode of administration. A total score is obtained by simply summing the ratings given by the test taker on all items. Interpretation is based on the total score, which may range from 0 to 63. Among depressed patients, scores in the 0-9 range denote “Minimal” depression, 10-16 suggest “Mild” depression, 17-29 are considered “Moderate”, and scores in the 30-63 range indicate “Severe” levels of depression. Within the normal population, total scores above 15 may be indicative of possible depression, although further assessment would be essential in order to confirm the presence of depression. Computer software is available from The Psychological Corporation, which offers the Beck Computer Scoring program (Beck & Steer, 1992). The program scores and interprets the BDI, as well as the several other instruments developed by Beck. The test authors note that in addition to the total score, responses to individual items should be considered in an effort to understand and extract clinically meaningful data. In particular, the clinical relevance of Item 2, which the authors term the pessimism/hopelessness item, and Item 9, the suicide ideation item, is noted. Because the severity of depression, as assessed by the BDI, is independent of symptom types, the test authors suggest that overall patterns of responses be considered in order to ascertain whether the individual patient demonstrates symptoms that are more cognitive, more somatic, more vegetative, or more affective. Elucidating the nature of the most prominent symptoms may have important treatment implications. TECHNICAL ASPECTS. The section of the test manual that covers psychometric characteristics presents information from six normative-outpatient samples used by the CCT to establish reliability estimates and provide evidence of scale validation. The total normative sample consisted of 944 outpatients with mixed diagnoses (n = 248), single episodes of major depression (n = 113), recurrent episodes of major depression (n = 168), Dysthymic disorder (n = 99), alcoholism (n = 105), or heroine addiction (n = 211). Gender and race distributions of the six normative samples are presented in the test manual, as are means, standard deviations, percentages of item endorsement, and item-total correlations. Alpha reliability coefficients across the six samples ranged from .79 to .90. Item-total correlations ranged from .07 to .68, with most values being in the .30 or better range. The authors cite several studies related to the stability of BDI scores over time, using patient and nonpatient groups. It appears that nonpsychiatric samples demonstrate somewhat more stable BDI scores (in the .60 to .90 range) than patient samples (in the .48 to .86 range). These results are not unexpected, given that depressed individuals who are undergoing treatment are expected to improve. Validation evidence for the BDI is provided in the test manual under the headings of content, discriminant, construct, and concurrent validity. The test authors report on two studies that compared the content of the revised BDI to DSM-III criteria for affective disorders and found that two-thirds of the criteria were addressed by the BDI items. The authors note that the lack of items related to the omitted criteria was intentional, and provide an appropriate rationale for not including items that reflected these criteria. The test authors cite several studies bearing on the ability of the BDI to differentiate psychiatric patients from normals, patients with Dysthymic Disorder from those with Major Depressive Disorder, and patients with Generalized Anxiety Disorders from those with Major Depressive Disorder. Based on research findings, the test authors extracted two subscales that can be calculated by summing the ratings for appropriate items. Items 1 through 13 comprise the cognitive-affective subscale, and Items 14 through 21 comprise the somatic-performance subscale. A score greater than 10 on the cognitive-affective subscale indicates moderate depression. Correlation coefficients between the BDI and the Beck Hopelessness Scale and the Hamilton Rating Scale for Depression ranged from .38 to .76 and from .40 to .87, respectively, across the six normative-outpatient samples. Among patients with mixed depressive disorders, correlation coefficients between the BDI and the Symptom Checklist-90-Revised and the Minnesota Multiphasic Personality Inventory–Depression scale were .76 and .61, respectively. Factor analyses