Cognitive Therapy and Research, VoL 4, No. 4, 1980, pp.

383-395

Cognitive Self-Statements in Depression:

Development of an Automatic Thoughts
Questionnaire 1
Steven D. Hollon 2 and Philip C. Kendall

University of Minnesota

A 30-item questionnaire was devised to measure the frequency of

occurrence of automatic negative thoughts (negative self-statements)
associated with depression. Male and female undergraduates were asked to
recall dysphoric experiences and to report associated cognitions. One
hundred representative cognitions were selected and administered to a
second sample, along with the MMPI D scale and the Beck Depression
Inventory. Thirty items discriminating between criterion groups of
psychometrically depressed and nondepressed subjects were identified. The
resultant 30-item automatic thoughts questionnaire (ATQ-30) was
cross-validated and found to significantly discriminate psychometrically
depressed from nondepressed criterion groups. No differences were found
between males and females on the measure. Factor analysts indicated a
four-factor solution, with a large first factor reflecting Personal Maladjust-
ment, a second factor indicative o f Negative Self-Concept and Negative
Expectations, and two lesser factors. The ATQ-30 may provide a means of
testing basic theory relating cognitive content to behavioral and affective
processes and assessing change in cognitions associated with experimental
manipulation or psychotherapeutic intervention.

~Preparation of this article was supported in part by two University of Minnesota Graduate
School Grants in Aid of Research (492-0325-4909-02; 440-0160-4909-02) to the first and
second authors, respectively. We wish to thank Auke Tellegen for his invaluable comments
and suggestions, and Cliff Johnson for his reliable assistance.
~Address all correspondence to Steven D. Hollon or Philip C. Kendall, Cognitive Assessment
Project, Department of Psychology, University of Minnesota, 75 East River Road, Minnea-
polis, Minnesota 55455. An expanded version of this manuscript, complete with correlation
matrix, is available from the authors.
383
0147-5916/80/1200-0383503.00/0 © 1980 Plenum Publishing Corporation
384 Hollon and Kendall

Recent theory and research has focused increasingly on the role of cognition
in the etiology, maintenance, and treatment of various types of psycho-
pathology (Bandura, 1977; Beck, 1976; Ellis & Grieger, 1977; Goldfried &
Davison, 1976; Kendall & Hollon, 1979; Mahoney, 1974, 1977;
Meichenbaum, 1977). With regard to depression, Beck ha s argued that the
various components of the clinical syndrome are the consequence of
pervasive, systematically negative distortions in both cognitive content and
process (Beck, 1963, 1964, 1967).
Emphasis on cognitive factors has led to the development of
therapeutic procedures designed to reduce depression by systematically
altering these negative beliefs (Beck, Rush, Shaw, & Emery, 1979; Hollon &
Beck, 1979). Outcome studies have indicated that such procedures are
superior to either traditional or strictly behavioral interventions in terms of
symptom reduction (Shaw, 1977; Taylor & Marshall, 1977) and superior to
tricyclic pharmacotherapy in terms of both symptom reduction and
prevention of relapse (Rush, Beck, Kovacs, & Hollon, 1977).
Despite the current interest in both cognitive theory and cog-
nitive-behavioral therapy in the affective disorders, there appears to
have been little systematic effort to assess changes in cognitive content and
process as a function of treatment (Kendall & Korgeski, 1979). One reason
for this failure to provide a "confirmation of treatment mechanisms" is the
absence of suitable specific measures of cognitions associated with
depression.
Jones's Irrational Beliefs Test (IBT) (1968) is one measure of those
beliefs presumed to underly neurotic emotionality. However, while the IBT
predicts affect-related arousal to specific situational stimuli (Goldfried &
Sobocinski, 1975), it correlates only moderately with self-reported
depression (Nelson, 1977), assesses beliefs associated with general
emotionality rather than specific affective states, and assesses the degree to
which people ascribe to the validity of various statements, not the frequency
with which they experience specific cognitions. EUis's concept of "irrational
beliefs" (1962) may well represent accurate summaries of what people act as
if they believe, but may be quite distinct from any cognitions that
individuals spontaneously experience. Measures such as the IBT may reflect
attitudes that people hold but rarely think about, at least in the format
presented.
The present paper reports on the development of any inventory
designed to identify the covert self-statements (see Kendall & Hollon, 1981)
reported by depressives as being representative of the kinds of cognitions
they experience. Such an inventory should prove of use both in theory
testing and as an independent measure of cognitive change associated with
either laboratory manipulations or clinical interventions.
Automatic Thoughts Questionnaire 385

