Content uploaded by Tian Po Oei
Author content
All content in this area was uploaded by Tian Po Oei on Sep 24, 2014
Content may be subject to copyright.
Behavioural and Cognitive Psychotherapy, 2002, 30, 351–360
Printed in the United Kingdom DOI: 10.1017/S1352465802003107
CHANGES TO AUTOMATIC THOUGHTS AND
DYSFUNCTIONAL ATTITUDES IN GROUP CBT FOR
DEPRESSION
Michele Furlong and Tian P. S. Oei
University of Queensland, Brisbane, Australia
Abstract. The present study sought to clarify the role of cognitive change in Cognitive Behavi-
oural Therapy (CBT) by examining the relationship between depression outcome and changes
to automatic thoughts and dysfunctional attitudes at different points of therapy. Thirty patients
suffering from Major Depression (MDD) or Dysthymia attended the 12 sessions of a group
CBT program. Multiple regressions found total scores on the Automatic Thoughts Question-
naire (ATQ) and cumulative change scores on the Dysfunctional Attitudes Scale (DAS) to pre-
dict scores on the Beck Depression Inventory (BDI) at later stages of therapy, though neither
form of cognition was predictive from earlier stages of therapy. Only scores on the ATQ were
significantly related to both cognitive and somatic subscales of the BDI, indicating that auto-
matic thoughts are more directly related to cognitive change than dysfunctional attitudes. Over-
all findings suggest that significant reductions in both automatic thoughts and dysfunctional
attitudes are related to non-clinical levels of depressive symptoms at the end of the treatment.
Keywords: Depression, psychotherapy, cognitive behaviour therapy, cognition, treatment
outcome.
Introduction
The need for further investigation into the active mechanisms of change in CBT is partly
due to equivalent treatment outcomes when comparing different therapies (Free & Oei,
1989; Oei & Free, 1995; Jacobson et al., 1996) and to the finding that cognitive change is
not exclusive to cognitive therapies, but is evident in the treatment outcomes of other therap-
ies such as pharmacotherapy and behavioural therapy (Oei & Free, 1995; Oei, Llamas, &
Devilly, 1999). In addition, previous studies have used ANOVA to analyse the roles of
cognitions during therapy and thus only infer their specific relationship during recovery
(Keller, 1983; DeRubeis et al., 1990; Oei & Sullivan, 1999), while other studies did not
separately examine the difference between cognitive and somatic outcome measures of
depression (Kwon & Oei, in press). Thus, while previous studies have added to the know-
ledge of cognitive change during recovery from depression, further investigation is required.
The current study firstly aimed to further clarify the role of cognitive change in the
recovery from depression by examining the relationship between BDI outcome scores and
Reprint requests and requests for extended report to Tian Oei, School of Psychology, University of Queensland
and Cognitive Behaviour Therapy Unit, Toowong Private Hospital, 496 Milton Road, Toowong, QLD 4066, Aus-
tralia. E-mail: oei@psy.uq.edu.au
2002 British Association for Behavioural and Cognitive Psychotherapies
M. Furlong and T. P. S. Oei352
changes to automatic thoughts and dysfunctional attitudes scores at different points in ther-
apy for depression. It was expected that outcome measures of depression would be predicted
at an earlier stage (Weeks 1 to 6) of assessment by automatic thoughts and at a later stage
(Weeks 7 to 12) by dysfunctional attitudes. This expectation was derived from the following
factors encompassed in cognitive theory of depression: changes in automatic thoughts are
proposed to have a direct effect on changes in depressive symptoms while changes in dys-
functional attitudes are said to have an indirect effect on changes in depressive symptoms,
mediated by automatic thoughts (Kwon & Oei, 1994, in press); automatic thoughts are
described as surface level and are therefore more accessible to change than the deeper
dysfunctional attitudes; automatic thoughts are the primary target of cognitive therapy in
the early stages of treatment while attention to dysfunctional attitudes is given later in
treatment (Jacobson et al., 1996).
