Carefully go over the instructions below and the sample proposals posted on Canvas.You can use the sample proposals and the examples posted in this document to guide you. However, write your study/pro

~ Pergamon 0005-7967(94)00075-1

BehaP. Res. Ther.

THE STRUCTURE OF NEGATIVE EMOTIONAL STATES:

COMPARISON OF THE DEPRESSION ANXIETY STRESS

SCALES (DASS) WITH THE BECK DEPRESSION AND

ANXIETY INVENTORIES

P. F.

LOVIBOND and

S. H.

LOVIBOND

School of Psychology, University of New South Wales, Sydney, N.S.W. 2052, Australia

(Received 10 September 1993; accepted 22 September 1994)

Summary--The psychometric properties of the Depression Anxiety Stress Scales (DASS) were evaluated in a normal sample of N = 717 who were also administered the Beck Depression Inventory (BDI) and the Beck Anxiety Inventory (BAI). The DASS was shown to possess satisfactory psychometric properties, and the factor structure was substantiated both by exploratory and confirmatory factor analysis. In comparison to the BDI and BAI, the DASS scales showed greater separation in factor loadings. The DASS Anxiety scale correlated 0.81 with the BAI, and the DASS Depression scale correlated 0.74 with the BDI. Factor analyses suggested that the BDI differs from the DASS Depression scale primarily in that the BDI includes items such as weight loss, insomnia, somatic preoccupation and irritability, which fail to discriminate between depression and other affective states. The factor structure of the combined BDI and BAI items was virtually identical to that reported by Beck for a sample of diagnosed depressed and anxious patients, supporting the view that these clinical states are more severe expressions of the same states that may be discerned in normals. Implications of the results for the conceptualisation of depression, anxiety and tension/stress are considered, and the utility of the DASS scales in discriminating between these constructs is discussed.

INTRODUCTION

Historically, the relationship between the negative affective conditions of depression and anxiety

has been of considerable theoretical and clinical interest (Akiskal, 1985; Clark, 1989; Clark &

Watson, 1990; Dobson, 1985; Stavrakaki & Vargo, 1986; Watson, Clark & Carey, 1988).

Conceptually, depression and anxiety are quite distinct, but the clinical overlap between the two

conditions has long exercised both clinicians and researchers. The concept of stress poses additional

problems in the study of negative affective conditions. In addition to precipitating episodes of

anxiety and depression, stressful life events are often thought to lead to a characteristic stress

response involving chronic arousal and impaired function (e.g. Selye, 1952). Considered as an

affective or emotional state (e.g. Lazarus, 1993), the concept of a stress response has clear affinities

with anxiety.

Recently Gotlib and Cane (1989) emphasised the limitations of existing self-report scales for

anxiety and depression, and Clark and Watson (1990) have summarised the evidence relating to

the assessment of the two conditions. The essential findings are as follows: (a) Self-report anxiety

and depression scales typically correlate between 0.40 and 0.70 across a wide range of patient and

non-patient samples; (b) Anxiety scales frequently correlate as highly with depression scales as with

other anxiety scales, and depression scales show equal lack of specificity; (c) Clinicians' ratings of

anxiety and depression duplicate the relationships found with self-report scales; and (d) Only about

half the patients diagnosed as having a depressive or an anxiety disorder exhibit relatively pure

syndromes of one type or the other.

The two most ambitious research programs designed to differentiate anxiety and depression were

conducted by Costelio and Comrey (1967) and Beck, Epstein, Brown and Steer (1988). Costello

and Comrey began with a large pool of items derived for the most part from existing anxiety and

depression scales. The items were administered to successive clinical samples, and iterative factor

analyses were performed in the search for items that would define orthogonal factors of anxiety

BRT 33/3--fj 335

336 P.F. Lovibond and S. H. Lovibond

and depression. In the case of the Beck

et al.

research, a widely used depression scale, the Beck

Depression Inventory (BDI), was already available (Beck, Ward, Mendelson, Mock & Erbaugh,

1961; Beck, Rush, Shaw & Emery, 1979). The task was to develop an anxiety scale that would

provide maximum discrimination from the BDI and other measures of depression. An initial pool

of items was drawn from pre-existing scales, and was refined on the basis of factor analyses of the

responses of a series of outpatient samples. The final 21-item anxiety scale (the Beck Anxiety

Inventory; BAI) was found to correlate in the region of 0.50 with the BDI. It is of interest that

the final form of the anxiety and depression scales of Costello and Comrey (1967) also correlated

in the region of 0.50, despite the stated goal of the authors to derive orthogonal anxiety and

depression factors.

