PSYCHOMETRIC PROPERTIES OF A PERSIAN-LANGUAGE

VERSION OF THE AUTOMATIC THOUGHTS QUESTIONNAIRE:

ATQ-PERSIAN

HABIBOLLAH GHASSEMZADEH, RAMIN MOJTABAI, NARGES

KARAMGHADIRI & NARGES EBRAHIMKHANI

ABSTRACT
Aims: To examine the psychometric properties of a Persian-language version of
Hollon and Kendall’s (1980) Automatic Thoughts Questionnaire (ATQ-Persian) –
a measure of negative automatic thoughts in depression.
Methods: In a sample of 125 student volunteers from two Iranian universities we
assessed the internal consistency and test–retest reliability of ATQ-Persian, and
examined its concurrent validity against the Beck Depression Inventory (Beck et
al., 1996). We also examined the factor structure of the test through comparing
the fit of various proposed factor-analytic models to the data using confirmatory
factor analysis.
Results: ATQ-Persian had excellent internal consistency (Cronbach’s alpha ¼
0.96), test–retest reliability (r ¼ 0.84) and correlation with the Beck Depression
Inventory (r ¼ 0.77). Using a receiver operating characteristics curve, a score
of 53 on ATQ-Persian was associated with a sensitivity of 89% and a specificity
of 78% for detecting moderate to severe depression defined through Beck
Depression Inventory. None of the proposed factorial models fitted the data
well. However, models with correlated factors generally provided better fit than
models with orthogonal factors, suggesting that the latent factors underlying
ATQ-Persian are highly correlated.
Conclusions: These data support the reliability and validity of ATQ-Persian as a
measure of negative automatic thoughts in depression.

INTRODUCTION

Automatic thoughts (ATs) are major elements of both normal and abnormal cognitive pro-
cessing. ATs are ready-made thoughts or self-statements that come to mind when the person
encounters a problem – hence ‘automatic’ (Best, 1992; Shiffrin & Schneider, 1977). Negative
ATs are also central to cognitive formulations of clinical depression (Beck, 1976; Beck et al.,
1979; Stein & Young, 1992; Teasdale & Barnard, 1993). Automatic thoughts in depression
are idiosyncratic, involuntary, but plausible self-statements that contribute to recurrence
or maintenance of depression and may mediate other features of depressive illness such as
dysphoria and loss of motivation. A major focus of cognitive therapy, therefore, is to
International Journal of Social Psychiatry. Copyright & 2006 Sage Publications (London, Thousand Oaks and
New Delhi) www.sagepublications.com Vol 52(2): 127–137. DOI: 10.1177/0020764006062095

Downloaded from isp.sagepub.com by guest on January 16, 2015

128 INTERNATIONAL JOURNAL OF SOCIAL PSYCHIATRY 52(2)

challenge and modify negative ATs (Beck, 1967). When the person can correctly identify
these distorted cognitions and acquire objectivity towards them, their pathogenic effect will
be neutralised. Thus, from a cognitive perspective, measurement of ATs is critical for assess-
ment of depression and response to treatment.
There have been a number of attempts to develop measures of negative automatic
thoughts, with promising results (Glass & Arnkoff, 1997). From their review of the literature,
Glass and Arnkoff (1997) concluded that: ‘Self-statement questionnaires have demonstrated
good to excellent internal consistency, are able to differentiate between adjusted (nonclinical)
and dysfunctional (clinical) groups, and correlate significantly with other measures of affect,
behavior, and symptomatology’ (p. 923).
Perhaps the most widely used of such measures is the Automatic Thoughts Questionnaire
(ATQ) developed by Hollon and Kendall (1980). Past research has found the ATQ to be a
reliable and valid measure of automatic thoughts in clinical as well as general population
samples (Dobson & Shaw, 1986; Harrell & Ryon, 1983; Hill et al., 1989; Hollon & Kendall,
1980; Hollon et al., 1986). The cross-cultural utility of the ATQ has also been investigated,
using foreign-language translations of the instrument (Charles et al., 1989; Revuelta Perez,
1995; Sahin & Sahin, 1992). In this article we report on the psychometric properties of a
Persian-language version of the ATQ (ATQ-Persian), which was administered to a sample
of Iranian college students. More specifically, we assess the internal consistency and test–
retest reliability of ATQ-Persian and examine its concurrent validity against the Beck Depres-
sion Inventory (BDI-II, Beck et al., 1996). We also examine the factor structure of the
ATQ-Persian through comparing the fit of various proposed factor models to our data using
confirmatory factor analysis.