The second aim of the current study was to analyse the relationship of two cognitive
measures with the Cognitive and Somatic subscales of the Beck Depression Inventory (BDI:
Beck & Steer, 1993). The expectation was that scores on the ATQ and DAS would predict
changes in outcome on the BDI Total and Cognitive scores, but that the scores on the
Somatic subscale would be better predicted by the ATQ than the DAS due to the direct
effect that automatic thoughts are said to have on changes to depressive symptoms (Kwon &
Oei, 1994). Although there have been a number of comprehensive reviews about the applica-
tion and psychometric properties of the BDI (Beck, Steer, & Garbin, 1988; Robinson &
Kelley, 1996), less attention has been given to the BDI subscales, which have been shown to
differentiate among psychiatric, medical and normal samples. In particular, previous studies
investigating the role of cognition in recovery from depression have used the BDI Total
score rather than subscales.
Method
Participants
The participants were 30 patients (21 males and 9 females) who had been diagnosed with
MDD or Dysthymia and who had completed all sessions of a 12-week, group CBT program.
The mean age of participants was 41.8 years (SD = 12.5) and ranged from 18 to 69 years.
Reports of marital status revealed 30.6% of subjects were single, 52.8% were married and
16.7% were divorced or separated. Although some patients were taking anti-depressant med-
ication during the course of therapy, recent studies have shown that, overall, pre-existing
medication regimes did not significantly affect the long-term outcome of a CBT program
(Oei, Llamas, & Evans, 1997; Oei & Yeoh, 1999).
Recruitment of participation
Participants were recruited from the Brisbane community through a local media release
seeking persons suffering from depression (Free, Oei, & Sanders, 1991). Participants were
screened through a series of interviews before being selected into the group CBT program.
Patients meeting DSM-IIIR criteria for current MDD (either single episode or recurrent) or
Dysthymia were selected for participation. Diagnosis was established through a 1.5 hour
interview that included the Structured Clinical Interview for Depression (SCID). Exclusion
Automatic thoughts and dysfunctional attitudes 353
from the study was based on participants meeting the criteria for the following DSM-IIIR
diagnoses: bipolar mood disorder or other major psychiatric disorder (e.g. schizophrenia,
personality disorder); organic brain disorder; abuse of drugs and/or alcohol. Exclusion from
the study was also based on the presence of a major physical illness or problems in reading
English. For more information on recruitment and exclusion criteria see Free et al. (1991)
and Oei and Shuttlewood (1997).
Procedure
A series of group CBT programs were run by the University of Queensland between 1992–
1994. CBT was delivered in a series of identical programs. Each program consisted of 12
two-hour group format sessions, held weekly. An outline of the program has been published
previously (Oei & Shuttlewood, 1997) and will therefore not be detailed here. Measures for
the study were taken pre-treatment (Week 1), post-treatment (Week 13) and approximately
fortnightly (Weeks 3, 5, 7,9, and 11) throughout the CBT program.
Assessment measures
Beck Depression Inventory. The 21 items of the Beck Depression Inventory (BDI; Beck &
Steer, 1993) are designed to assess the severity of the affective, cognitive, motivational, psy-
chomotor and vegetative components of depression (Beck, Ward, Mendelson, Mock, &
Erbaugh, 1961; Beck et al., 1988). Alpha reliability coefficients of the BDI have been found to
exceed .90 in a range of populations (Beck et al., 1988). The 21 items of the BDI can also be
separated into two subscales: a cognitive-affective subscale (the sum of the first 13 items) and
a somatic-performance subscale (the sum of the last eight items) (Beck & Steer, 1993). These
subscales, as well as the Total score of the BDI, were used in the analyses of the current study.
Finally, in accordance with guidelines stipulated by Beck and Steer (1993), the cut-offs for
assessing different levels of depression were as follows: 0–9 = minimal depression, 10–16 =
mild depression, 17–29 = moderate depression; and 30–36 = severe depression.
Automatic Thoughts Questionnaire. The Automatic Thoughts Questionnaire (ATQ;
Hollon & Kendall, 1980) is a 30-item inventory in which clients are asked to indicate, on a
scale of 1 (not at all) to 5 (all the time), how frequently negative automatic thoughts such
as ‘‘I’m worthless’’ have occurred in the past week. Scores on the ATQ can range from 30
to 150: scores for depressed persons usually range from 90–130 and between 40–60 for
non-depressed persons (DeRubeis et al., 1990). The ATQ is a widely accepted scale and
has been found to have good reliability and validity (Hill, Oei, & Hill, 1989).