The present paper focuses on a third research program which aimed to develop self-report

anxiety and depression scales that would (a) cover the full range of core symptoms of anxiety and

depression, (b) meet high psychometric standards, and (c) provide maximum discrimination

between the two scales. The research was carried out over the period 1979-1990 and data were

obtained from 30 samples. During testing of the depression and anxiety scales, a new factor

emerged from analysis of the non-discriminating anxiety and depression items, with the highest

loading items referring to difficulty relaxing, nervous tension, irritability and agitation. Further

items were tested to explore the limits of this factor, leading to the development of a new scale

which was labelled 'stress' by virtue of perceived similarity of the items to the symptoms of tension

or stress (cf. Selye, 1952, 1974). The resulting scales, now referred to as the Depression Anxiety

Stress Scales (DASS), were first described by Lovibond (1983; see also Wilson, 1980, 1982). A

detailed account of the scale development, normative data, and research applications is provided

in the DASS manual (Lovibond & Lovibond, 1993), available from the authors. The DASS items

are listed in Table 3.

The DASS research strategy differed in several ways from both previous studies (Costello &

Comrey, 1967) and subsequent studies (Beck

et al.,

1988). First, the scales were developed using

a boot-strapping strategy in which factors were defined initially in terms of clinical consensus, but

were refined empirically using a confirmatory factor analytic technique, multiple groups factor

analysis (Harman, 1976). Items were added as well as deleted over successive samples in an iterative

procedure which led to the emergence of the stress factor (cf. Tellegen, 1985, pp. 685-688). Second,

because of the overlap and inconsistencies between existing anxiety and depression scales, and

between the various diagnostic systems for anxiety and depression, no external criteria were used

in the development of the DASS scales. In contrast, Beck

et al.

(1988) selected items in part on

the basis of their relationship to

DSM-III

(American Psychiatric Association, 1980) diagnostic

categories for anxiety and depression. Finally, the major development of the DASS scales was

carried out with normal, non-clinical samples. Thus, the central aim underlying development of

the DASS scales was to generate measures of general negative affective syndromes, guided by

existing conceptions but ultimately determined on empirical grounds.

The present study was designed to evaluate the psychometric properties of the DASS scales in

an independent sample of normal Ss. In particular, it was considered important to assess the

generality of the factor structure derived from multiple groups factor analysis by using conventional

exploratory and confirmatory factor analyses. Inclusion of the Beck scales (BDI and BAI) provided

an interesting comparison in that these scales were developed with a different research strategy but

with similar aims to the DASS scales. Furthermore, the factor structure obtained for the Beck scales

in the present normal sample could be compared with the factor structure reported by Beck

et al.

(1988) in a clinical sample, thus providing information about the degree of convergence between

the emotional states experienced by normals and by depressed and anxious patients.

METHOD

Subjects

The Ss were 717 (486 female and 231 male) first year psychology students at the University of

New South Wales. The mean age was 21.0 years.

Negative emotional states 337

Table 1. Means, standard deviations and alpha coefficients Table 2. Intercorrelations between the DASS scales, the BDI, and the for the DASS scales, the BDI, and the BAI BAI

Mean SD Alpha DASS

DASS Depression 7.19 6.54 0.91 Depression Anxiety Stress BDI BAI

Anxiety 5.23 4.83 0.81 DASS anxiety 0.54 -- Stress 10.54 6.94 0.89 DASS stress 0.56 0.65 -- BDI 7.72 6.47 0.84 BDI 0.74 0.58 0.60 -- BAI 9.15 7.41 0.88 BAi 0.54 0.81 0.64 0.59 --

Measures

The DASS (Lovibond & Lovibond, 1993) consists of 42 negative emotional symptoms (see

Table 3). Ss rate the extent to which they have experienced each symptom over the past week, on

a 4-point severity/frequency scale. Scores for the Depression, Anxiety and Stress scales are

determined by summing the scores for the relevant 14 items. Internal consistencies (coefficient

alpha) for each scale for the DASS normative sample were: Depression 0.91; Anxiety 0.84; Stress

0.90.

The revised BDI (Beck & Steer, 1987) is a 21-item self-report questionnaire in which each item

consists of four statements indicating different levels of severity of a particular symptom

experienced over the past week. Scores for all 21 items are summed to yield a single depression

score. The internal consistency of the BDI, based on a number of clinical samples, is 0.86 (Beck

& Steer, 1987). The BAI (Beck & Steer, 1990) consists of 21 symptoms that are rated on a 4-point

severity scale referring to experience of symptoms over the past week. Scores for the 21 items are

summed to yield a single anxiety score. The internal consistency of the BAI over a number of

samples has been found to be in the range 0.85 to 0.94 (Beck & Steer, 1990).

