Journal of Abnormal Psychology Copyright 1987 by the American Psychological Association, Inc.

1987, Vol. 96, No. 3,179-183 0021-843X/87/S00.75

Differentiating Anxiety and Depression: A Test of the Cognitive

Content-Specificity Hypothesis
Aaron T. Beck and Gary Brown Robert A. Steer
Center for Cognitive Therapy Department of Psychiatry
Department of Psychiatry University of Medicine and Dentistry of New Jersey-
University of Pennsylvania Medical School School of Osteopathic Medicine
Judy I. Eidelson John H. Riskind
Philadelphia, Pennsylvania George Mason University

The development and initial psychometric properties of the Cognition Checklist (CCL), a scale to
measure the frequency of automatic thoughts relevant to anxiety and depression, are described in
this article. Item analyses of the responses of 618 psychiatric outpatients identified a 14-item depres-
sion and a 12-item anxiety subscale that were significantly related, respectively, to the revised Hamil-
ton Rating Scales for Depression and Anxiety. Patients diagnosed according to the Diagnostic and
Statistical Manual of Mental Disorders (DSM-III; American Psychiatric Association, 1980) with
anxiety disorders had higher mean CCL anxiety scores than patients with DSM-III depression disor-
ders who, in turn, had higher mean CCL depression scores. The validity of the CCL supports the
content-specificity hypothesis of the cognitive model of psychopathology (Beck, 1976).

The cognitive model of psychopathology stipulates that each Method

neurotic disorder can be characterized by a cognitive content
specific to that disorder (Beck, 1976). Thus the transient auto- Patients
matic thoughts, interpretations, and imagery of the depressed A total of 618 consecutive outpatients received the CCL prior to ad-
patient center around the theme of self-depreciation and nega- mission to the Center for Cognitive Therapy, Philadelphia, as part of a
tive attitudes toward the past and future. Anxiety disorders, on standard pretreatment evaluation. A subsample of 210 patients was set
the other hand, are characterized by the theme of danger; anx- aside for cross-validation analyses, leaving 408 patients in the index
ious patients tend to misread their experiences as constituting sample. The overall sample consisted of 278 (45%) men and 340 (55%)
either a physical or psychosocial threat and to overestimate both women. The mean age was 36.20 years (SD = 12.16 years).
the probability and intensity of anticipated harm in future situ-
ations (Beck & Emery, 1985). Automatic thoughts are denned Instruments
as nonvolitional, stream-of-consciousness cognitions (Beck,
Diagnostic interview. A clinician diagnosed each patient according
1976). to the Diagnostic and Statistical Manual of Mental Disorders (DSM-
This study describes the development and psychometric III; American Psychiatric Association, 1980). For a significant portion
properties of the Cognition Checklist (CCL), which was de- of these patients (approximately 415, or 68%) the Structured Clinical
signed to measure the frequency of automatic thoughts. Previ- Interview for DSM-III (SCID; Spitzer & Williams, 1985) was used. All
ous scales of this type, such as the Automatic Thoughts Ques- of the patients in the cross-validation sample were diagnosed with the
tionnaire (Hollon & Kendall, 1980) and the Crandell Cogni- SCID. The SCID provides a standardized format for questioning pa-
tions Inventory (Crandell & Chambless, 1981), have con- tients about their symptoms, and the sequence of questions approxi-
centrated exclusively on the loss and failure cognitions charac- mates the DSM-III decision rules. DSM-III criteria are embedded di-
teristic of depression. In contrast, the CCL includes cognitions rectly in the SCID interview, thus ensuring adequate coverage of the
related to danger, which are said to be characteristic of anxiety relevant criteria. Evidence for the reliability of SCID-based diagnoses
on a portion of the present sample is provided by Riskind, Beck, Ber-
disorders (Beck & Emery, 1985), in addition to cognitions re- chick, Brown, and Steer (in press) who found kappa coefficients of .72
lated to depression. Consequently, the CCL can be used to test for major depression and .79 for generalized anxiety disorder.
explicitly the content-specificity hypothesis of the cognitive In the index sample (N = 408), 99 patients (24%) were given a primary
model (Beck, 1967,1976). diagnosis of a DSM-III anxiety disorder, 207 patients (51%) were given
a primary diagnosis of a depression disorder, and 102 (25%) were not
given a depression or an anxiety disorder as a diagnosis. In the cross-
validation sample (N = 210), 75 (35.7%) of the patients were given pri-
This study was supported by National Institute of Mental Health mary anxiety diagnoses, 96 (45.7%) were given primary depression di-
Grant MH38843 to Aaron T. Beck. agnoses, and the remaining 39 patients (18.6%) received diagnoses other
We gratefully acknowledge the contributions of Paul McDermott and than anxiety or depression.
Auke Tellegen. Hamilton psychiatric rating scales for depression and anxiety. The
Correspondence concerning this article should be addressed to Aaron clinician also rated each patient on the Hamilton Psychiatric Rating
T. Beck, Center for Cognitive Therapy, Room 602, 133 South 36th Scales for Anxiety (Hamilton, 1959) and Depression (Hamilton, 1960).
Street, Philadelphia, Pennsylvania 19104. Because the standard Hamilton scales overlap substantially in content,