Dysfunctional Attitudes Scale. The Dysfunctional Attitudes Scale (DAS; Weissman & Beck,
1978; Weissman, 1979; Beck, Brown, Steer, & Weissman, 1991) uses a 7-point rating scale
ranging from ‘‘totally agree’’ to ‘‘totally disagree’’ to measure those stable cognitive schemas
associated with depression (Beck et al., 1991). For example, ‘‘I should be happy all the time’’
and ‘‘My life is wasted unless I am a success’’. The 66-item version of the DAS was used in
the current study with scores ranging from 66 to 462. Lower scores indicate less dysfunction
in attitudes, with 236 being the lowest clinical level (Oei & Sullivan, 1999). Finally, the DAS
has been used widely and has good reliability and validity (Hill et al., 1989).
M. Furlong and T. P. S. Oei354
Statistical analyses
Transformations to the univariate and multivariate outliers that were present had little or no
effect on the interpretation of the results. Thus an untransformed, mean substituted data set
was chosen for analysis. Missing values were minimal (about 5%) and thus mean substitu-
tions were used. A series of repeated measures ANOVA with polynomial contrasts was
performed to analyse the significance between pre-treatment scores and changes in scores
that occurred at each fortnightly assessment time. ANOVA were performed for scores on
the ATQ, DAS, BDI Total, Cognitive, and Somatic measures.
A series of Linear Multiple Regressions were also performed using actual scores and
cumulative change scores as predictors of outcome. Actual scores were the raw scores taken
from the ATQ and DAS at each measurement time. Cumulative change scores were calcu-
lated by subtracting the pre-treatment raw scores from subsequent weeks of measurement
(i.e., from Week 3, Week 5, Week 9, etc.). Multiple regressions were run for each week of
assessment using the post-treatment Total, Cognitive and Somatic measures from the BDI
as the dependent variables each time.
Results
Table 1 displays the means and standard deviations of scores for the measures taken fort-
nightly across therapy. At Weeks 9 and 11 of therapy the DAS and ATQ, respectively,
showed that on average the scores were within non-depressed ranges (i.e. DAS < 236; ATQ
< 60) (Oei & Sullivan, 1999; De Rubeis et al., 1990). Although it appears that the mean
scores on the DAS reached a non-depressed range sooner than those of the ATQ, the latter
group mean was on the border of the non-depressed range at Week 9. In addition, the mean
scores on the ATQ up until Week 11 were within a range of scores that were below the range
for depressed persons, but above the range for non-depressed persons, possibly indicating a
lower level of depression. Repeated measures ANOVAs with polynomial contrasts revealed
that from Week 9 onwards subjects’ ATQ scores showed a significant change from pre-
treatment baseline scores, Fs (1,26) > 24.04, ps<.01; while for DAS scores a significant
change from baseline scores was not found until Week 11, Fs (1,25) > 7.19, ps<.01.
Table 1. Means and standard deviations of the scores on the ATQ, DAS, and BDI Total, Cognitive
and Somatic scales at the seven times of assessment
Pre Week 3 Week 5 Week 7 Week 9 Week 11 Post
ATQ Mean 81.77 77.87 76.73 71.11 61.5 58.70 50.03
SD 27.62 28.88 29.70 30.41 26.77 26.05 25.68
DAS Mean 242.80 260.08 251.58 247.27 229.6 220.27 193.07
SD 58.85 47.13 56.44 58.50 65.2 67.37 72.76
BDI Tot Mean 20.33 15.91 15.65 13.76 11.98 10.72 7.16
SD 7.62 7.61 8.37 8.13 8.64 9.46 7.23
BDI Cog Mean 12.98 10.04 9.95 8.62 7.43 7.02 4
SD 4.98 5.30 6.35 5.78 6.99 6.94 4.35
BDI Som Mean 7.19 5.87 5.75 5.22 5.54 4 3.16
SD 3.13 4.00 3.32 3.43 3.70 3.25 3.74
Automatic thoughts and dysfunctional attitudes 355
For the BDI Total, baseline scores ranged from 12–37 and on average were within the mod-
erately depressed range (17–29). BDI Total scores were not within the recovered/minimally-
depressed range (0–9) until the completion of therapy and, at this time, 22 of the 30 participants
had BDI Total scores within the range of 0–9. Repeated measures ANOVAs with polynomial
contrasts revealed that the mean BDI Total scores were significantly different from pre-
treatment scores at all of the measurement times across therapy, Fs (1,15) > 9.40, ps<.01.