METHOD

Item Generation

An initial pool of items was generated by asking 788 male and female
undergraduate students to recall an experience in their lives that they had
found to be depressing. Subjects were asked to recreate the situation in
memory as vividly as they could, as if they were reexperiencing the situation
at that moment. Subjects were then instructed to record whatever thoughts
had "popped into their head" in that situation. Subjects were asked to
record their cognitions exactly the way they thought them, regardless of
grammar or syntax.
The authors screened the pool of reported cognitions to eliminate
redundant or incomprehensible responses. A total of 100 reported self-
statements were selected for subsequent use, forming the initial automatic
thoughts questionnaire (ATQ-100).

Subjects

A total of 348 male and female undergraduate college students were

recruited to participate in item selection and cross-validation. All subjects
received extra credit points toward course grades in return for participation.
Protocols from 36 subjects who failed to complete the full test battery
were excluded from all further analyses, leaving a sample of 312 (167 males
and 145 females). The mean age of the full sample was 20.22 (SD = 4.34).
Subjects were randomly divided into two subsamples of 156 each. The
first subsample was used for item selection, while the second subsample was
used for subsequent scale cross-validation. The item selection sample was
composed of 85 males and 71 females and the cross-validation sample
consisted of 82 males and 74 females. These two subsamples did not differ
significantly in terms of sexual composition, X2(1) = .05, or age, t(310) =
.69.

Measures of Depression and Anxiety

Two self-report inventories, recently recommended by Rehm (1976),

were used to assess levels of depression. The Beck Depession Inventory
(BDI) is a 21-item inventory designed to assess affective, behavioral,
cognitive, motivational, and vegetative aspects of depression (Beck, Ward,
Mendelson, Mock, & Erbaugh, 1961; Beck, 1967). Each item consists of
four statements, scored on a range from 0 to 3. Subjects are instructed to
386 Hollon and Kendall

complete each item in terms o f how they felt over the preceding week. A
total score, ranging from 0 to 63, is obtained by summing over the items.
The greater the score, the greater the severity of syndrome depression. The
BDI has shown good concurrent validity when compared to psychiatric
ratings o f severity of depression in both clinical populations (r = .79, N =
226 and r = .67, N = 183) (Beck et al., 1961) and college student
populations (r = .79, N = 56) (Bumberry, Oliver, & McClure, 1978). Beck
recommends a cutting score o f 10 and above for designating a subject as
depressed.
The Minnesota Multiphasic Personality Inventory Depression scale
(MMPI-D) consists o f 60 true-false items, 49 initially selected to
discriminate normals from hospitalized manic-depressives and 11 initially
selected to distinguish between depressed and other psychopathological
conditions (Hathaway & McKinley, 1940). Total scores can range from 0 to
60, with high scores reflecting greater severity of depression. The MMPI-D
has been widely used as a screening instrument for selecting depressed
samples for research purposes.
It should be noted that both measures provide estimates o f the
severity o f syndrome depression. Depression can be measured as a symptom
(generally defined as sadness or dysphoric mood), a syndrome (consisting o f
covarying signs and symptoms), or a discrete nosological category (Beck,
1967). With the exception o f self-reported anxiety, no attempt was made to
evaluate the presence or absence of associated indices o f psychopathology
for the sample.
The State-Trait Anxiety Inventory (STAI) A-Trait scale (Spielberger,
Gorsuch, & Lushene, 1970) is a 20-item inventory asking subjects to report
how frequently they had experienced various phenomena related to trait
anxiety. Total possible scores range from 20 to 80, with higher scores
indicating greater trait anxiety. A measure of trait anxiety was included in
an effort to check on the specificity of negative cognitions to depressive
affect.

Procedure

All 348 subjects were administered the tests described above as part o f
a larger battery. For the ATQ, subjects were instructed as follows.

Listed beloware a varietyof thoughts that pop into people's heads. Please read each
thought and indicate how frequently, if at all, the thought occurred to you over the
last week. Please read each item carefully and flu in the appropriate circle on the
answer sheet in the following fashion (1 = "not at all," 2 = "sometimes," 3 =
"moderately often," 4 = "often," and 5 = "all the time").