METHODS

Sample
The sample for this study comprised 125 students from Tehran University of Medical
Sciences and the Allameh Tabatabai University, Tehran, Iran, who volunteered to participate
in the study. Of these, 82 (65.6%) were medical students and 43 (34.4%) were graduate
students of psychology. Average age of the sample was 22.1 years (SD ¼ 2:7, range ¼ 19–37);
50.4% (N ¼ 63) were female and 49.6% (N ¼ 62) male. The majority of students were single
(N ¼ 105, 84.0%).

Assessments
ATQ-Persian is a translation of the 30-item Hollon and Kendall (1980) ATQ questionnaire
into colloquial Persian (Farsi) language. The original Persian translation was fine-tuned
through back-translation into English by a bilingual mental health professional not familiar
with the instrument. Similar to the English version, on the ATQ-Persian, respondents are
asked to read 30 statements about various negative thoughts and then report how frequently
those thoughts have crossed their minds during the past week and specify the degree of their
belief in these statements. Each frequency response is scored on a scale of 1 (not at all) to 5 (all
the time). Scores are added across the 30 items to compute a total score (range ¼ 30–150).

Downloaded from isp.sagepub.com by guest on January 16, 2015

 GHASSEMZADEH ET AL.: PSYCHOMETRIC PROPERTIES OF ATQ-PERSIAN 129

One to two weeks after the baseline administration, the test was readministered to 123
(98.4%) of the baseline sample to assess test–retest reliability.
BDI-II is the second edition of the widely used 21-item Beck Depression Inventory.
Originally introduced in the 1960s (Beck et al., 1961), BDI-II was revised to approximate
the DSM-IV (American Psychiatric Association, 1994) criteria for major depression. BDI
is perhaps the most widely used self-report measure of depression both in English-speaking
countries and across the world. The reliability and validity of various non-English versions
of the BDI were established in numerous previous studies. The Persian translation of this
test is widely used in various settings in Iran. The version of the BDI-II used in this study
had excellent internal consistency (
¼ 0:87) and 1–2 week test–retest reliability (r ¼ 0:73).

Statistical analysis
Analyses were conducted in five stages. First, we compared the total and item scores on
ATQ-Persian with data from a sample of US college students who were administered the
English-language versions of ATQ (Bryant & Baxter, 1997). Second, we assessed the internal
consistency of the ATQ-Persian by computing Cronbach’s alpha. Third, we assessed the test–
retest reliability of the measure by computing the Pearson correlation coefficient for scores
obtained across two assessment points 1–2 weeks apart. Fourth, we assessed the concurrent
validity of the ATQ-Persian against BDI-II by computing the correlation between the two
tests and by examining the sensitivity and specificity of ATQ-Persian for predicting moderate
to severe depression (defined as a score of 20 or higher on BDI-II) (Beck et al., 1996). The
Receiver Operating Characteristics (ROC) curve method was used for the analyses of the sen-
sitivity and specificity of the test (Metz, 1978). The results of all these analyses are reported
along with data from previous studies of ATQ as benchmarks.
Finally, we assessed the factor structure of ATQ-Persian by comparing the fit of various
proposed models to the ATQ-Persian data, using the confirmatory factor analysis method.
In comparing various models we used change in the model chi-square (
x2 ) as an index of
improvement in model fit when comparing nested models. We also used six other indices
(model x2 , x2 /df, root mean square residual [RMSR], goodness-of-fit index [GFI], adjusted
goodness-of-fit index [AGFI], and Bentler-Bonett normed fit index) as indicators of goodness
of fit for each model (Loehlin, 1992). Although a discussion of these indices and their limita-
tions is beyond the scope of this article, a brief description is in order.
Model x2 is perhaps the most widely used measure of goodness of fit. Statistically non-
significant values of x2 and values of x2 /df that are close to 1 or smaller indicate a good fit of
the model to the data. RMSR, GFI and AGFI are all based on residuals, a smaller value of
RMSR and larger values of GFI and AGFI indicate better fit. A RMSR less than 0.10, a GFI
larger than 0.90 and an AGFI larger than 0.80 are conventionally regarded as indicators of
good fit. The Bentler-Bonett normalised fit index is also based on the model x2 compared to
the x2 of the null model. Values larger than 0.90 indicate good fit of the model.