Excluding Week 5, changes in mean BDI Cognitive scores were significant throughout assess-
ment, Fs (1,25) > 14.40, ps<.001; and for the BDI Somatic measure, Fs (1.25) > 8.34, ps<.01
for Weeks 5, 9, 11 and 13. These findings indicate that at these times of assessment, levels of
depression were significantly less than pre-treatment levels of depression. Lastly, although on
average the scores on the ATQ and the Total and subscales of the BDI all showed a steady
decrease over the course of therapy, scores on the DAS worsened for the first three weeks of
therapy and remained above pre-treatment levels until Week 9 of assessment.
The multiple regression analyses using raw scores from the ATQ and DAS as predictors
and the BDT Total, Cognitive and Somatic scores as outcome variables showed significant
results only for Week 9, Fs (2,27) > 5.42, p<.01; for Week 11, Fs (2,27) > 5.69, p<.01; and
for Week 13, Fs (2,27) > 7.04, p<.01. Weeks 1–7 were not significant (p>.05). Table 2
summarizes the findings for the significant regression analyses at Weeks 9, 11, and 13. At
Week 9, 41% of the variance in BDI Total scores was accounted for by the ATQ and DAS
Table 2. Multiple regression analyses predicting BDI Total, BDI Cognitive and BDI Somatic scores
from ATQ and DAS scores at Weeks 9, 11, and 13 of assessment
Criterion Predictor β 95% CI for β sr
2
pr R
2
Adj R
2
Week 9 Lower Upper
BDI Tot ATQ .62** .24 1.01 .24 .54 .41*** .37
DAS .03 −.35 .41 .0005 .03
BDI Cog ATQ .54** .15 .93 .18 .48 .39*** .35
DAS .12 −.26 .51 .0009 .13
BDI Som ATQ .59** .17 1.01 .22 .48 .29** .23
DAS −.09 −.51 .33 .0005 −.09
Week 11 Lower Upper
BDI Tot ATQ .57** .18 .95 .19 .50 .44*** .40
DAS .14 −.24 .52 .01 .14
BDI Cog ATQ .48** .09 .88 .14 .14 .42*** .38
DAS .22 −.16 .61 .03 .03
BDI Som ATQ .54** .11 .97 .17 .44 .30** .24
DAS .01 −.42 .44 .00005 .01
Week 13 Lower Upper
BDI Tot ATQ .70** .38 1.01 .32 .65 .56*** .53
DAS .08 −.24 .40 .004 .10
BDI Cog ATQ .62** .31 .93 .25 .62 .59*** .56
DAS .23 −.08 .53 .03 .28
BDI Som ATQ .64** .25 1.03 .27 .54
DAS −.11 −.28 .50 .008 −.11 .34** .29
* p<.05 ** p<.01 *** p<.001.
M. Furlong and T. P. S. Oei356
Table 3. Correlations between the predictors (ATQ and DAS); and between the predictors and the
criterions (BDI Total, Cognitive and Somatic) for Weeks 9, 11, and 13
Predictors DAS BDI Total BDI Cognitive BDI Somatic
Week 9 — ATQ .61*** .64*** .62*** .53***
— DAS — .41** .46** .27
Week 11 — ATQ .64*** .66*** .63*** .54***
— DAS — .60** .53** .36
Week 13 — ATQ .58*** .74*** .75*** .58***
— DAS — .48*** .58*** .26
* p<.05 ** p<.01 *** p<.001.
scores with up to 56% of the variance by Week 13. Similarly, the ATQ and DAS scores
from Weeks 9–13 accounted for 39–59% of the variance in BDI Cognitive scores and 29–
34% in BDI Somatic scores. As a predictor, the ATQ accounted for a greater proportion of
the total variance in each of the BDI criterion scores from Weeks 9, 11, and 13. The DAS
was not a significant predictor despite its significant correlations (p<.01 level) with the BDI
Total and BDI Cognitive scores from Week 9 onwards. Table 3 shows the correlations
between the predictors (ATQ and DAS); and between the predictors and the criterions (BDI
Total, Cognitive and Somatic) for Weeks 9, 11, and 13. The correlations amongst the pre-
dictors (p<.001 level) may account for the lack of prediction by the DAS. A non-significant
correlation was found between the DAS and the BDI Somatic measure at these times.