Procedure

The students completed the DASS, BAI and BDI, in that order, during tutorial classes. Ratings

were made in pencil on answer sheets that were read by an optical scanner. The data were analysed

by means of principal components and confirmatory factor analyses. A criterion of ~ = 0.05 was

used in significance tests.

RESULTS

Descriptive statistics

Means, standard deviations and alpha coefficients for the five scales are shown in Table 1.

Inter-correlations between the scales are shown in Table 2. Initial analyses indicated a similar factor

structure for males and females, so the factor analyses reported below are all based on the full

sample.

Principal components factor analysis

The first approach used to test the factor structure of the DASS, which had been developed using

multiple groups factor analysis, was a principal components analysis. Three factors were specified,

in order to allow factors corresponding to the three scales to emerge, if supported by the data.*

The first three factors together accounted for 41.3% of the item variance. Oblique rotation yielded

three correlated factors that reproduced the three DASS scales, with one exception: Anxiety item

10 ("I feared that I would be 'thrown' by some trivial but unfamiliar task") loaded more highly

on the Stress factor than on the Anxiety factor. In general, most items received a moderate to high

loading on their own factor and low loadings on the other two factors (see Table 3). The

correlations between factors were: Depression-Anxiety r = 0.42; Anxiety-Stress r = 0.46; and

Depression-Stress r = 0.39.

*When two factors were specified, the anxiety and stress factors collapsed into one. When four factors were specified, the items describing impatience in the stress scale formed a separate factor, and when five factors were specified, the anxiety scale split into two factors, corrresponding to autonomic and subjective symptoms.

338 P.F. Lovibond and S. H. Lovibond

For comparison purposes, a similar principal components analysis, with two factors specified,

was carried out on the BDI and BAI items. The first two factors accounted for 22.5% and 6.6%

of the variance. Rotation yielded two factors corresponding to Depression and Anxiety, with three

exceptions: BAI items 5 ("Fear of the worst happening") and 14 ("Fear of losing control") loaded

more highly on the Depression factor than on the Anxiety factor, and BDI item 19 (Loss of weight)

loaded more highly on the Anxiety factor than on the Depression factor. The correlation between

the two factors was r = 0.47. In comparison to the DASS analysis, there was a greater degree of

overlap in the loadings of the Depression and Anxiety factors in the BDI/BAI analysis. Several

BAI items loaded moderately on both factors. The BDI items did not tend to load strongly on the

Anxiety factor, but several BDI items had low loadings on the Depression factor. The BDI items

with loadings below 0.3 were item 19 (Weight loss: 0.0), item 21 (Loss of libido: 0.18), item 11

(Irritability: 0.27), item 18 (Loss of appetite: 0.27) and item 20 (Somatic preoccupation: 0.27).

Confirmatory factor analysis

The second approach to testing the DASS factor structure was confirmatory factor analysis, The

statistical program Lisrel7 (Joreskog & Sorbom, 1988) was used to test the adequacy of the

allocation of items to the three DASS scales. The first model tested was a single factor model, which

yielded a very large and significant chi-square value \[Zz(819)= 5413, P <0.05\], indicating a

significant discrepancy between the model and the data. The adjusted goodness of fit index was

0.60. Next, a two-factor model was tested in order to assess the validity of the distinction between

Depression and the other two DASS scales. This model yielded an improved fit \[X2(818)= 3942,

P < 0.05; adjusted goodness of fit---0.74\], and differed significantly from the one-factor model

\[z2(l) = 1471, P < 0.05\]. Finally, three factors were defined, corresponding to the three DASS

scales. This model yielded a lower chi-square value again \[X2(816)= 3559, P <0.05; adjusted

goodness of fit = 0.76\]. The phi coefficients, which assess the strength of the links between the three

factors, were: Depression-Anxiety 0.61; Anxiety-Stress 0.76; Depression-Stress 0.62. The three-fac-

tor model provided a significantly better fit than the two-factor model \[Z2(2)= 383, P < 0.05\].

These comparisons indicate that distinguishing between depression and the other two scales yields

a large improvement in fit to the data, and that distinguishing between anxiety and stress yields

a further moderate improvement in fit.