RESULTS

Total and item scores on ATQ-Persian

The total score on ATQ-Persian was 49.78 (SD ¼ 18:41) for the whole sample. There were no

Downloaded from isp.sagepub.com by guest on January 16, 2015

130 INTERNATIONAL JOURNAL OF SOCIAL PSYCHIATRY 52(2)

differences across genders (t ¼ 0:61, df ¼ 123, p ¼ 0:55) or student groups (medical versus
graduate psychology students; t ¼ 0:88, df ¼ 123, p ¼ 0:38) on total score. ATQ-Persian
total score was also not associated with age (r ¼
0:002, p ¼ 0:99). The average item scores
in the whole group and in males and females separately are reported in Table 1. The total
score of ATQ-Persian in this sample is very close to the total score of 48.57 (SD ¼ 10:89)
reported from the non-depressed subsample in the Hollon and Kendall (1980) study and
the total score of 48.84 (SD ¼ 16:16) reported from a sample of midwestern factory workers
(Deardorf et al., 1984). However, the total score in our study was somewhat lower than the
total score of 52.84 (SD ¼ 14:92) in a sample of 345 Turkish university students who were
administered the Turkish-language version of ATQ.

Table 1
Descriptive statistics and item-total correlations of the ATQ-Persian items

Item means and standard deviations

Items Male Female Total Item-total

Mean SD Mean SD Mean SD correlation

1. I feel like I am up against the world 1.58 0.80 1.60 0.76 1.59 0.78 0.50
2. I’m no good 1.37 0.63 1.51 0.76 1.44 0.70 0.68
3. Why can’t I ever succeed? 1.61 0.80 1.79 1.03 1.70 0.92 0.65
4. No one understands me 1.85 0.90 1.90 1.05 1.88 0.98 0.52
5. I’ve let people down 1.45 0.74 1.61 0.76 1.53 0.75 0.62
6. I don’t think I can go on 1.44 0.90 1.69 0.98 1.57 0.95 0.79
7. I wish I were a better person 2.53 1.21 2.51 1.35 2.52 1.28 0.55
8. I’m so weak 1.51 0.96 1.54 0.84 1.52 0.90 0.72
9. My life’s not going the way I want 2.13 0.96 2.23 1.14 2.18 1.05 0.69
it to
10. I’m so disappointed in myself 1.44 0.80 1.62 0.85 1.53 0.83 0.80
11. Nothing feels good anymore 1.56 0.84 1.54 0.74 1.55 0.79 0.70
12. I can’t stand this anymore 1.48 0.84 1.67 0.95 1.58 0.90 0.72
13. I can’t get started 1.56 0.82 1.59 0.96 1.58 0.89 0.68
14. What’s wrong with me? 1.61 0.94 1.73 1.07 1.67 1.00 0.64
15. I wish I were somewhere else 2.21 1.16 1.95 1.21 2.08 1.19 0.50
16. I can’t get things together 1.60 0.80 1.76 0.95 1.68 0.88 0.72
17. I hate myself 1.31 0.59 1.37 0.77 1.34 0.68 0.69
18. I’m worthless 1.26 0.68 1.35 0.60 1.30 0.64 0.70
19. Wish I could just disappear 1.53 1.11 1.32 0.59 1.42 0.89 0.58
20. What is the matter with me? 1.70 0.88 1.95 1.04 1.83 0.97 0.65
21. I’m a loser 1.37 0.89 1.32 0.78 1.35 0.84 0.76
22. My life is a mess 1.61 0.94 1.79 0.99 1.70 0.96 0.82
23. I’m a failure 1.55 0.88 1.67 0.92 1.61 0.90 0.80
24. I’ll never make it 1.24 0.72 1.33 0.62 1.29 0.67 0.77
25. I feel so helpless 1.53 0.95 1.65 0.83 1.59 0.89 0.74
26. Something has to change 2.58 1.18 2.65 1.15 2.62 1.16 0.60
27. There must be something wrong 1.79 0.89 1.66 0.87 1.73 0.88 0.65
with me
28. My future is bleak 1.39 0.78 1.43 0.92 1.41 0.85 0.71
29. It’s just not worth it 1.63 1.12 1.70 1.13 1.66 1.12 0.59
30. I can’t finish anything 1.50 0.86 1.65 1.05 1.58 0.96 0.62
Total score 48.77 19.44 50.78 17.45 49.78 18.41 –