Similarly, analyses of cumulative change scores showed that only two regression analyses
were significant while the other multiple regressions were not significant at a p<.01 level.
The two significant regressions occurred when Week 9 ATQ and DAS scores minus pre-
treatment ATQ and DAS scores, respectively, were used as cumulative change scores with
the BDI Total as the criterion, F(2,27) = 5.35, p<.01; and similarly when Week 13 ATQ and
DAS scores minus pre-treatment ATQ and DAS scores, respectively, were used as cumulat-
ive change scores with the BDI Cognitive scores as the criterion, F(2,27) = 5.35, p<.01.
Table 4 summarizes the findings from these two significant regression analyses using cumu-
lative change ATQ and DAS scores as predictors. As indicated 28% of BDI Total and 28%
Table 4. Multiple regression analyses predicting BDI Total scores and BDI Cognitive scores from
cumulative ATQ and DAS scores
Criterion Predictor change scores β 95% CI for β sr
2
pr R
2
Adj R
2
Lower Upper
BDI Tot Week 9 ATQ minus .01** −.39 .41 .0006 .02 .28** .23
pre-treatment ATQ
Week 9 DAS minus .53 .12 .93 .19 .14
pre-treatment DAS
BDI Cog Week 13 ATQ minus .15* −.39 .54 .02 .15 .28** .23
pre-treatment ATQ
Week 13 DAS minus .44 .05 .82 .14 .40
pre-treatment DAS
* p<.05 ** p<.01 *** p<.001.
Automatic thoughts and dysfunctional attitudes 357
Table 5. Correlations between the predictors (Cumulative ATQ and DAS scores); and between the
predictors and the criterions (BDI Total, Cognitive and Somatic scores) for Week 9 minus
pre-treatment and Week 13 minus pre-treatment
Change Scores DAS BDI Total BDI Cognitive BDI Somatic
Week 9 ATQ minus .58*** .31* .29 .24
pre-treatment ATQ
Week 9 DAS minus — .53*** .24 .21
pre-treatment DAS
Week 13 ATQ minus .52*** .29 .38 .26
pre-treatment ATQ
Week 13 DAS minus — .23 .52** .21
pre-treatment DAS
* p<.05 ** p<.01 *** p<.001.
of BDI Cognitive outcome scores could be predicted by scores measuring the amount of
improvement in patients’ automatic thoughts and dysfunctional attitudes between pre-
treatment and Week 9 of therapy and between pre-treatment and Week 13 (post-treatment).
The cumulative-change DAS scores showed to be better predictors and accounted for a
greater proportion of total variance in the BDI Total measure, (when using Week 9 minus
pre-treatment scores as predictors) and in the BDI Cognitive measure, (when using Week
13 minus pre-treatment scores as predictors) than did the cumulative-change ATQ scores at
these times of assessment. The lack of significance of the cumulative change ATQ scores
as a predictor in these regression analyses may be due to the correlations of .58 (p<.001)
and .52 (p<.01), respectively, between the predictors (see Table 5 for correlations between
the cumulative-change predictors; and between the cumulative-change predictors and the
criterion for Week 9 minus pre-treatment and Week 13 minus pre-treatment).
Discussion
Examination of the mean scores on the ATQ and DAS indicates that changes in automatic
thoughts occur early in treatment while improvements in dysfunctional attitudes occur later
after an initial worsening. The early change to scores on the ATQ supports previous findings
that automatic thoughts change relatively quickly during therapy (Oei & Shuttlewood, 1997;
Oei & Sullivan, 1999); however, in the current study such changes were not significantly
different from baseline scores until Week 9. At this time, ATQ scores were also significantly
related to outcomes in both cognitive and somatic depression as measured by BDI Total,
Cognitive and Somatic scales. The significant relationship between later ATQ scores and
depression outcome may reflect the greater changes in automatic thoughts that have occurred
by the end of therapy (Oei & Sullivan, 1999); alternatively the relationship may be due to
the correlation of two similar depression measures (the ATQ and BDI) when completed
closer in time to each other. Thus there appears to be some support for the association of
depression outcome with significant changes to be automatic thoughts over therapy, but
some caution must be taken when making conclusions about these findings.