with clinical and nonclinical samples have been conducted. Some of the findings are presented briefly in the test manual. The test authors note the number of factors extracted varied with the characteristics of the samples used, and with the method of extraction employed. Brown, Schulberg, and Madonia (1995) indicate that the many factor analyses of the BDI have found anywhere from three to seven factors and note that “[s]tudies using latent structure analysis suggest that the BDI represents one general syndrome of depression that subsumes three highly inter-correlated factors … reflect[ing] negative attitudes toward self, performance impairment, and somatic disturbance, as originally described by Beck and Lester” (p. 59). Compared to other measures of depression, the BDI appears to tap more of the cognitive or cognitive-affective components of depression than other instruments such as the Hamilton Rating Scale for Depression (263) and the Zung Self-Rating Scale (e.g., Brown, Schulberg, & Madonia, 1995; Lambert, Hatch, Kingston, & Edwards, 1986; Santor, Zuroff, Ramsay, Cervantes, & Palacios, 1995). Because so much additional information concerning the psychometric properties of the BDI is contained in the literature and is not contained in the test manual, potential users would be well advised to consult the literature for further information. Prior reviews of the earlier editions of the BDI have indicated the importance of doing so as well. One previous reviewer (Conoley, 1992) summarized the findings of a reference that the reviewer found particularly useful, in that it presents a review and meta-analysis of the reliability and validity literature for the BDI (Beck, Steer, & Garbin, 1988). Readers are referred to these sources for elaboration. CRITIQUE. In general, the information referenced in the test manual and elsewhere is favorable and supportive of the BDI as far as its use with the intended population and, perhaps, others (Brown, Schulberg, & Madonia, 1995; Conoley, 1992; Santor, Zuroff, Ramsay, Cervantes, & Palacios, 1995; Stehouwer, 1985; Sundberg, 1992). The BDI has been widely used over at least the last 25 years, and has been recognized for its solid contributions to the measurement of depression. Potential users must be prepared, however, to ferret out some information beyond that presented in the test manual that may affect their choice of instruments. In particular, test users should bear in mind that the BDI is not intended to be a diagnostic instrument. Thus, in clinical applications, it is best regarded as a screening instrument for depression or as an indicator of the extent of depression, and should not serve as the sole means by which depression is assessed. The availability of software to score and interpret the BDI makes its use as a screening and/or research instrument quite practical and it seems likely that the BDI will continue to enjoy widespread applications in these venues. In situations where the ratings given to somatic or performance items could be attributable to another cause (e.g., a medical condition where fatigue is a symptom or side effect of treatment), the use of the cognitive-affective subscale may be particularly helpful, although further evidence of the validity and reliability of the subscale is needed. Similarly, validity and reliability evidence to establish the probity of various modifications to administration noted in the test manual would be useful additions. All items on the BDI contribute to the total score (i.e., there are no filler items) and items clearly are aimed at assessing aspects of depression. The four response options presented for each item are numbered from 0 to 3, in order of increasing severity, and scoring blanks are presented on each side of the test form. These characteristics make administration, scoring, and interpretation straightforward and, for many users, probably contribute to the appeal of this instrument. But these traits make faking rather easy as well. In cases where test takers might be motivated to deceive (e.g., competency to stand trial, custody hearings, involuntary commitment procedures, social desirability), the test user is advised to use additional or less transparent means of assessment. SUMMARY. The BDI has made, and is likely to continue making, a noteworthy contribution to the assessment of depression. The expansion of its use beyond that originally intended by the test authors is likely attributable to its obvious