179
180 BECK, BROWN, STEER, EIDELSON, RISKIND

they were rescored as suggested by Riskind, Beck, Brown, and Steer (in Table 1
press) to enhance discrimination of anxious and depressive disorders. Cognition Checklist (CCL) Items by Discriminant
Cognition checklist. The initial pool of items for the CCL was com- Function Loading and Labeling of Affect
piled from the verbatim reports of automatic thoughts provided by pa-
tients during the course of treatment with cognitive therapy. These cog- CCL item Loading Labeling Situation
nitions are routinely recorded by cognitive therapy patients on the Daily
Record of Dysfunctional Thoughts (Beck, Rush, Shaw, & Emery, 1979). There's no one left to help me. .28 Depressed" 5
A preliminary 43-item version of the scale was extracted from a pool of I'm worse off than they are. .27 Depressed11 5
nearly 100 potential items by including only cognitions that were judged I'll never be as good as other people
to be most typical of those encountered either in anxious or depressed are. .23 Depressed" 3
patients. In addition, cognitions that were confounded with symptom- I'm falling behind. .23 Anxious 4
atology (e.g., "I have become unable to manage on my own") or that Life isn't worth living. .20 Depressed" 2
reflected a disability typical of a variety of disorders (e.g., "I can't cope") There's no point in trying, I'm sure
were excluded, as were redundant and overlapping cognitions. Respon- to fail. .20 Anxious 3
I don't deserve to be loved. .18 Depressed" 5
dents rated how often each thought typically occurred to them on a 5-
He(she) won't want to see me again. .18 — 5
point scale ranging from 0 (never) to 4 (always) in the context of one of Nothing ever works out for me
four specific situations (attending a social occasion, with a friend, work- anymore. .15 Depressed" 2
ing on a project, and experiencing pain or physical discomfort) and re- I won't know what to say. .15 Anxious 3
gardless of the situation. In addition to rating the frequency of each People don't respect me anymore. .15 Depressed" 5
cognition, the first 212 patients were asked to label the predominant I'll never be as capable as I should
affect they experienced while thinking each of the thoughts contained be. .15 — 4
in the CCL. I'm not worthy of other people's
attention or affection. .14 Depressed" 5
I will never overcome my problems. .14 Depressed" 2
I won't have enough time to do a
Results good job. .13 Anxious 4
I have become physically
Cognition-Affect Consistency Analysis unattractive. .12 Depressed" 2
I'm worthless. .12 Depressed" 3
Labelings of predominant affect supplied by the respondent I'm a social failure. .11 Depressed" 3
that were associated with nonzero ratings were categorized as I've lost the only friends I've had. .11 Depressed" 5
Other things might get in the way. .09 Anxious 4
depressed ("depressed," "sad," etc.), anxious ("anxious," "ner- No one cares whether I live or die. .09 Depressed" 5
vous," etc.), angry ("angry," "mad," etc.), or other. Thirty-five I will hurt someone I care about. .08 2