As predicted, cumulative change scores on the DAS taken later in assessment were signi-
ficantly related to depression outcome. A similar finding was not found for the actual DAS
M. Furlong and T. P. S. Oei358
scores. Possibly, cumulative change scores reduce much of the initial differences between
individuals and are in fact measuring a rate of change throughout recovery. It is suggested
that cumulative change scores are a more conservative measure of cognitive change and are
less affected by depression severity compared to actual scores. Furthermore, the ability of
cumulative change DAS scores to predict BDI total and Cognitive scores rather than Somatic
scores supports the proposal that dysfunctional attitudes have an indirect effect on somatic
symptoms of depression while their predictive ability towards the end of therapy rather than
early in therapy supports the proposal that dysfunctional attitudes are of a deeper more rigid
structure compared to automatic thoughts (Rude & Rehm, 1991; Kwon & Oei, 1994).
While early changes in ATQ scores lend support for their surface structure, it is possible
that these changes are too small to predict depression outcome scores taken at the completion
of therapy. Although scores on the DAS were not significantly different from baseline scores
until Week 11, the finding that at Week 9 both ATQ and DAS mean scores reached non-
clinical levels and were able to predict depression outcome scores may suggest that signific-
ant reductions in both automatic thoughts and dysfunctional attitudes may be required before
a predictive relationship between cognitions and depression symptoms can be found. In fact,
regression statistics may require critical changes in measurement scores before detection of
change is statistically and clinically possible.
Alternatively, limitations in sample size may have prevented a true detection of cognitive
change in the process of recovery from depression. Any results (significant or
non-significant) may be confounded with the lack of variability that a small sample size
offers. Clients who begin therapy with moderate levels of depression may not show signi-
ficant changes to cognitions or somatic symptoms until later in therapy.
Finally, the initial worsening of dysfunctional attitudes over the first 7 weeks of therapy
may be due to an increased awareness of cognitions brought about by therapeutic techniques
(Oei & Shuttlewood, 1997). Dysfunctional attitudes may remain rigid and resistant to change
until this type of cognition is addressed later in therapy. Automatic thoughts, having a
surface level, situation-specific nature are addressed early in therapy and may be more
accessible to change though increased awareness (Rude & Rehm, 1991; Kwon & Oei, 1994).
Findings in the current study must be taken with caution due to the small sample (N = 30)
that affected the power of analyses.
In summary, support is found for the proposal that automatic thoughts are more strongly
related to changes to depressive symptoms and would be more closely related to both cognit-
ive and somatic change in depression compared to dysfunctional attitudes (Kwon & Oei,
1994). That is, despite the correlation between actual DAS scores and the BDI Total and
BDI Cognitive measures from Week 9 onwards, the ATQ was still the better predictor. This
may be due to the correlation amongst the predictors, but also to the effect that automatic
thoughts mediate between dysfunctional attitudes and depressive symptoms (Kwon & Oei,
1994, 2002). Finally, the current study supports previous findings that cognitive change
coincides with the recovery from depression (De Rubeis et al., 1990, Oei & Shuttlewood,
1997; Oei & Sullivan, 1999) as decreases in both cognitive and somatic symptoms of depres-
sion to a ‘‘recovered’’ or ‘‘non-depressed’’ level were found.
Acknowledgements
This project was supported by a grant from the Australian Research Council. We would like
to thank the patients and Ms Anna Lamberton and Dr Michael Free for their help.
Automatic thoughts and dysfunctional attitudes 359
References
B
ECK
, A. T., B
ROWN
, G., S
TEER
,R.A.,&W
EISSMAN
, A. (1991). Factor Analysis of the Dysfunctional
Attitudes Scale in a clinical population. Psychological Assessment, 3, 478–483.
B
ECK
, A. T., & S
TEER
, R. A. (1993). Beck Inventory Manual. The Psychological Corporation. Harcourt
Brace & Co.
B
ECK
, A. T., S
TEER
,R.A.,&G
ARBIN
, M. (1988). Psychometric properties of the Beck Depression
Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77–100.
B
ECK
, A. T., W
ARD
, C. H., M
ENDELSON
, M., M
OCK
, J., & E
RBAUGH
, J. (1961). An inventory for
measuring depression. General Archives of Psychiatry, 4, 561–571.
D
E
R
UBEIS
, R. J., E
VANS
, M. D., H
OLLON
, S. D., G
ARVEY
, M. J., G
ROVE
,W.M.,&T
UASON
,V.B.