In order to provide a reference point to evaluate the goodness of fit of the three-factor

confirmatory model, a three-factor exploratory factor analysis was carried out using the same

estimation method as the confirmatory analysis: maximum likelihood. This analysis also generated

a large chi-square value \[Xz(738)= 3025, P < 0.05\]. This value represents the lowest possible

chi-square that could be obtained for a three-factor solution, where all items are free to load on

all factors. By comparison, therefore, the chi-square value of 3559 for the highly constrained

three-factor confirmatory analysis is relatively low, particularly in comparison to the one-factor and

two-factor values of 5413 and 3942, respectively.

A final analysis addressed the issue of how much of the variance in the three scales is due to

a common factor. A second-order factor analysis was carried out, in which a common factor was

allowed to influence all three scales, but the three scales were not allowed to influence each other.

This model generated an identical fit and identical item loadings to the three-factor confirmatory

analysis described above \[zz(816) = 3559, P < 0.05; adjusted goodness of fit = 0.76\]. The gamma

coefficients, which represent the strength of the links between the common factor and each of the

scales, were: Depression 0.71; Anxiety 0.86; Stress 0.88. That is, the common factor accounted for

50.4% of the variance in Depression, 74.0% of the variance in Anxiety, and 77.4% of the variance

in Stress. The larger values for Anxiety and Stress reflect the fact that these two scales are more

highly correlated, and therefore dominated the definition of the common factor.

DASS subscale analysis

During development of the DASS, items in each scale were categorised into subscales of 2-6

items, on the basis of judgement of common content. In order to test the validity of the subscales,

a second-order factor analysis model was tested using Lisrel. This model defined 16 lower-level

factors corresponding to the subscales. Each lower-level factor was allowed to contribute to the

relevant higher-order factor corresponding to the three scales, Depression, Anxiety and Stress. The

Negative emotional states 339

Table 3. List of DASS items with factor Ioadings from 3-factor principal components analysis with oblique rotation

SCALE

Subscale

Factor Ioadings Item I 2 3

DEPRESSION

Dysphoria

I felt downhearted and blue* 1 felt sad and depressed

Hopelessness

1 could see nothing in the future to be hopeful about I felt that I had nothing to look forward to*

Devaluation q\[' I(\[b

1 felt that life was meaningless* 1 felt that life wasn't worthwhile

SelJ:deprecation

I felt I was pretty worthless I felt I wasn't worth much as a person*

Lack of interest/involvement

I felt that I had lost interest in just about everything I was unable to become enthusiastic about anything*

Anhedonia

I couldn't seem to experience any positive feeling at all* I couldn't seem to get any enjoyment out of the things I did

Inertia

I just couldn't seem to get going 1 found it difficult to work up the initiative to do things*

ANXIETY

Autonomic arousal

1 was aware of the action of my heart in the absence of physical exertion (e.g, sense of heart rate increase, heart missing a beat)* I perspired noticeably (e.g. hands sweaty) in the absence of high temperatures or physical exertion 1 was aware of dryness of my mouth* 1 experienced breathing difficulty (e.g. excessively rapid breathing, breathlessness in the absence of physical exertion)* I had difficulty in swallowing

Skeletal musculature effects

I had a feeling of shakiness (e.g. legs going to give way) I experienced trembling (e.g. in the hands)*

Situational anxiety

I was worried about situations in which I might panic and make a fool of myself* 1 found myself in situations which made me so anxious I was most relieved when they ended I feared that I would be "thrown" by some trivial but unfamiliar task

Subjective experience of anxious affect

I felt 1 was close to panic* I felt terrified I felt scared without any good reason* I had a feeling of faintness

STRESS

Difficulty relaxing

I found it hard to wind down* I found it hard to calm down after something upset me I found it difficult to relax*

Nervous arousal

I felt that I was using a lot of nervous energy* I was in a state of nervous tension

Easily upset/agitated

I found myself getting upset rather easily I found myself getting upset by quite trivial things I found myself getting agitated*

Irritable/over-reactive

1 tended to over-react to situations* 1 found that I was very irritable 1 felt that I was rather touchy*

Impa t ien t

I was intolerant of anything that kept me from getting on with what I was doing* I found myself getting impatient when I was delayed in any way (e.g. lifts, traffic lights, being kept waiting) 1 found it difficult to tolerate interruptions to what I was doing