.07 —
of the items were labeled as expected: Items reflecting themes What if I fail? Anxious 4
of hopelessness and loss were most often labeled as depressed, He(she) will reject me. .06 — 5
People will keep me from getting
whereas items describing themes of danger and threat of loss what I want. .05 2
were most often labeled as anxious. Only two items were not .04 —
He(she) will be irritated with me. Anxious 5
labeled as expected, and six items were labeled as neither anx- I'm losing my mind. .04 Anxious" 2
ious nor depressed (see Table 1). They won't be there when I need
them. .03 5
— 4
I might make a mistake. .03 Anxious
Item Selection I will make a fool out of myself. .01 Anxious 3
I am a defective human being. -.01 Depressed 1
People will laugh at me. -.02 Anxious" 3
For the following item-selection analyses, the anxious group Something might happen that will
was denned as those patients with revised Hamilton Anxiety ruin my appearance. -.04 Anxious" 2
Rating Scale (HARS-R) scores greater than or equal to 0.5 stan- There's something very wrong with
dard deviations above their revised Hamilton Rating Scale for me. -.05 Anxious" 1
I'm going to have an accident. -.05 Anxious" 1
Depression (HRSD-R) scores. Similarly, the depressed group was Something awful is going to
defined as those patients with HRSD-R scores greater than or happen. -.08 Anxious" 2
equal to 0.5 standard deviations above their HARS-R scores. I am going to be injured. -.09 Anxious" 1
The total number of patients meeting either of the aforemen- Something will happen to someone
tioned criteria in the index sample was 202, 106 (52.5%) in the I care about. -.10 Anxious" 2
I might be trapped. -.12 Anxious" 1
anxious group and 96 (47.5%) in the depressed group; 206 met I am not a healthy person. -.14 Anxious" 1
neither of the criteria and so were not included in the initial set What if no one reaches me in time
of item-selection analyses. to help? -.14 Anxious* 1
The 43 initial CCL items were entered into the SPSS" Dis- What if I get sick and become an
invalid? -.20 Anxious" 1
criminant program (SPSS, 1983). The classification variable I am going to have a heart attack. -.30 Anxious" 1
was coded 1 for anxiety and 2 for depression. The resulting dis-
criminant function loadings are shown in Table 1. Sixteen Note. For situations; 1 = feeling pain or physical discomfort; 2 = regard-
(76.2%) of the 21 highest loading positive items had depressed less of the situation; 3 = attending a social occasion; 4 = working on a
content and 12 (92.3%) of the 13 items with negative loadings project; 5 = with a friend. TV = 202 for discriminant analysis; N = 212
had anxious content. for labeling of affect. Values have been rounded to the nearest hun-
dredth.
Items were next assigned tentatively to anxiety and depres- " Included on tentative Anxiety scale.
sion subscales if the direction of the discriminant loading was " Included on tentative Depression scale.
 DIFFERENTIATING 181