Downloaded from isp.sagepub.com by guest on January 16, 2015

 GHASSEMZADEH ET AL.: PSYCHOMETRIC PROPERTIES OF ATQ-PERSIAN 131

To assess the ‘difficulty level’ of the ATQ-Persian items, we compared the item score pro-
files from our study with those from a study of 304 undergraduate students in a metropolitan
university in the USA (Bryant & Baxter, 1997). To make scores more comparable, mean item
scores on each test were standardised to the average of all items. Thus, for example, for
calculating the standardised mean item score for item 1, the difference between the mean
on this item and the average of all item means was computed and divided by the standard
deviation of all the item means. Standardised mean item scores calculated in this way repre-
sent deviations for each item mean from the average of all item means in standard deviation
units (i.e. z-transformed scores).
The standardised mean item scores for ATQ-Persian and the English-language ATQ
followed a similar pattern overall (Figure 1). The Spearman rank correlation between the
mean item scores across the two tests was 0.70, indicating a strong correspondence in the
pattern of responses to items on the ATQ-Persian and the English version of the test.

Internal consistency
Cronbach’s alpha for the ATQ-Persian was 0.96. Item-total correlations (Table 1, last
column) ranged from 0.50 to 0.82, and had an average of 0.67, indicating a high level of
internal consistency among the items and also a cohesive latent structure underlying these
items. The alpha coefficient obtained in this study was equal to the one obtained in the
original Hollon & Kendall (1980) study and the study by Bryant & Baxter (1997), but was
somewhat higher than that obtained in the Deardorf et al. (1984) study (0.92). Cronbach’s
alpha in both the Turkish and the French versions of the ATQ was 0.93 (Charles et al.,
1989; Sahin & Sahin, 1992).

Figure 1. The profiles of mean item scores on ATQ-Persian compared to mean item scores in the study by
Bryant & Baxter (1997) based on the English version of ATQ. The z-transformed scores represent deviation of
the mean score for each item from the average of the mean item scores on each test

Downloaded from isp.sagepub.com by guest on January 16, 2015

132 INTERNATIONAL JOURNAL OF SOCIAL PSYCHIATRY 52(2)

Test–retest reliability
ATQ-Persian also revealed remarkable consistency across time. The Pearson correlation
coefficient for the ATQ-Persian total scores across the two time points was 0.84. Test–
retest reliability of ATQ has rarely been examined. In a study of the French version of
ATQ, test–retest correlation of two administrations, 15 days apart, was 0.71 (Charles et
al., 1989).

Concurrent validity
ATQ-Persian was highly correlated with the BDI-II (r ¼ 0:77). The results of the ROC
analysis revealed that ATQ-Persian can effectively detect moderate to severe depression,
defined by a BDI-II score of 5 20, in this population (area under the curve ¼ 0:92, 95%
CI ¼ 0:86–0.99) (Figure 2). A score of 53 on ATQ-Persian was associated with a sensitivity
of 89% and specificity of 78% for detecting moderate to severe depression. The association
between ATQ and BDI scores has been examined in a number of previous studies (Harrell &
Ryon, 1983; Hollon & Kendall, 1980; Joseph, 1994; Peden et al., 2000; Sahin & Sahin, 1992).
The Pearson correlation coefficients between ATQ and BDI in these studies range from
r ¼ 0:75 to r ¼ 0:83. Thus, the correlation coefficient from our study falls within this range.