57 56 22

75 80

76 73

69 72

67 67

74 60

36 45

34 47 53

61 64

49 46 24 20 27

53 48 22 48 20 52

48 20 46 27 47

34 41 38 40

61 58 63

62 60 67

73 56 76

Note I: Decimal points and values less than 0.2 in factor Ioadings omitted for clarity. Note 2: The DASS scales are in the public domain. The above items are presented to Ss in random order, with a four-point scale for each item labelled "'Did not apply to me at all'" (0), "'Applied to me to some degree, or some of the time" (1), "'Applied to me a considerable degree, or a good part of the time (2), and "Applied to me very much, or most of the time" (3). Instructions at the top of the sheet read: "For each of the statements below, please circle the number which best indicates how much the statement applied to you OVER THE PAST WEEK. There are no right or wrong answers. Do not spend too much time on any one statement." Scores for each scale are obtained by summing the scores for the 14 items in the scale. Note 3: Items which are included in the short (21-item) version are marked with an asterisk. The items were selected such that scale scores for the short version may be converted to full scale scores by multiplying by 2. Note 4: The DASS is a minor revision of an earlier version of the scales, the SAQ (Lovibond, 1983), which contained 14 Depression items, 12 Anxiety items and 16 Stress items. SAQ scores may be converted to DASS scores by multiplying the Anxiety score by 1.037 and the Stress score by 0.921. The Depression scales are identical.

340 P.F. Lovibond and S. H. Lovibond

first attempt to test this model failed, as the Lisrel program could not generate a solution (psi matrix

not positive definite). Examination of the standard errors suggested that Anxiety item 14 (faintness)

was problematic. Reassignment of this item from the Autonomic to the Subjective subscale allowed

a solution to be generated. This solution produced a better fit than the first-order models

L~2(801) = 2399, P < 0.05; adjusted goodness of fit = 0.84\]. No further adjustments were made to

item assignment, due to the likelihood of sample-specific artefacts (MacCallum, Roznowski &

Necowitz, 1992). The final allocation of items to subscales is listed in Table 3. Since this allocation

was based primarily on logical considerations, it cannot be defended as an optimal solution, since

other groupings may have yielded similar or superior fit to the data. However, the subscales provide

a means of reducing the information contained in the 42 items, and may be useful for descriptive

and research purposes.

The relative contribution of each subscale to the three scales was examined to see whether the

various types of symptoms tended to be associated with particular levels of severity. Mean subscale

scores were calculated as a function

the Inertia and Dysphoria subscales

at low levels of Depression than the

a full scale score of approximately

of full scale score, for each scale. In the case of Depression,

had the highest mean values and were also more responsive

other subscales, which did not show substantial scores until

6. However, whereas the Dysphoria subscale continued to

increase as the full scale Depression score increased, the Inertia subscale showed a slower rate of

increase past a full scale score of 8. In the case of Anxiety, the Situational subscale had a higher

mean than the other subscales, but all subscales discriminated across the full range of Anxiety scale

scores. Finally, all of the Stress subscales showed similar means and discriminated across the full

range of Stress scale scores.

BDI and BAI factor analyses

Two further factor analyses were carried out on the present data to replicate directly those

reported by Beck et al. (1988),* which were based on a clinical sample of N = 160. Firstly, a

four-factor principal factor analysis with orthogonal rotation was conducted on the 42 BDI and

BAI items. The first three factors generated were virtually identical to those obtained by Beck: a

depression factor, a somatic anxiety factor, and a subjective anxiety factor. The fourth factor was

a somatic depression factor (e.g. loss of appetite, insomnia, somatic preoccupation), whereas in

Beck's analysis this factor contained a mixture of BDI and BAI somatic items. The second

replication analysis was an oblique two-factor principal factor analysis of the 21 BAI items. This

analysis yielded a somatic factor (14 items) and a subjective factor (7 items), with identical item

composition in each case to that reported by Beck et al. (1988). These results indicate that the factor

structure of the BDI and BAI is essentially the same in a normal sample as in a clinical population

with diagnosed depressive and anxiety disorders.

DISCUSSION

Psychometric data

This study yielded means, standard deviations and intercorrelations for the DASS scales similar

to those observed in previous samples (Lovibond & Lovibond, 1993), and confirmed the

satisfactory reliability of the three scales. The factor structure of the DASS was confirmed with

two different approaches. First, principal components factor analysis reproduced the three scales,

with the exception of one Anxiety item that performed more weakly than in previous data sets.

Second, confirmatory factor analysis indicated that the three scales provided a better fit to the data

than either a one-factor or a two-factor solution, and yielded a chi-square value close to that of

a completely unrestricted analysis. The analyses confirmed that while the DASS successfully

discriminates between three negative emotional syndromes, these syndromes are still moderately

highly correlated with each other, and in particular the Stress scale is more closely associated with

Anxiety than with Depression.

*We thank A. T. Beck for providing an earlier draft of this paper containing additional details of the analyses conducted.