(1990). How does cognitive therapy work? Cognitive change and symptom change in cognitive
therapy and pharmacotherapy for depression. Journal of Consulting and Clinical Psychology, 58,
862–869.
F
REE
, M. L., & O
EI
, T. P. S. (1989). Biological and psychological processes in the treatment and
maintenance of depression. Clinical Psychology Review, 9, 653–688.
F
REE
, M. L., O
EI
, T. P. S., & S
ANDERS
, M. R. (1991). Treatment outcome of a group therapy program
for depression. International Journal of Group Psychotherapy, 41, 533–547.
H
ILL
, C. V., O
EI
, T. P. S., & H
ILL
, M. A. (1989). An empirical investigation of the specifity and
sensitivity of the Automatic Thoughts Questionnaire and the Dysfunctional Attitudes Scale. Journal
of Psychopathology and Behavioural Assessment, 11, 291–311.
H
OLLON
,S.D.,&K
ENDALL
, P. C. (1980). Cognitive self-statements in depression: Development of
an Automatic Thoughts Questionnaire. Cognitive Therapy and Research, 4, 383–395.
J
ACOBSON
, N. S., D
OBSON
, K. S., T
RUAX
, P. A., A
DDIS
, M. E., K
OERNER
, K., G
OLLAN
, J. K., G
ORTNER
,
E., & P
RINCE
, S. E. (1996). A component analysis of cognitive-behavioural treatment for depression.
Journal of Consulting and Clinical Psychology, 64, 295–304.
K
ELLER
, K. E. (1983). Dysfunctional attitudes and the cognitive therapy for depression. Cognitive
Therapy and Research, 7, 437–444.
K
WON
,S.M.,&O
EI
, T. P. S. (1994). Roles of two levels of cognition in the development, maintenance
and treatment of depression. Clinical Psychology Review, 14, 331–358.
K
WON
,S.M.,&O
EI
, T. P. S. (in press). Cognitive change processes in depression following a group
cognitive behaviour therapy treatment. Journal of Behaviour Therapy and Experimental Psychiatry.
O
EI
, T. P. S., & F
REE
, M. L. (1995). Do cognitive behaviour therapies validate cognitive models of
mood disorders? A review of the empirical evidence. International Journal of Psychology, 30, 145–
179.
O
EI
, T. P. S., L
LAMAS
,M.,&D
EVILLY
, G. J. (1999). The efficacy and cognitive processes of cognitive
behaviour therapy in the treatment of panic disorder with agoraphobia. Behavioural and Cognitive
Psychotherapy, 27, 63–88.
O
EI
, T. P. S., L
LAMAS
,M.,&E
VANS
, L. (1997). Does concurrent drug intake affect the long-term
outcome of group-cognitive behaviour therapy in panic disorder with or without agoraphobia? Beha-
viour Research and Therapy, 35, 851–857.
O
EI
, T. P. S., & S
HUTTLEWOOD
, G. J. (1997). Comparison of specific and nonspecific factors in a group
cognitive therapy for depression. Journal of Behavior Therapy and Experimental Psychiatry, 28,
221–231.
O
EI
, T. P. S., & S
ULLIVAN
, L. M. (1999). Cognitive changes following recovery from depression in a
group cognitive-behaviour therapy program. Australian and New Zealand Journal of Psychiatry,
33, 407–415.
O
EI
, T. P. S., & Y
EOH
, A. E. O. (1999). Pre-existing antidepressant medication and the outcome of
group cognitive-behavioural therapy. Australian and New Zealand Journal of Psychiatry, 33, 70–76.
R
OBINSON
, B. E., & K
ELLEY
, L. (1996). Concurrent validity of the Beck Depression Inventory as a
measure of depression. Psychological Reports, 79, 929–930.
M. Furlong and T. P. S. Oei360
R
UDE
, S. S., & R
EHM
, I. P. (1991). Response to treatment for depression. The role of initial status on
targeted cognitive and behavioral skills. Clinical Psychology Review, 11, 495–514.
W
EISSMAN
, A. N. (1979). The Dysfunctional Attitude Scale: A validation study. Dissertation Abstracts
International, 40, 1389–1390.
W
EISSMAN
,A.N.,&B
ECK
, A. T. (1978). Development and validation of the Dysfunctional Attitude
Scale: A preliminary investigation. Paper presented at the meeting of the American Educational
Research Association, Toronto, Ontario, Canada.