Factor structure
A number of exploratory and confirmatory factor-analytic studies have investigated the
factor structure of ATQ (Bryant & Baxter, 1997; Charles et al., 1989; Deardorf et al., 1984;
Hollon & Kendall, 1980; Joseph, 1994, Netemeyer et al., 2002; Sahin & Sahin, 1992) with
different results. The number of factors in the exploratory factor analyses ranged from 1 to
5 and the number of items loading on these factors from 15 to 30. To assess the fit of the
various proposed models to our data we used a stepwise approach similar to the one used
by Bryant & Baxter (1997), in which each of the proposed models were compared to a

Figure 2. Receiver Operating Characteristics curve for ATQ-Persian predicting moderate to severe depression
(BDI-II 5 20)

Downloaded from isp.sagepub.com by guest on January 16, 2015

 GHASSEMZADEH ET AL.: PSYCHOMETRIC PROPERTIES OF ATQ-PERSIAN 133

model with one global factor. In addition, models with orthogonal (uncorrelated) factors
were compared to models with correlated factors.
The first model examined was a model with one global factor for all items (Table 2). While
the RMSR for this model was rather low, the other fit indices indicated a poor fit. Next,
the original model of ATQ proposed by Hollon & Kendall (1980) with four orthogonal fac-
tors was examined. The model provided a poor fit to data as judged by all fit indices. Allowing
for the factors to correlate greatly improved the model fit (
x2 ¼ 241:27, df ¼ 6, p < 0:001).
This model, however, provided a poorer fit to data when compared with a one-factor model
based on the same 16 items (
x2 ¼ 27:71, df ¼ 4, p < 0:001). Judged by the fit indices, neither
of these models provided a good fit.
The next model examined was the one proposed by Deardorf et al. (1984) with three ortho-
gonal factors (Table 2). The model provided a poor fit to data as judged by all fit indices.
Allowing for the factors to correlate improved the fit of the model considerably
(
x2 ¼ 292:51, df ¼ 3, p < 0:001). But, even for this improved model, the indices did not
indicate a good fit to data. Based on their review of the correlations among the factors in
the model by Deardorf et al. (1984), Bryant & Baxter (1997) proposed combining two of
the factors into one. In our data, this two-factor model was very similar to the three-factor
model with correlated factors and the model with one global factor.
The third model tested was a model with five orthogonal factors proposed by Sahin &
Sahin (1992) based on a Turkish version of the ATQ. The original model tested provided a
poor fit to the data judged by all indices. Allowing for the factors to correlate improved
the fit of the model considerably (
x2 ¼ 622:56, df ¼ 10, p < 0:001). However, even the
model with correlated factors did not provide a good fit to the data, as judged by most fit
indices. The model with one global factor based on the same items was associated with an
even poorer fit than the model with multiple correlated factors (
x2 ¼ 70:62, df ¼ 10,
p < 0:001).
Finally, the model with two correlated factors proposed by Joseph (1994) was tested.
Similar to all other models tested, this model provided a poor fit to the data judged by
most indices. However, it fit the data better than the model with one global factor based
on the same items (
x2 ¼ 11:70, df ¼ 1, p < 0:001).

DISCUSSION

ATQ-Persian appears to be a reliable and valid measure of negative automatic thoughts in

this non-clinical sample of Iranian students. The psychometric properties of ATQ-Persian
are quite similar to those reported for the English and other language versions of the test.
ATQ-Persian has high internal consistency and is reliable across time. Furthermore, the
pattern of responses to items on the Persian-language version administered to Iranian college
students is remarkably similar to the profile on the English-language version administered to
a sample of US college students. ATQ-Persian is also strongly correlated with self-reported
depression as measured by BDI-II and can detect moderate to severe depression with remark-
able sensitivity and specificity.
The results of the confirmatory factor analyses reported here did not provide a clear solu-
tion. However, this finding is not unique to the ATQ-Persian because various analyses based

Downloaded from isp.sagepub.com by guest on January 16, 2015

 Table 2

134
Results of the comparisons of different factorial models for ATQ-Persian

Model fit indices

Number of Bentler-Bonett
Model items v2 df v2/df RMSR GFI AGFI normed fit index

INTERNATIONAL JOURNAL
1 global factor 30 1122.15 405 2.80 0.079 0.59 0.52 0.62
a
Hollon and Kendall, 1980
b
4 factors, uncorrelated 16 623.77 106 b 5.88 0.381 0.61 0.49 0.53
b
100 b
Downloaded from isp.sagepub.com by guest on January 16, 2015