strengths. Working in its favor are the facts that it is a simple measure that is easily and rapidly administered, encompasses the majority of symptoms associated with depression, has been well researched, can be scored and interpreted via computer software, and can be considerably less expensive than other screening or research tools that require individual administration. REVIEWER’S REFERENCES Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561-571. Stehouwer, R. S. (1985). [Review of Beck Depression Inventory]. In D. J. Keyser & R. C. Sweetland (Eds.), Test critiques (Vol. II) (pp. 83-87). Kansas City, MO: Test Corporation of America/Westport Publishers. Lambert, M. J., Hatch, D. R., Kingston, M. D., & Edwards, B. C. (1986). Zung, Beck, and Hamilton rating scales as measures of treatment outcome: A meta-analytic comparison. Journal of Consulting and Clinical Psychology, 54, 54-59. Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77-100. Beck, A. T., & Steer, R. A. (1992). Beck computer scoring. San Antonio, TX: The Psychological Corporation. Conoley, C. W. (1992). [Review of the Beck Depression Inventory (Revised Edition)]. In J. J. Kramer & J. C. Conoley (Eds.), The eleventh mental measurements yearbook (pp. 78-79). Lincoln, NE: The Buros Institute of Mental Measurements. Sundberg, N. D. (1992). [Review of the Beck Depression Inventory (Revised Edition)]. In J. J. Kramer & J. C. Conoley (Eds.), The eleventh mental measurements yearbook (pp. 79-81). Lincoln, NE: The Buros Institute of Mental Measurements. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Brown, C., Schulberg, H. C., & Madonia, M. J. (1995). Assessing depression in primary care practice with the Beck Depression Inventory and the Hamilton Rating Scale for Depression. Psychological Assessment, 7, 59-65. Santor, D. A., Zuroff, D. C., Ramsay, J. O., Cervantes, P., & Palacios, J. (1995). Examining scale discriminability in the BDI and CES-D as a function of depressive severity. Psychological Assessment, 7, 131-139. Assignment: Evaluating Assessment Instruments

Review of the Beck Depression Inventory by NIELS G. WALLER, Associate Professor of Psychology, University of California, Davis, CA: The Beck Depression Inventory (BDI) is a brief self-report measure of depressive symptoms in adolescents and adults. It has been a clinical mainstay for more than 35 years (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961) and, not surprisingly, the BDI has spawned a rich and extensive research literature. To date, the BDI-in one form or another-has been used in over 3,000 studies. Most of these have used a 21-item long form (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961), although authorized (Beck & Beck, 1972) and unauthorized short forms have also been used. A new edition of the BDI, called the BDI-II (Beck, Steer, & Brown, 1996) is scheduled for release in 1996. At the time of this writing the BDI-II was not available; thus my review focuses on the BDI, although I also discuss some promised features of the BDI-II. The BDI can be administered via paper and pencil, a computer, or orally in approximately 5-15 minutes. In other words, in less time than it takes to drink a cup of coffee, a clinician or researcher can gather information on 21 signs of depression severity: (1) Sadness, (2) Pessimism, (3) Sense of Failure, (4) Dissatisfaction, (5) Guilt, (6) Punishment, (7) Self-dislike, (8) Self-accusations, (9) Suicidal ideas, (10) Crying, (11) Irritability, (12) Social withdrawal, (13) Indecisiveness, (14) Body Image Change, (15) Work difficulty, (16) Insomnia, (17) Fatigability, (18) Loss of Appetite, (19) Weight Loss, (20) Somatic Preoccupation, and (21) Loss of Libido. These symptoms were not selected to reflect any particular theory of depression. The items of the BDI-II, on the other hand, are tailored to the current criteria of depression as outlined in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). BDI items comprise four self-descriptive statements. For example, Item 10-which measures Crying-reads as follows: “(0) I don’t cry any more than usual; (1) I cry more now than I used to; (2) I cry all the time now; (3) I used to be able to cry, but now I can’t cry even though I want to” (p. 1, BDI Questionnaire). Notice that each statement is preceded by a scoring weight. The weights range from 0 to 3; consequently BDI total scores range from 0 to 63. The scoring weights were rationally derived such that higher weights signify greater