Subjects were instructed to complete the other tests according to standard

instructions.
Automalic Thoughts Questionnaire 387

Method of A nalysis

Depressed and nondepressed criterion groups were formed with each

subsample on the basis of scores on both depression measures. Subjects
scoring 1 standard deviation or more above the mean on both the BDI
(score of 11 and above) and the MMPI-D (score of 26 and above) were
defined as depressed. Subjects scoring at or near the mean on both scales
(BDI scores equal to 3, 4, or 5 and MMPI-D scores from 15 through 20)
were defined as nondepressed.
In the item selection sample, 12 subjects (7 males and 5 females) met
the criteria for inclusion in the depressed group, while 20 subjects (9 males
and 11 females) met the criteria for inclusion in the nondepressed group. In
the cross-validation sample, 14 subjects (5 males and 9 females) and 21
subjects (12 males and 9 females) met the criteria for the depressed and
nondepressed groups, respectively. There were no significant gender
differences among the four groups, ~(2(3) = 1.97.
Independent t tests were computed on each of the ATQ-100 items
between depressed and nondepressed subjects for the item selection sample.
Items discriminating between the two groups were then combined into a
single scale and cross-validated on the cross-validation sample-depressed
and cross-validation sample-nondepressed groups.

RESULTS

Sample Characteristics and Subsample Comparability

One-way analyses of variance evidenced no significant differences

between the two subsamples on the BDI, the MMPI-D, the STAI A-Trait,
and the ATQ-30, all F's(1,308)< 1. Males and females did not differ on
three of the four measures: BDI, STAI A-Trait, and ATQ-30, all
F's(1,308)< .05. Females did score significantly higher than males on the
MMPI-D scale, F(1,308) -- 5.06, p < .05. There were also no significant
Subsample X Sex interactions on all variables, all F's(1,308)< 1. It appears
that the two subsamples were comparable.

Characteristics of Depressed and Nondepressed Criterion Groups

As expected, depressed subjects had significantly higher mean scores

than nondepressed subjects on both the BDI, F(1,63) = 11.00, p < .001,
and the MMPI-D, F(1,63) = 126.47, p < .001. Males did not differ from
females on either the BDI or the MMPI, all F's(1,63) < 1, nor were there any
significant interactions between sex and criterion group on the depression
388 Hoilon and Kendall

©
Automatic ThoughtsQuestionnaire 389

~ ~ . ~ .~ • . .

.g

C~

0
z
390 Hoilon and Kendall

measures, all F's < 1, respectively. Mean BDI scores for both depressed cri-
terion groups were comparable to the Bumberry et al. (1978) mildly to
moderately depressed college students. In that study, BDI scores were
significantly correlated with psychiatric ratings of depression. Thus the
depressed criterion groups can be said to represent a mixed sex sample of
psychometrically identified mildly to moderately depressed college students.

A TQ Item Selection

Using the item selection sample data, 30 of the original 100 ATQ
items were found to significantly discriminate between the depressed and
nondepressed criterion groups at the .01 level. The 30 items, with means
(and standard deviations) and t-test values are presented in Table I. On each
of the 30 automatic thoughts, the depressed subjects reported significantly
more frequent instances of the specific cognitions. None of the remaining
items were endorsed significantly more often by the nondepressed group
than by the depressed group.
Although the 30 items were selected because they significantly
discriminated depressed from nondepressed subjects, the items were not
differentially endorsed by males and females. An analysis of variance
indicated that, while depressed and nondepressed group did differ on total
ATQ-30 score, F(1,27) = 43.48, p < .001, there was no significant effect for
sex, F(1,27)< 1, and no significant Depression X Sex interaction,
F(1,27) < 1.

Cross- Validation

Total scores on the ATQ-30 were computed for all subjects in the
depressed and nondepressed criterion group from the cross-valida-
tion sample. An independent t test indicated significantly higher scores,
t(17) = 4.85, p < .001, for the depressed subjects. The mean ATQ-30 for
the depressed subjects was 79.64 (SD = 22.29), while the mean ATQ-30 for
the nondepessed subjects was 48.57 (SD = 10.89). These data cross-validate
the finding that these 30 items are endorsed more frequently by depressed
than by nondepressed subjects. Differences in scores on the ATQ-30 in the
cross-validation sample are not inflated by capitalization on chance factors
in item selection. The means (and standard deviations) and t-test values for
the cross-validation of each item are presented in Table I.