Negative emotional states 341

Relationship between the DASS, BDI and BAI

The DASS Anxiety scale and the BAI were highly correlated (r =0.81), while the DASS

Depression scale and the BDI were somewhat less strongly correlated (r = 0.74). However, both

of these correlations were much higher than the corresponding cross-correlations of r = 0.58 and

r = 0.54 (respectively, Z = 12.41 and Z = 9.25, P < 0.05), indicating a greater degree of convergent

validity than is typically the case for self-report scales (Clark & Watson, 1990). The primary reason

for the lower correlation between the DASS Depression scale and the BDI appears to be the

inclusion in the BDI of several items which are not strongly or uniquely related to depression. In

particular, weight loss, loss of libido, irritability, loss of appetite and somatic preoccupation

received low loadings on the Depression factor in the analysis of the combined BDI and BAI items,

replicating the analysis reported by Beck et al. (1988). This difference is also reflected in the lower

alpha coefficient for the BDI (0.84) compared to the DASS Depression scale (0.91), despite the

larger number of items in the BDI (21 versus 14).

It is of interest that many of the BDI items with low factor loadings were among those that were

tested during the development of the DASS, but had been rejected since they failed to discriminate

between the Depression scale and the other two scales. These items included disturbance of

appetite/weight loss, sleeping difficulties, tiredness for no reason, lack of energy, poor concen-

tration, indecisiveness, agitation, guilt, lack of interest in sex, early awakening, feeling worst in the

morning, mood changes during the day, crying, restlessness, and irritability. Some of these items

were in fact more closely associated with the DASS Stress scale and form part of the final version

of that scale. It should be noted that the first eight of the above symptoms are included among

the criteria for diagnosis of major depression listed in the Diagnostic and Statistical Manual

(DSM-III-R) of the American Psychiatric Association (1987).

The implication of the present data, that several traditional depressive symptoms are only poorly

associated with other components of the depressive syndrome, is of course a contentious one. In

particular, clinicians have long regarded somatic symptoms as core features of depression.

However, the present findings are supported by previous research on the discriminant validity of

individual symptoms. In a review of studies of both clinician-rated and self-reported symptoms of

depression ands anxiety, Clark (1989, Table 4.3) reported that four symptoms (loss of sexual

interest, loss of appetite, overeating/weight gain, and increased sleep) were weak or inconsistent

discriminators between patients diagnosed with depressive and anxiety disorders. Thus it would

appear that current conceptions of depression, reflected both in popular self-report instruments

such as the BDI and in formal diagnostic systems, may include symptoms that are poorly associated

with the core depressive syndrome, or that are not unique to depression. By contrast, these

symptoms were excluded from the DASS Depression scale during scale development, where

discrimination between the three scales was a primary criterion for symptom inclusion. Further

research is necessary with clinical samples to determine systematically which symptoms are best

related to the overall diagnosis of depression, and which discriminate depression from other

diagnoses such as anxiety disorders.

Normal versus clinical emotional states

In the present normal sample, the factor structure obtained from factor analysis of the BDI and

BAI was highly similar to that reported by Beck et al. (1988) for these instruments on a clinical

sample. Similarly, in the development of the DASS, not only the factor structure but also the

relative performance of individual items was found to be virtually the same in clinical and

non-clinical samples. Furthermore, in the present study, the subscale analysis indicated that the

contribution of the various symptoms within each syndrome was relatively constant across severity.

All of these findings strongly suggest that the depression, anxiety and tension/stress manifested by

non-psychotic clinical outpatients and by normal non-clinical groups differ primarily in severity.

This conclusion does not, of course, deny the possibility that specific clinical sub-categories may

be distinguished within these general affective disturbances. However, the results do add to evidence

suggesting that emotional disorders fall on a continuum with less extreme emotional disturbance

(e.g. Farmer & McGuffin, 1989; Vredenburg, Flett & Krames, 1993). That is, clinical disorders may

342 P.F. Lovibond and S. H. Lovibond

represent the severe, inappropriate or chronic manifestation of syndromes whose essential structure

may be discerned in normal Ss.

The nature and differentiation of depression, anxiety and stress

The present study confirms that three syndromes, labelled depression, anxiety and stress, may

be distinguished from self-report data by the DASS scales. Descriptively, the Depression scale is

characterised principally by a loss of self-esteem and incentive, and is associated with a low

perceived probability of attaining life goals of significance for the individual as a person. The

coherence of these symptoms supports the view that the depressive syndrome is considerably

broader than sadness of mood, although as discussed above, several symptoms traditionally

regarded as part of this syndrome are not represented in the DASS Depression scale, since they

do not appear to be specific to depression.