Table 2
Varimax- Rotated Principal-Factor Loadings for Cognition Checklist (CCL) Items
CCL item Depression Anxiety Communality Situation
I'm worthless. .77 .63 1
I'm not worthy of other people's attention or affection. .73 .57 2
I'll never be as good as other people are. .73 .56 3
I'm a social failure. .71 .51 3
I don't deserve to be loved. .69 .49 2
People don't respect me anymore. .66 .48 2
I will never overcome my problems. .63 .36 .52 1
I've lost the only friends I've had. .62 .40 2
Life isn't worth living. .61 .31 .47 1
I'm worse off than they are. .61 .45 2
There's no one left to help me. .60 .32 .46 1
No one cares whether I live or die. .60 .41 2
Nothing ever works out for me anymore. .56 .35 .44 1
I have become physically unattractive. .50 .35 .37 1
What if I get sick and become an invalid? .71 .49 4
I am going to be injured. .70 .51 4
What if no one reaches me in time to help? .70 .50 4
I might be trapped. .66 .48 4
I am not a healthy person. .65 .49 4
I'm going to have an accident. .63 .44 1
There's something very wrong with me. .34 .62 .50 1
Something might happen that will ruin my appearance. .56 .35 4
I am going to have a heart attack. .56 .32 1
Something awful is going to happen. .36 .53 .42 1
Something will happen to someone I care about. .51 .35 1
I'm losing my mind. .36 .51 .39 1
% total variance .80 .20
% common variance .55 .45
Note. For situations, 1 = regardless of the situation; 2 = with a friend; 3 = attending a social occasion; 4 = feeling pain or physical discomfort. N =
408. Values have been rounded to the nearest hundredth. Loadings less than .30 are not shown.

consistent with the labeling of affect. The disposition of each Reliability

item is shown in Table 1.
Cronbach coefficient alphas were calculated in the cross-vali-
dation sample for both CCL subscales to estimate their internal
Treatment of Marginal Items
consistency. The alpha coefficient for CCL-A (Anxiety) was .90,
We constructed tentative subscales by summing the unit- and the average corrected item-total correlation was .62. The
weighted 14 depression and 12 anxiety items that met both the overall alpha coefficient for CCL-D (Depression) was .92, and
labeling-of-affect and discriminant-loading criteria. Partial cor- the average corrected item-total correlation was .65.
relations were calculated between each of the unassigned items To estimate test-retest reliability, the CCL was readminis-
that had an absolute discriminant function loading less than .10 tered to a subsample of 66 patients during their sixth week of
and each subscale controlling for the correlation of the remain- treatment. The correlation between intake and 6-week CCL-A
ing subscale. One item, "I'm losing my mind," which loaded in scores was .79 (p< .001); the correlation between intake and 6-
the depressed direction on the discriminant function but was week CCL-D scores was .76 (p < .001). The test-retest reliabili-
labeled as anxious in the cognition-affect consistency analysis, ties of the subscales did not differ.
had a higher partial correlation with the tentative anxiety scale
than with the depression scale; it was therefore added to the Discriminant Validity
anxiety scale.
To confirm that dimensions of anxious and depressed cogni- We calculated simple and partial correlations between each
tion were underlying the 13 anxiety and 14 depression items, CCL subscale and each revised Hamilton scale, controlling, re-
we conducted a principal-factor analysis with iterations to es- spectively, for the correlation with the remaining Hamilton
tablish communalities on the entire index sample (N = 408), scale. Because we had used the Hamilton scales in the item-
and two factors were retained for rotation. After a varimax rota- selection process, this analysis was cross-validated in a separate
tion, one item ("People will laugh at me") that had been as- sample (N= 210).
signed to the anxiety subscale now loaded on the depression The intercorrelation of the CCL subscales was .58 in the in-
factor and was therefore removed from the anxiety scale. The dex sample and .57 in the cross-validation sample. Although
analysis was repeated without this item. The resulting factor both the CCL-D and the CCL-A correlated significantly with
pattern of the final 12 anxiety subscale (CCL-A) and 14 depres- both revised Hamilton scales, the relationship was stronger be-
sion subscale (CCL-D) items is shown in Table 2. tween the CCL subscale and the same-affect Hamilton scale
182 BECK, BROWN, STEER, EIDELSON, RISKIND