Interrelationships Among Measures

All correlations among the BDI, the MMPI-D, and the ATQ-30 (as
well as the 100-item ATQ) using the full sample were statistically
Automatic Thoughts Questionnaire 391

significant, p ' s < .01. Intercorrelations calculated separately from the item
selection sample and cross-validation sample data were also significant. Of
note were the comparabilities of ATQ-30/BDI and MMPI-D correlations
(r's range from .45 to .70) across both samples (r's within .02 of each other)
and the relative lack of differences in the magnitude of the correlations
involving the ATQ-30 and the 100-item ATQ (largest difference between r's
was .05). The comparability of the interrelationships speaks to the stability
of the endorsement of automatic negative cognitions, while the similarity of
the ATQ-30 and ATQ-100 indicates a retention of concurrent validity
despite the reduction in the total number of items.

Anxiety, Depression, and the A TQ-30

The STAI A-Trait had been administered to provide a basis for

selecting ATQ cognitions associated with depression but not associated with
anxiety. However, the correlation between trait anxiety and the ATQ-30 was
actually quite high (r = .79, for the full sample), precluding any
interpretation of specificity to depression-related cognitions. It is not clear,
however, whether this lack of specificity reflects flaws in the ATQ-30 or
overlap in the measures of syndrome psychopathology used. Correlations
between the established depression measures and the anxiety measure were
also high (r's ranged from .60 to .78), versus an r of .63 between the two
depression measures, suggesting that it was not possible to meaningfully
discriminate between self-reported depression and self-reported anxiety in
this sample.

Reliability

Both split-half and coefficient alpha (Kuder & Richardson, 1937)

reliability coefficients were used to estimate the reliability of the ATQ-30.
The split-half reliability coefficient, calculated on odd versus even items,
was .97, p < .001. Calculation of coefficient alpha yielded a correlation of
.96, p < .001. In general, both reliability estimates arequite satisfactory,
and the ATQ-30 appears to possess sufficient reliability to justify further
use.
There are theoretical grounds for not collecting test-retest data on the
ATQ-30 without simultaneously collecting comparable data on self-re-
ported levels of syndrome depression. Without specifying the direction of
causality, the frequency of specific depression-relevant cognitions would be
expected to be closely related to the level of experienced depression.
Depression itself is a transitory phenomenon; thus test-retest coefficients on
the ATQ-30 should be evaluated only when accompanied by reliability
coefficients on syndrome depression measures. Such data are currently
being collected and will be reported in a later paper.
392 Hollon and Kendall

Table II. ATQ-30 Scale Items with Factor Loadings Over ± .50 a

ATQ-30 scale Factors loadings

item number Items I II III IV
Factor I
26 Something has to change. .64 .23 .20 .30
20 What's the matter with me? .63 .39 .23 .19
7 I wish I were a better person. .62 .26 .16 .13
14 Whats wrong with me? .62 .29 .31 .29
10 I'm so disappointed in myself. .57 .22 .24 .17
Factor II
28 My future is bleak. .19 .70 .13 .35
23 I'm a failure. .23 .68 .33 .21
24 I'U never make it. .21 .57 .27 .45
9 My life's not going the way
I wanted it to. .49 .52 .27 .28
21 I'm a loser. .24 .52 .46 .18
3 Why can't I ever succeed? .46 .52 .31 .22
2 I'm no good. .38 .52 .45 .03
Factor III
18 I'm worthless. .10 .37 .66 .23
17 I hate myself. .30 .14 .65 .06
Factor IV
30 I can't finish anything. .27 .27 .11 .73
29 It's just not worth it. .17 .23 .43 .58

aATQ-30 = Automatic Thoughts Questionnaire.

1tern A n a l y s i s

All A T Q - 3 0 i t e m - t o - t o t a l correlations were significant at o r b e y o n d

the .001 level. I n d i v i d u a l i t e m - t o t a l c o r r e l a t i o n s r a n g e f r o m r = .47 ( " I ' v e
let people d o w n " ) to r = .78 ( " M y life's n o t g o i n g the way I w a n t it t o " ) .
These c o n s i s t e n t l y m o d e r a t e t o strong c o r r e l a t i o n s indicate t h a t each item is
s i g n i f i c a n t l y related to the t o t a l score.