4 factors, correlated 16 382.50 3.83 0.128 0.70 0.59 0.71

1 global factor 16 354.79 104 3.41 0.078 0.71 0.62 0.73

Deardorf et al., 1984 c

3 factors, uncorrelated 15 560.08 90 6.22 0.400 0.66 0.55 0.56
3 factors, correlated 15 267.57 87 3.08 0.065 0.76 0.67 0.79
2 factors, correlated d 15 268.99 89 3.02 0.065 0.76 0.68 0.79
1 global factor 15 268.99 90 2.99 0.063 0.76 0.69 0.79

OF
Sahin & Sahin, 1992 e

SOCIAL PSYCHIATRY 52(2)

5 factors, uncorrelated 27 1468.45 324 4.53 0.416 0.51 0.43 0.44
5 factors, correlated 27 845.89 314 2.69 0.078 0.64 0.56 0.68
1 global factor 27 916.51 324 2.83 0.076 0.61 0.55 0.65

Joseph, 1994 f
2 factors, correlated 27 940.52 323 2.91 0.080 0.60 0.54 0.64
1 global factor 27 952.22 324 2.94 0.080 0.60 0.53 0.63
a
Factors and items loading on them included: ‘personal maladjustment and desire for change’ (items 7, 10, 14, 20, 26), ‘negative self-concept and
negative expectations’ (items 2, 3, 9, 21, 23, 24, 28), ‘low self-esteem’ (items 17, 18), and ‘giving up/helplessness’ (items 29, 30).
b
Unique variances for items 18 and 29 were set to zero to avoid negative variances (Heywood case) in estimating this model (Loehlin, 1992).
c
Factors and items loading on them included: ‘negative self-statements’ (10, 17, 18, 21, 22, 23), ‘giving up’ (6, 12, 15, 19, 29), ‘weakness’ (8, 13, 25, 30).
d
This modification of the Deardorf et al. (1984) model was introduced by Bryant & Baxter (1997) and combined the ‘giving up’ and ‘weakness’ factors.
e
Factors and items loading on them included: ‘negative self-concept’ (items 2, 3, 7, 8, 17, 18, 21, 23, 24, 27), ‘confusion and escape fantasies’ (items 13,
14, 15, 19, 20, 22), ‘personal maladjustment and desire for change’ (9, 26, 29), ‘loneliness/isolation’ (1, 4, 10, 28), ‘giving up/helplessness’ (6, 11, 12, 25).
f
Factors and items loading on them included: ‘negative self-concept and negative expectations’ (items 2, 6, 7, 8, 10, 11, 12, 17, 18, 20, 21, 23, 24, 27, 28,
29, 30) and ‘personal maladjustment and desire for change’ (items 1, 4, 5, 9, 14, 15, 16, 19, 22, 26).
 GHASSEMZADEH ET AL.: PSYCHOMETRIC PROPERTIES OF ATQ-PERSIAN 135