The DASS Anxiety scale emphasises the links between the relatively enduring state of anxiety

and the acute response of fear. It is of interest that the conventionally derived anxiety scale of

Costello & Comrey (1967) does not share this feature. Indeed, the content of the Costello & Comrey

anxiety scale is very narrow, and the scale appears to tap a dimension that might be termed

'nervousness'. By contrast, the BAI, which was constructed by a multi-stage procedure emphasising

discrimination from depression at every stage, shares with the DASS Anxiety scale an emphasis

on fear-related symptoms. Both scales give weight to somatic and subjective symptoms (cf. Morris,

Davis & Hutchins, 1981), and the DASS Anxiety scale additionally addresses situational anxiety.

The content of the Stress scale suggests that it is measuring a state of persistent arousal and

tension with a low threshold for becoming upset or frustrated. The items that showed the greatest

overlap in factor loadings with the Anxiety scale (see Table 3) were those involving nervous tension

and nervous energy. This overlap suggests that there is a natural continuity between the syndromes

assessed by the Anxiety and Stress scales, and that the point of division between the two may be

somewhat arbitrary. Nonetheless, the analyses confirmed that the Stress scale as a whole contains

a coherent set of symptoms that may be differentiated from depression and anxiety. The existence

of such a syndrome has important implications for any attempt to provide a full analysis of negative

emotional states. However, the present results do not comment on the appropriateness of labelling

the scale as measuring 'stress'. Further research is clearly necessary to clarify the external validity

of this scale, in particular its relationship to constructs in stress research such as life events,

appraisal and coping (Coyne & Downey, 1991), and its discriminant validity from anxiety.

A central theoretical question that arises in the present research is the basis of the measured

association between depression, anxiety and tension/stress. Our findings concur with those of other

investigators who have expended considerable effort in attempts to obtain the maximum meaning-

ful discriminations between measures of anxiety and depression. Indeed, the findings of Costelio

and Comrey (1967) and Beck

et al.

(1988), together with those of the present study, suggest that

-t-0.50 may well be an irreducible minimum correlation between self-report scales designed to

measure depression and anxiety. In the case of the DASS scales, no item was retained which loaded

substantially (greater than 0.25) on the other scales. The absence of overlapping items is largely

confirmed by the present data set. Nonetheless, the intercorrelations between the scales are still

moderately high. It therefore may be argued that the associations between the DASS scales are

not the result of the scales measuring

overlapping constructs.

Rather, these correlations may reflect

common causes

of anxiety, depression and stress (cf. the correlation between height and weight).

There would seem to be two possible types of common cause of negative affective states: a common

vulnerability factor which influences all three states, such as neuroticism (Eysenck & Eysenck, 1964)

or negative affectivity (Watson & Clark, 1984), and common environmental activation. We are

currently engaged in research aimed at identifying the common and specific factors, both trait and

environmental, that underlie the states of depression, anxiety and stress.

In summary, the present research provides support for the psychometric properties of the DASS

scales and their convergent and discriminant validity with other instruments developed on clinical

populations. The DASS scales not only provide measures of anxiety and depression that have been

specifically designed to maximise internal consistency and differentiation, but both of these scales

are further differentiated from the related state of tension/stress. The capacity to separately measure

these three related states may be of considerable use for researchers dealing with the complex links

Negative emotional states 343

between environmental demands and emotional and physical disturbance. The scales may also be

useful to the clinician in clarifying the locus of emotional disturbance, as part of the broader task

of clinical assessment (cf. Kendall, Hollon, Beck, Hammen & Ingram, 1987).

Acknowledgements--This research was supported by Australian Research Council grants A28316103, A78831924 and A79131809.