Factor Analysis

A p r i n c i p a l - c o m p o n e n t s factor analysis with V a r i m a x r o t a t i o n s was

p e r f o r m e d o n the 312 s u b j e c t s ' responses to the A T Q - 3 0 . Since males a n d
females d i d n o t differ in t e r m s o f total scores i n either s a m p l e , d a t a f r o m
b o t h sexes were i n c l u d e d in the analysis. As s h o w n i n T a b l e II, f o u r factors
emerged with eigenvalues > 1, a c c o u n t i n g for a total o f 58.9o7o o f the
Automatic Thoughts Questionnaire 393

variance. The first factor accounted for the bulk of the variance, 45.9°70,
while factors II, III, and IV accounted for 5.2070, 4.4070, and 3.5070 of the
variance, respectively.
Items loading on the first factor appear to reflect perceptions of
Personal Maladjustment and Desirefor Change. The second factor consists
of items reflecting Negative Self-Concept and Negative Expectations. Items
on the third factor reflect Low Self-Esteem, while items on the fourth factor
reflect Giving Up/Helplessness. Items loading on Factor II seem
particularly consistent with two of the three components of Beck's negative
cognitive triad (Beck, 1963): negative views of the self and negative views of
the future. Similarly, these cognitions appear quite consistent with the
tendency to attribute nonsuccess to internal, global, stable-over-time
factors, a tendency regarded by Abramson, Seligman, and Teasdale (1978)
to be central to depression in their recent reformulation of the learned-
helplesness model of depression.

DISCUSSION

A 30-item inventory that identifies cognitions reported as being

present in association with mild to moderate depression was developed and
cross-validated. The inventory appears to have sufficient internal reliability
and concurrent validity to justify use as a general measure of
depression-related automatic negative thoughts. Specific items, as well as
the total score on the inventory, appear to reliably separate depressed from
nondepressed criterion groups. It appears that the ATQ-30 is a relatively
brief measure of automatic thoughts (self-statements) in depression, one that
applies to either males or females. As with research investigating the self-
statements associated with assertive behavior (Schwartz & Gottman, 1976)
and stressful medical procedures (Kendall, Williams, Pechacek, Graham,
Shisslak, & Herzoff, 1979), the present study supports the use of self-state-
ment inventories to assess certain cognitive events (see also Kendall &
Hollon, 1981).
It remains to be determined whether clinically formed samples of
depressed patients will differ in terms of reported frequency of automatic
negative thoughts on the ATQ-30. The samples utilized in the present study
were composed of subclinical volunteers likely to be somewhat less
depressed, younger, and better educated than many populations of
potential interest. A closely related issue concerns the ubiquitousness of
negative cognitions across subtypes of depression. It is not unreasonable to
speculate that relationships between cognitive content and affective states
may hold for some subtypes of depression, but not all. If so, it is possible
that the nonbipolar rather than bipolar (Depue & Monroe, 1978) and
394 Hollon and Kendall

primary rather than secondary (Robins & Guze, 1972) depressives would
evidence more extreme levels of negative cognitions.
Finally, the present data did not speak to the utility of the ATQ-30 as
a measure sensitive to change. As stated earlier, a major reason for
developing a measure of depression-related self-statements or negative
automatic thoughts was to provide an independent means of assessing the
adequacy of manipulations or therapies targeted at the cognitions of de-
pressives. It currently remains to be demonstrated whether the ATQ-30 is
sensitive to changes in the nature and frequency of cognitions that result
from experimental manipulation and/or psychotherapeutic interventions.
At this time it appears that the ATQ-30 possesses good concurrent
validity as a measure of the automatic negative thoughts presumed by
cognitive theories to be related to the state of depression. Additional work
needs to be done addressing further issues of reliability (e.g., test-retest),
sensitivity to change, and applicability to clinical populations, but the
ATQ-30 appears to be a promising tool for assessing cognitions associated
with depression.

REFERENCES

Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. D. Learned helplesness in humans:

Critique and reformulation. Journal of Abnormal Psychology, 1978, 87, 49-74.
Bandura, A. Self-efficacy: Toward a unifying theory of behavioral change. Psychological Re-
view, 1977, 84, 191-215.
Beck, A. T. Thinking and depression: I. Idiosyncratic content and cognitive distortions.
Archives o f General Psychiatry, 1963, 9, 324-333.
Beck, A. T. Thinking and depression: II. Theory and therapy. Archives o f General Psychiatry,
1964, 10, 561-571.
Beck, A. T. Depression: Clinical, experimental, and theoretical aspects. New York: Hoeber,
1967.
Beck, A. T. Cognitive therapy and the emotional disorders. New York: International Uni-
versities Press, 1976.
Beck, A. T., Rush, A. J., Shaw., B. F., & Emery, G. Cognitive therapy of depression: A
t r e a t m e n t manual New York: Guilford Press, 1979.
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. An inventory for measur-
ing depression. Archives of General Psychiatry, 1961, 4, 561-571.
Bumberry, W., Oliver, J. M., & McClure, J. N. Validation of the Beck Inventory in a uni-
versity population using psychiatric estimate as the criterion. Journal o f Consulting
and Clinical Psychology, 1978, 46, 150-155.
Depue, R. A., & Monroe, S. M. The unipolar-bipolar distinction in the depressive disorders.
Journal o f Abnormal Psychology, 1978, 87, 3-20.
Ellis, A. E. Reason and emotion in psychotherapy. New York: Lyle Stuart, 1962.
Ellis, A., & Grieger, R. Handbook of rational-emotive therapy. New York. Springer, 1977.
Goldfried, M. R., & Davison, G. C. Clinical behavior therapy. New York: Holt, Rinehart &
Winston, 1976.
Goldfried, M. R., & Sobocinski, D. Effect of irrational beliefs on emotional arousal. Journal
of Consulting and Clinical Psychology, 1975, 43, 504-510.
Automatic Thoughts Questionnaire 395