on the English-language version of the test have also produced conflicting results. This may
be due to the cohesive latent factor structure of the test as suggested by the uniformly strong
item-total correlations for all test items, and as noted by other investigators (Bryant & Baxter,
1997). This conclusion is further supported by our observation that, in every case, a solution
with correlated factors considerably improved the model fit, suggesting that the underlying
latent structure of ATQ-Persian is composed of strongly correlated factors. Hence, no ortho-
gonal solutions would adequately depict this structure.
Another noteworthy finding from our analyses was that two models with correlated factors
provided a better fit to the data than models with one global factor based on the same items
(models by Sahin & Sahin [1992] and Joseph [1994]). This finding suggests that the items do
cluster into certain distinct (albeit highly correlated) factors. Thus, it seems unlikely that one
global factor could adequately explain the latent structure of ATQ-Persian. It is also of note
that while neither of the models provided a good fit to the data, the modified Deardorf et al.
(1984) model provided the closest fit. This finding is consistent with that of Bryant & Baxter
(1997) in their sample of US college students. These authors also proposed that the cognitive
phenomenon of ‘spreading activation’ may explain the high correlation among the under-
lying constructs, ‘whereby one type of thought triggers a host of related types of thoughts’
(p. 255). The phenomenon of ‘spreading activation’ is consistent with the ‘automatic’ nature
of these cognitions.
Another possible explanation for the highly correlated latent structure of the ATQ is simi-
larity in the content of items. Beck’s formulation of cognitive biases in depression includes
negative ideations about self, future and the world (Beck, 1967), whereas the ATQ items
mostly focus on negative ideations about self. Although aberrations in self-evaluation and
self-esteem are likely to be central to depression (Fennell, 1997; Roberts & Monroe, 1994),
assessment of thoughts concerning the future and the world may provide a somewhat differ-
ent picture of the latent structure of automatic thoughts in depression.
In summary, the results of this study suggest that ATQ-Persian is a reliable and valid
measure of negative cognitions. Future studies need to examine the validity of ATQ-Persian
for assessing the frequency of depressogenic thoughts in clinical samples and the utility of this
test as a process or an outcome measure in clinical settings.
This report is part of an ongoing study of negative cognitions associated with major
depression in Iranian patients. The ultimate goal of this study is to adapt and improve the
application of cognitive-behavioural therapy methods originally designed in US and Euro-
pean countries to the Iranian cultural setting. Characterising the negative ATs in Iranian
patients would help to define appropriate targets for such intervention.
Assuming that ATs are mostly the product of schematic processing, which, in turn, is
affected by cultural factors (D’Andrade, 1989), one could expect differences in the content
of ATs in depressed individuals belonging to different cultures. For example, the cultural
emphasis on professional accomplishments in US society versus family relationships and
obligations in Iranian society (Ghassemzadeh, 1993) might translate into different negative
ATs among depressed patients in the two cultures. These differences in turn might imply
different targets for cognitive-behavioural therapy. Availability of a measure of negative
ATs with comparable reliability and validity across the two settings would greatly facilitate
such comparative studies.

Downloaded from isp.sagepub.com by guest on January 16, 2015

136 INTERNATIONAL JOURNAL OF SOCIAL PSYCHIATRY 52(2)

ACKNOWLEDGEMENT

We thank Dr Steven Hollon for providing us with a copy of the Automatic Thoughts Ques-
tionnaire and for his encouragement, and Drs Fred Bryant and Jeff Baxter for providing us
with unpublished data from their study.

REFERENCES

AMERICAN PSYCHIATRIC ASSOCIATION (1994) Diagnostic and Statistical Manual of Mental Disorders
(4th edition). Washington, DC: American Psychiatric Association.
BECK, A.T. (1967) Depression – Causes and Treatment. Philadelphia, PA: University of Pennsylvania Press.
BECK, A.T. (1976) Cognitive Therapy and the Emotional Disorders. New York: International Universities
Press.
BECK, A., RUSH, A., SHAW, B. & EMERY, G. (1979) Cognitive Therapy of Depression. New York:
Guilford.
BECK, A.T., STEER, R.A. & BROWN, G.K. (1996) Manual for the Beck Depression Inventory-II. San
Antonio, TX: Psychological Corporation.
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.
BEST, J.B. (1992) Cognitive Psychology. St. Paul, MN: West Publishing Company.
BRYANT, F.B. & BAXTER, W.J. (1997) The structure of positive and negative automatic cognition.
Cognition & Emotion, 11, 225–258.
CHARLES, S.B.M., MOLLARD, E. & COTTRAUX, J. (1989) Validation and factorial structure of the
French version of the Automatic Thoughts Questionnaire. Poster presented at the World Congress of
Cognitive Therapy, Oxford, UK.
D’ANDRADE, R. (1989) Cultural cognition. In Foundations of Cognitive Science (ed. M.H. Posner, pp. 795–
830). Cambridge, MA: MIT Press.
DEARDORF, P.A., HOPKINS, L.R. & FINCH, A.J. (1984) Automatic Thoughts Questionnaire: a reliability
and validity study. Psychological Reports, 55, 708–710.
DOBSON, K.S. & SHAW, B.F. (1986) Cognitive assessment of major depressive disorders. Cognitive Therapy
and Research, 10, 13–29.
FENNELL, M.Y. (1997) Low self-esteem: a cognitive perspective. Behavioural and Cognitive Psychotherapy,
25, 1–25.
GHASSEMZADEH, H. (1993) Reflexiones sober el estudio de la personalidad en Iran. Boletin de Psicologia,
40, 25–35.
GLASS, C.R. & ARNKOFF, D.B. (1997) Questionnaire methods of cognitive self-statement assessment.
Journal of Consulting & Clinical Psychology, 65, 911–927.
HARRELL, T.H. & RYON, N.B. (1983) Cognitive-behavioral assessment of depression: clinical validation of
the Automatic Thoughts Questionnaire. Journal of Consulting and Clinical Psychology, 51, 721–725.
HILL, C.V., OEI, T.P.S. & HILL, M.A. (1989) An empirical investigation of the specificity and sensitivity of
the Automatic Thoughts Questionnaire and Dysfunctional Attitudes Scale. Journal of Psychopathology
and Behavioral Assessment, 11, 291–311.
HOLLON, S.D. & KENDALL, P.C. (1980) Cognitive self-statements in depression: development of an auto-
matic thoughts questionnaire. Cognitive Therapy and Research, 4, 383–395.
HOLLON, S.D., KENDALL, P.C. & LUMRY, A. (1986) Specificity of depressotypic cognitions in clinical
depression. Journal of Abnormal Psychology, 95, 52–59.
JOSEPH, S. (1994) Subscales of the Automatic Thoughts Questionnaire. Journal of Genetic Psychology, 155,
367–368.
LOEHLIN, J.C. (1992) Latent Variable Models (2nd edition). Hillsdale, NJ: Lawrence Erlbaum.
METZ, C.E. (1978) Basic principles of ROC analysis. Seminars in Nuclear Medicine, 8, 283–298.
NETEMEYER, R.G., WILLIAMSON, D.A., BURTON, S., BISWAS, D., JINDAL, S., LANDRETH, S.,
MILLS, G. & PRIMEAUX, S. (2002) Psychometric properties of shortened versions of the Automatic
Thoughts Questionnaire. Educational and Psychological Measurement, 62, 111–129.