Table 3 had higher mean CCL-A scores than did depressed patients in
Simple and Partial Correlations of Cognition Checklist both the index study and cross-validation study, whereas de-
(CCL) Subscales With Revised Hamilton Scales pressed patients had higher mean CCL-D scores than did anx-
ious patients.
Cross-validation sample To determine the accuracy with which patients could be as-
Index sample (JV= 408) (N =210)
signed to their correct diagnostic group on the basis of CCL
CCL scale HRSD-R HARS-R t HRSD-R HARS-R t subscale scores, we performed a discriminant classification
analysis. By applying the discriminant function derived on the
Depression index sample to the cross-validation sample, 30 of 38 (79%, or
r .56** .38** 4.79** .62** .37** 6.89** 59% above chance) anxious patients and 34 of 41 (83%, or 65%
Partial r .45** .08 .54** .03
Anxiety above chance) depressed patients were correctly classified.
r .43** .55** .36* .54** 3.29**
Partial r .16* .41** 3.20* .08 .44**
Discussion
Note. HRSD-R = Hamilton Depression Scale-Revised. HARS-R = Ham-
ilton Anxiety Scale-Revised. The present set of results supports the content-specificity hy-
*p<.01. **/><.001. pothesis of the cognitive model that anxious and depressed
groups could be distinguished by the types of cognitive content
intrinsic to the two conditions. It is clear that the items retained
(Table 3). The differences in magnitude of the correlations be- on the anxiety and depression subscales of the CCL are consis-
tween same- and different-affect scales were all significant be- tent in content with the cognitive themes ascribed to them by
yond the .05 level using Hotelling's t test. When partial corre- the cognitive model of psychopathology (Beck, 1976).
lations were calculated between each CCL and each revised The content of the subscales also conforms to a broader
Hamilton scale controlling, in turn, for the remaining revised framework of affective thought processes recently proposed by
Hamilton scale, all of the correlations between each CCL sub- Tellegen (1985). Tellegen proposed that depressive states are
scale and the same-affect Hamilton scale in both samples re- characterized by affective disengagement and that the associ-
mained significant. All opposite-affect correlations were not sig- ated cognitions are indicative of an "oriented" or "knowing"
nificant, with the exception of the partial correlation of the mode; in contrast, affectively engaged states such as anxiety re-
CCL-A with the HRSD-R in the index sample, controlling for flect an "orienting" or "asking" mode. Thus the anxiety cogni-
HARS-R (partial r = . 16, p< .05). tions on the CCL embody a greater degree of uncertainty and
Next, each sample was regrouped according to both DSM- an orientation toward the future, whereas depressive cognitions
III diagnosis and salience of affect. Patients were included in are either oriented toward the past or reflect a more absolute
the second anxiety grouping if they had a primary diagnosis of negative attitude toward the future.
a DSM-III anxiety disorder (generalized anxiety disorder, panic The CCL items were subjected to a variety of statistical pro-
disorder, social phobia, etc.) and their HARS-R standard score (z cedures. Multivariate analyses using both internal and external
score) was at least 0.5 standard deviations higher than their HR- criteria yielded a 12-item subscale of anxious cognitions and
SD-R standard score. Likewise, patients were included in the de- a 14-item subscale of depressed cognitions. Evidence for the
pression grouping if they had a primary diagnosis of a DSM- discriminant and convergent validity of the two measures was
III depression disorder (major depression, dysthymic disorder, demonstrated by (a) correlations with a set of independent rat-
etc.) and their HRSD-R standard score (z score) was at least 0.5 ings of anxiety (the HARS-R) and depression (HRSD-R); (b) the
higher than their HARS-R standard score. mean scores of the two scales, which differentiated samples of
The mean CCL subscale scores for the criterion groups are patients diagnosed with anxious and depressed DSM-III disor-
shown in Table 4. rtests indicated that the CCL subscale scores ders; and (c) a good classification rate of patients into their cor-
differentiated the groups in both the index and the cross-valida- rect DSM-III diagnostic category on the basis of their CCL sub-
tion studies (all one-tailed ps < .025). Thus anxious patients scale scores.