Hathaway, S. R., & McKinley, J. C. The measurement of symptomatic depression with the
Minnesota Multiphasic Personality Schedule. Psychological Bulletin, 1940, 37, 425.
Hollon, S. D., & Beck, A. T. Cognitive therapy of depression. In P. C. Kendall & S. D.
Hollon (Eds.), Cognitive-behavioral interventions: Theory, research, and procedures.
New York: Academic Press, 1979.
Jones, R. G. A factored measure of Ellis" irrational belief system. Wichita, Kansas: Test
Systems, 1968.
Kendall, P. C., & Hollon, S. D. (Eds.). Cognitive-behavioral interventions: Theory, research,
and procedures. New York: Academic Press, 1979.
Kendall, P. C., & Hollon, S. D. Assessing self-referent speech: Methods in the measurement of
self-statements. In P. C. Kendall and S. D. Hollon (Eds.), Assessment strategies for
cognitive-behavioral intervention. New York: Academic Press, 1981.
Kendall, P. C., & Korgeski, G. P. Assessment and cognitive-behavioral interventions.
Cognitive Therapy and Research, 1979, 3, 1-21.
Kendall, P. C., Williams, L., Pechacek, T. F., Graham, L. E., Shesslak, C., & Herzoff, N.
Cognitive-behavioral and patient education interventions in cardiac catheterization
procedures: The Palo Alto medical psychology project. Journal of Consulting and
Clinical Psychology, 1979, 47, 49-58.
Kuder, G. F., & Richardson, M. W. The theory of estimation of test reliability. Psycho-
metrika, 1937, 2, 151-160.
Mahoney, M. Cognition and behavior modification. Cambridge, Massachusetts: Ballinger,
1974.
Mahoney, M. Reflections on the cognitive-learning trend in psychotherapy. American
Psychologist, 1977, 32, 5-13.
Meichenbaum, D. Cognitive-behavior modification. New York: Plenum, 1977.
Nelson, R. E. Irrational beliefs in depression. Journal o f Consulting and Clinical Psychology,
1977, 45, 1190-1191.
Rehm, L. P. Assessment of depression. In M. Hersen & A. Bellack (Eds.), Behavioral assess-
ment: A practical handbook. New York: Pergamon Press, 1976.
Robins, E., & Guze, S. B. Classification of affective disorders: The primary-scondary, the
endogenous-reactive, and the neurotic-psychotic concepts. In T. A. Williams, M. M.
Katz, & J. A. Shields (Eds.), Recent advances in psychobiology of the depressive
illnesses. Washington, D.C.: U.S. Government printing Office, 1972.
Rush, A. J., Beck, A. T., Kovacs, M., & Hollon, S. D. Comparative efficacy of cognitive
therapy and pharmacotherapy in the treatment of depressed outpatients. Cognitive
Therapy and Research, 1977, 1, 17-37.
Schwartz, R. M., & Gottman, J. M. Toward a task analysis of assertive behavior. Journal o f
Consulting and Clinical Psychology, 1976, 44, 910-920.
Shaw, B. F. Comparison of cognitive therapy and behavior therapy in the treatment of de-
presion. Journal of Consulting and Clinical Psychology, 1977, 45, 543-551.
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. Manual for the State-Trait Anxiety
Inventory (Self-Evaluation Questionnaire). Palo Alto, California: Consulting Psy-
chologists Press, 1970.
Taylor, F. G., & Marshall, W. L. Experimental analysis of a cognitive-behavioral therapy for
depression. Cognitive Therapy and Research, 1977, 1, 59-72.