REFERENCES

American Psychiatric Association (1980). Diagnostic and statistical manual of mental disorders (3rd edn). Author: Washington, D.C. American Psychiatric Association (1987). Diagnostic and statistical manual of mental disorders (3rd edm rev). Author: Washington, D.C. Akiskal, H. S. (1985). Anxiety: Definition, relationship to depression and proposal for an integrative model. In Tuma, A. H. & Maser, J. D. (Eds), Anxiety and the anxiety disorders. Erlbaum: Hillsdale, N. J. Beck, A. T., Epstein, N., Brown, G. & Steer, R. A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56, 893-897. Beck, A. T., Rush, A. J., Shaw, B. F. & Emery, G. (1979). Cognitive therapy of depression. Guilford Press: New York. Beck, A. T. & Steer, R. A. (1987). Manual for the revised Beck Depression Inventory. The Psychological Corporation: San Antonio, Tex. Beck, A. T. & Steer, R. A. (1990). Manual for the Beck Anxiety Inventory. The Psychological Corporation: San Antonio, Tex. 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. Clark, L. A. (1989). The anxiety and depressive disorders: Descriptive psychopathology and differential diagnosis. In Kendall, P. C. & Watson, D. (Eds), Anxiety and Depression: Distinctive and overlapping features. Academic Press: San Diego, Calif. Clark, L. A. & Watson, D. (1990). Theoretical and empirical issues in differentiating depression from anxiety. In Becker, J. & Kleinman, A. (Eds), Advances in mood disorders, Vol. 1. Psychosocial aspects of depression. Erlbaum, Hillsdale, N.J. Costello, C. G. & Comrey, A. L. (1967). Scales for measuring anxiety and depression. Journal of Psychology, 66, 303 313. Coyne, J. C. & Downey, G. (1991). Social factors and psychopathology: Stress, social support, and coping processes. Annual Review of Psychology, 42, 401-425. Dobson, J. S. (1985). The relationship between anxiety and depression. Clinical Psychology Review, 5, 307-324. Eysenck, H. J. & Eysenck, S. B. G. (1964). Manual of the Eysenck Personality Inventor),. University of London Press: London. Farmer, A. & McGuffin, P. (1989). The classification of the depressions: Contemporary confusion revisited. British Journal of Psychiatry, 155, 437-443. Gotlib, 1. H. & Cane, D. B. (1989). In Kendall, P. C. & Watson, D. (Eds), Anxiety and depression; Distinctive and oterlapping features. Academic Press: San Diego. Harman, H. H. (1976). Modern .factor analysis. Chicago University Press: Chicago. Joreskog, K. G. & Sorbom, D. G. (1988). LISREL 7: A guide to the program and applications. SPSS: Chicago. Kendall, P. C., Hollon, S. D., Beck, A. T., Hammen, C. L. & Ingram, R. E. (1987). Issues and recommendations regarding use of the Beck Depression Inventory. Cognitive Therapy and Research, 11, 289-299. Lazarus, R. S. (1993). From psychological stress to the emotions: A history of changing outlooks. Annual Review of Psychology, 44, 1-21. Lovibond, S. H. (1983). The nature and measurement of anxiety, stress and depression. Paper presented at the 18th Annual Conference of the Australian Psychological Society, University of Western Australia, May. Lovibond, S. H. & Lovibond, P. F. (1993). Manual for the Depression Anxiety Stress Scales (DASS). Psychology Foundation Monograph. (Available from The Psychology Foundation, Room 1005 Mathews Building, University of New South Wales, NSW 2052, Australia). MacCallum, R. C., Roznowski, M. & Necowitz, L. B. (1992). Model modifications in covariance structure analysis: The problem of capitalization on chance. Psychological Bulletin, 111, 490-504. Morris, L. W., Davis, M. A. & Hutchins, C. H. (1981). Cognitive and emotional components of anxiety: Literature review and a revised worry-emotionality scale. Journal of Educational Psychology, 73, 541-555. Selye, H. (1952). The story of the adaptational syndrome. Acta Inc.: Montreal. Selye, H. (1974). Stress without distress. Lippincott: New York. Stavrakaki, C. & Vargo, B. (1986). The relationship of anxiety and depression: A review of the literature. British Journal of Psychiatry, 149, 7-16. Tellegen, A. (1985). Structures of mood and personality and their relevance to assessing anxiety, with an emphasis on self-report. In Tuma, A. H. & Maser, J. (Eds), Anxiety and the Anxiety Disorders (pp. 681-706). Erlbaum: Hillsdale, N.J. Vredenburg, K., Flett, G. L. & Krames, L. (1993). Analogue versus clinical depression: A critical reappraisal. Psychological Bulletin, 113, 327-344. Watson, D. & Clark, L. A. (1984). Negative Affectivity: The disposition to experience aversive emotional states. Psychological Bulletin, 96, 465-490. Watson, D., Clark, L. A. & Carey, G. (1988). Positive and negative affectivity and their relation to anxiety and depressive disorders. Journal of Abnormal Psychology, 97, 346-353. Wilson, P. H. (1980). Behavioural and pharmacological treatment of depression (Doctoral dissertation, University of New South Wales, Australia). Dissertation Abstracts International 41, 4063. Wilson, P. H. (1982). Combined pharmacological and behavioural treatment of depression. Behaviour Research and Therapy, 20, 173-184.