symptom severity in clinical populations. For some items, as discussed below, the scoring weights are not appropriate for nonclinical samples. All items are face valid (transparent) and scored in the same direction. Thus, BDI scores can be easily distorted with respect to faking (Beck & Beamesderfer, 1974; Dahlstrom, Brooks, & Peterson, 1990; Lees-Haley, 1989); obviously, the scale should be administered only to cooperative examinees with no motivation to malinger. The BDI was originally designed to measure the severity of depressive symptoms in clinical samples. Nevertheless, it is frequently used as a screening instrument in nonclinical samples (Barrera & Garrison-Jones, 1988) even though the manual author warns “there is considerable debate concerning the use of the BDI for screening” (p. 2). Santor, Ramsay, and Zuroff (1994) have recently shown why the BDI can perform nonoptimally in nonclinical populations. Using a nonparametric item response model (Ramsay, 1991), these authors compared the BDI item-scoring weights in clinical and nonclinical samples. They found the estimated and original (rationally derived) weights were similar in their clinical sample but that “differences implied by the a priori weights may not be warranted for some options in the nonpatient college sample” (p. 266). For example, regarding Item 10, their analyses indicated that “at any level of depression, it is more likely that an individual [from a nonclinical sample] will choose Option 3 than Option 2″ (p. 259, italics added). The aforementioned scoring problems may be alleviated in the BDI-II. The product bulletin notes that “Dr. Beck … used Item Response Theory (IRT) to examine how well the four response options are differentiated from each other, and how well the set of response options measures the underlying dimension (latent trait) of self-reported depression. Based on these analyses … several response options were reworded and subsequently tested on a large clinical sample (N = 500).” By relying on IRT while revising the BDI, Beck and his colleagues have set a new standard for clinical test revision. I hope Beck or other researchers will publish IRT item parameter estimates (Reise & Waller, 1990) for the BDI and BDI-II so that these scales can be administered by a computerized adaptive test (Waller & Reise, 1989). The BDI manual contains a weak and inadequate summary of the scale’s properties. Rather than reviewing the psychometric characteristics of the BDI in sufficient detail, the manual directs readers elsewhere for more comprehensive reviews (e.g., Beck, Steer, & Garbin, 1988). Much of the manual is a presentation of BDI summary statistics-such as item means, standard deviations, endorsement percents, and item-total correlations-for six normative outpatient samples: (a) mixed diagnostic; (b) major depression, single episode; (c) major depression, recurrent episode; (d) dysthymic disorder; (e) alcoholism; and (f) heroin addiction. Data from nonclinical samples are conspicuously absent. For the clinical samples, the BDI reliabilities are uniformly high (mean coefficient alpha = .86). The product bulletin promises that the BDI-II is even more reliable. The weakest parts of the manual are the sections on test validity. Five types of validity are discussed: (a) content, (b) discriminant, (c) construct, (d) concurrent, and (e) factorial. The content and organization of these sections leaves much to be desired. Consider, for example, that under the heading: Discriminant Validity, we find evidence that the BDI discriminates psychiatric patients from normal controls. This is surely a desiderata, but it does not speak to the discriminant validity of the instrument. As Campbell and Fiske (1959) remind us, tests are invalidated when they yield high correlations with other tests purporting to measure different things. In other words, they are invalidated when they fail to discriminate between conceptually distinct constructs-such as depression and anxiety (Clark & Watson, 1991). I hope a future manual will report correlations between the BDI and scales that are not designed to measure depression, such as the Beck Anxiety Inventory (Beck & Steer, 1990; 000). Our review materials also included a copy of the Beck Computer Scoring Program (BCS, version 1). This stand-alone product can be used to (a) administer, (b) score, and (c) interpret the BDI and several other Beck scales (The Beck Hopelessness Scale [000], the Beck Anxiety Inventory [000], and The Beck Scale for Suicidal Ideation [000]). Overall, the program passed a very rigorous test: I was able to load the program, take the BDI, and generate an interpreted report without opening the manual. Noteworthy features of the BCS include (a) the ability to return to previously administered items and (b) the ability to display the test instructions during any point of the scale administration. I found the interpretive summaries to be both useful and conservative, though they err on the side of redundancy. For example, a sample report for Jane Doe concludes that: “Ms Doe expresses higher levels of hopelessness and anxiety than depression. Viewed together, the moderate hopelessness, severe anxiety, and mild depression indicate that she views the future with considerable hopelessness and anxiety but is not very depressed by this view.” Scores on the computer-administered BDI are comparable to those from the paper-and-pencil version (Steer, Rissmiller, Ranieri, & Beck, 1994). As a clinician trained in psychometric theory, I am skeptical of 5-minute assessment scales. Nevertheless, as a Minnesota-trained dust-bowl empiricist, I also listen to the data. In the case of the BDI, the data speak loud and clearly. The BDI is an excellent measure of depressive symptoms when it is used in clinical samples with cooperative subjects. The BDI-II promises to be a useful screening measure in nonclinical samples. In situations where subjects may be motivated to malinger, other scales with validity checks, such as the Minnesota Multiphasic Personality Inventory-2 (MMPI-2) (Butcher, Dahlstom, Graham, Tellegen, & Kaemmer, 1989; T4:1645) should provide more trustworthy data. REVIEWER’S REFERENCES Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81-105. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561-571. Beck, A. T., & Beck, R. W. (1972). Screening depressed patients in a family practice. Postgraduate Medicine, 52(6), 81-85. Beck, A. T., & Beamesderfer, A. (1974). Assessment of depression: The depression inventory. In P. Pichot (Ed.), Modern problems in pharmacopsychiatry (pp. 151-169). Basel, Switzerland: Karger. Barrera, M., & Garrison-Jones, C. V. (1988). Properties of the Beck Depression Inventory as a screening instrument for adolescent depression Journal of Abnormal Child Psychology, 16(3), 263-273. Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77-100. Butcher, J. N., Dahlstrom, W. G., Graham, J. R., Tellegen, A., & Kaemmer, B. (1989). Minnesota Multiphasic Personality Inventory-2 (MMPI-2): Manual for administration and scoring. Minneapolis: University of Minnesota Press. Lees-Haley, P. R. (1989). Malingering traumatic mental disorder on the Beck Depression inventory: Cancerphobia and toxic exposure. Psychological Reports, 65, 623-626. Waller, N. G., & Rise, S. (1989). Computerized adaptive personality assessment: An illustration with the Absorption scale. Journal of Personality and Social Psychology, 57, 1051-1058. Beck, A. T., & Steer, R. A. (1990). Manual for the Beck Anxiety Inventory. San Antonio, TX: The Psychological Corporation. Dahlstrom, W. G., Brooks, J. D., & Peterson, C. D. (1990). The Beck Depression Inventory: Item order and the impact of response sets. Journal of Personality Assessment, 55, 224-233. Reise, S. P., & Waller, N. G. (1990). Fitting the two-parameter model to personality data. Applied Psychological Measurement, 14, 45-58. Clark, L. A., & Watson, D. (1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100, 316-336. Ramsay, J. O. (1991). Kernel smoothing approaches to nonparametric item characteristic curve estimation. Psychometrika, 56, 611-630. Santor, D. A., Ramsay, J. O., & Zuroff, D. C. (1994). Nonparametric item analyses of the Beck Depression Inventory; Evaluating gender item bias and response option weights. Psychological Assessment, 6, 255-270. Beck, A. T., Steer, R. A., & Brown, G. (1996). Beck Depression Inventory II. San Antonio, TX: Harcourt Brace & Company, The Psychological Corporation. Assignment: Evaluating Assessment Instruments

 

*** Copyright © 2014. The Board of Regents of the University of Nebraska and the Buros Center for Testing. All rights reserved. Any unauthorized use is strictly prohibited. Buros Center for Testing, Buros Institute, Mental Measurements Yearbook, and Tests in Print are all trademarks of the Board of Regents of the University of Nebraska and may not be used without express written consent.

 

 

American Accent

 

Choose Language