Downloaded from isp.sagepub.com by guest on January 16, 2015

 GHASSEMZADEH ET AL.: PSYCHOMETRIC PROPERTIES OF ATQ-PERSIAN 137

PEDEN, A.R., HALL, L.A., RAYENS, M.K. & BEEBE, L. (2000) Negative thinking mediates the effect of
self-esteem in depressive symptoms in college women. Nursing Research, 49, 201–207.
REVUELTA PEREZ, F. (1995) Adaptacion al espanol del ‘Cuestionario de Auto-Manifestaciones anxiosas’
(Anxious Self-Statements Questionnaire (ASSQ)) de Kendall y Hollon (1989). Analisis y Modificacion de
Conducta, 21, 547–570.
ROBERTS, J.E. & MONROE, S.M. (1994) A multi-dimensional model of self-esteem in depression. Clinical
Psychology Review, 14, 161–181.
SAHIN, N.H. & SAHIN, N. (1992) Reliability and validity of the Turkish version of the Automatic Thoughts
Questionnaire. Journal of Clinical Psychology, 48, 334–340.
SHIFFRIN, R.M. & SCHNEIDER, W. (1977) Controlled and automatic human information processing: II.
Perceptual learning, automatic attending, and a general theory. Psychological Review, 84, 127–190.
STEIN, D.J. & YOUNG, J.E., eds (1992) Cognitive Science and Clinical Disorders. San Diego, CA: Academic
Press.
TEASDALE, J.D. & BARNARD, P. (1993) Affect, Cognition and Change. Hove: Erlbaum.

Habibollah Ghassemzadeh, Clinical Psychology Unit, Department of Psychiatry, Tehran University of Medical
Sciences, Tehran, Iran.
Ramin Mojtabai, Department of Psychiatry, Beth Israel Medical Center, New York, NY, USA.
Narges Karamghadiri, Clinical Psychology Unit, Department of Psychiatry, Tehran University of Medical Sciences,
Tehran, Iran.
Narges Ebrahimkhani, Clinical Psychology Unit, Department of Psychiatry, Tehran University of Medical Sciences,
Tehran, Iran.
Correspondence to H. Ghassemzadeh, PhD, Roozbeh Hospital, Kargar Ave., Tehran 13185, Iran.
Email: hghasemzadeh@yahoo.com

Downloaded from isp.sagepub.com by guest on January 16, 2015