Table 4
Means of Cognition Checklist (CCL) Subscales for DSM-IIIAnxiety and Depression Diagnostic Groups
CCL-Anxiety CCL-Depression
Group N M SD / M SD t
Index sample
Anxiety 41 52.28 7.84 . „„ 47.93 9.72
Depression 71 49.14 8.54 3
53.46 8.50
3.14**
Cross-validation sample
Anxiety 38 52.97 11.21 2 52,». 45.02 7.71
Depression 41 47.56 6.50 ^ 54.78 8.98 5.05***

Note. Scores have been converted to /"scores [(z score X 10) + 50].
*p< .025. **p< .005. ***/>< .001.
 DIFFERENTIATING 183

Although the correlation between the subscales was substan- nitions Inventory. Paper presented at the meeting of the Association
tial, they afforded moderate discrimination between the crite- for Advancement of Behavioral Therapy, Toronto, Ontario, Canada.
rion groups on the basis of mean scores and good above-chance Dobson, K. S. (1985). Relationship between anxiety and depression.
classification rates. The discrimination achieved with the CCL Clinical Psychology Review, 5, 307-324.
Hamilton, M. (1959). The assessment of anxiety states by rating. British
is comparable to the best results that have been obtained with
Journal of Medical Psychology, 32, 50-55.
symptom-based psychometric measures of anxiety and depres- Hamilton, M. (1960). A rating scale for depression. Journal of Neurol-
sion, in which a high degree of overlap is commonly found (see ogy, Neurosurgery, and Psychiatry, 23, 56-61.
Dobson, 1985, for a review). It is hoped that the CCL will be Hollon, S. D., & Kendall, P. C. (1980). Cognitive self-statements in de-
used in conjunction with symptom-based measures of psycho- pression: Development of an automatic thoughts questionnaire. Cog-
pathology to afford enhanced discrimination of the two syn- nitive Therapy and Research, 4, 383-395.
dromes. In addition, the CCL would have utility in a variety of Riskind, J. H., Beck, A. T., Berchick, R. J., Brown, G., & Steer, R. A.
studies relating cognitive factors to diagnostic groups as well as (in press). Interrater reliability of the Structured Clinical Interview
in process studies of psychotherapy. for DSM-III (SCID) for major depression and generalized anxiety dis-
order. Archives of General Psychiatry.
Riskind, J. H,, Beck, A. T, Brown, G., & Steer, R. A. (in press). Taking
References the measure of anxiety and depression: Validity of reconstructed
Hamilton Scales. Journal of Nervous and Mental Disease.
American Psychiatric Association. (1980). Diagnostic and statistical Spitzer, R. L., & Williams, J. B. W. (1985). Instruction manual for the
manual of mental disorders (3rd ed.). Washington, DC: Author. Structured Clinical Interview for DSM-III (SCID). New York: Bio-
Beck, A. T. (1967). Depression: Clinical, experimental, and theoretical metrics Research Department, New \fork State Psychiatric Institute.
aspects. New 'York: Hoeber. SPSS, Inc. (1983). SPSS* user's guide. New York: McGraw-Hill.
Beck, A. T. (1976). Cognitive therapy and the emotional disorders. New Tellegen, A. (1985). Structures of mood and personality and their rele-
York: International Universities Press. vance to assessing anxiety, with an emphasis on self-report. In A. H.
Beck, A. T, & Emery, G. (1985). Anxiety disorders and phobias: A cog- Tuma & J. D. Maser, (Eds.), Anxiety and the anxiety disorders (pp.
nitive perspective. New York: Basic Books. 681-706). Hillsdale, NJ: Erlbaum.
Beck, A. T, Rush, A. J., Shaw, B. E, & Emery, G. (1979). Cognitive
therapy of depression. New \brk: Guilford Press. Received August 18,1986
Crandell, C. J., & Chambless, D. L. (1981, November). The validation Revision received February 16,1987
of an inventory for measuring depressive thoughts: The Crandell Cog- Accepted February 27,1987 •