ORIGINAL ARTICLES

Cross-Cultural Validation of the DMI-10 Measure of State
Depression
The Development of a Chinese Language Version
Bibiana Chan, BAppSc, MA,* Gordon Parker, MD, PhD, DSc, FRANZCP,*
Lucy Tully, BA (Hons) Psych,* and Maurice Eisenbruch, MD, MPhil†

Abstract: Depression measurement tools in cross-cultural research
require careful design and thorough validation to ensure that cognitive concepts in one culture can be appropriately translated and
applied to a differing culture. The aim of this study was to validate
the Chinese version of a screening measure of state depression, the
10-item Depression in Medically Ill (DMI-10), and we report three
interdependent studies. An initial bilingual test-retest study identified four (of the 10) items as having poor cross-cultural validity. A
second study involved focus groups participants exploring the meaning of translated items with Chinese speakers. The third study
repeated the bilingual test-retest analyses on the modified DMI-10
form and demonstrated improved correlation coefficients on all
items and an excellent overall correlation (r ⫽ 0.87) between the
Chinese and English versions. The Chinese DMI-10 should prove
useful as a tool in cross-cultural research to understand the Chinese
experience of depression. The findings of this study have methodological implications for cross-cultural research on depression.
Key Words: Depression, culture, Chinese, validation,
measurement.
(J Nerv Ment Dis 2007;195: 20 –25)

A

re human emotions universal across cultures? Anthropologists have long been interested in the study of specific
emotions and expression of distress in different cultures
(Levy, 1984; Lutz and White, 1986; Rosaldo, 1983). Crosscultural studies of mental disorders, such as depression, are
important to challenge the basic assumptions of existing
psychiatric theory and practices, and examine the relevance
of transposing Western concepts of mental illness into nonWestern cultures (Kirmayer and Minas, 2004). The Chinese
make up the planet’s largest ethnic group with individuals
born in different regions, and are also heterogeneous in terms

*School of Psychiatry, University of New South Wales; and Black Dog
Institute, Sydney, Australia; and †Institute for Health and Diversity,
Victoria University, Melbourne, Victoria, Australia.
Send reprint requests to Bibiana Chan, BappSc, MA, Black Dog Institute,
Hospital Road, Randwick, NSW 2031, Australia.
Copyright © 2007 by Lippincott Williams & Wilkins
ISSN: 0022-3018/07/19501-0020
DOI: 10.1097/01.nmd.0000252008.95227.88

20

of culture and language spoken (Parker et al., 2001a). Crosscultural studies on depression in the Chinese are particularly
important given reports on the low prevalence of depression
among Chinese in different parts of the world (Hwu et al.,
1996; Kleinman, 1986; Lin, 1985; Zhang et al., 1998). Parker
et al. (2001a) reviewed rates of depression in community
studies undertaken in China and Taiwan and proposed a
number of explanations for the lower prevalence rates, including (1) symptom reporting being restricted by stigma and
the Chinese view that emotional illness is just part of “life”;
(2) sociocultural values and family cohesiveness serving as
protective factors against depression in Chinese communities;
and (3) the tendency for Chinese to “somatize,” representing
an idiomatic style for reporting emotional distress. In relation
to the latter explanation, the high prevalence of “somatization” among depressed Chinese patients has been detailed by
many authors (e.g., Cheung, 1995; Kirmayer et al., 1993;
Kleinman, 1986; Lee, 1998).
A final explanation for the low prevalence of depression in Chinese communities relates to the lack of culturally
sensitive case-finding tools, which may make detection difficult (Chen et al., 1998). Although quantitative self-report
measures for assessing depression in the Chinese have been
developed (e.g., Beck Depression Inventory) and demonstrate
satisfactory internal reliability and concurrent validity (Cheung
and Bagley 1998, Yeung et al., 2002), cultural biases still
exist. For example, Zheng et al. (1988) reported that, in the
validation of the Chinese Beck Depression Inventory, three
out of six factors extracted by principal component analysis
did not satisfactorily explain clinical features of depression.
Chen et al. (1998) reported a study using the DSM Scale for
Depression (DSD-26), which is derived from the Diagnostic
Interview Schedule and the DSM-IV (American Psychiatric
Association, 1994) criteria for “caseness.” Results indicated
that Chinese and Anglo-Americans with similar level of
depression responded differently to five items assessing somatic complaints and guilt, suggesting that these items were
likely to have differing cultural meanings.
Parker and colleagues developed the 10-item Depression in Medically Ill (DMI-10) measure as a clinical depression screening tool (Parker et al., 2001b, 2002). During the
development of the DMI-10, 81 provisional items were refined to 16 items by examining their concurrent validity with

The Journal of Nervous and Mental Disease • Volume 195, Number 1, January 2007

The Journal of Nervous and Mental Disease • Volume 195, Number 1, January 2007

the Hospital Anxiety and Depression Scale and the Beck
Depression Inventory for Primary Care (Parker et al., 2001b).
Parker et al. (2002) subsequently developed the brief 10-item
DMI-10. As the DMI was developed for use with medically
ill samples, it uses a cognitive-focused item set to assess depression, and thus avoids any confounding influence of physical
symptoms. Its psychometric properties have been assessed using
different samples, including hospital inpatients and outpatients,
attendants of general practice, and psychiatric outpatients
(Parker and Gladstone, 2004; Parker et al., 2001b, 2002, 2003).
As this measure focuses on cognitive symptoms rather than
physical symptoms to assess depression, it may be well-suited
for use in Chinese samples, since somatic complaints appear to
be highly prevalent among the Chinese. Thus, validating the
Chinese version of the DMI-10 may assist cross-cultural research in depression as well as enabling future application in
clinical settings.
In relation to establishing the cross-cultural validity of
measures, issues regarding psycholinguistic equivalence
(Poortinga, 1989) seem to offer the greatest challenge for
researchers. Westermeyer and Janca (1997) describe the use
of a bilingual test-retest methodology to establish psycholinguistic validity of different language versions of a measurement tool. This strategy determines whether the original
version (e.g., the English version) of the instrument and the
translated version are measuring the same constructs. The
same bilingual subjects complete both language versions on
separate occasion but within close proximity of time, and
resulting correlation coefficients are used to assess the significance of the test-retest data. Typically, test-retest studies
are used to assess the reliability of a measure, but for
cross-cultural studies, this methodology is deemed appropriate for assessing the validity of a translated measure. Another
issue relevant to cross-cultural validation of measures relates
to the use of a top-down approach (involving a priori hypotheses) versus a bottom-up approach (hypotheses not well
defined but guided by the data collected). Many research
projects translating Western psychiatric instruments or depressive measurement tools have followed a top-down approach to establish content validity and/or concurrent validity
(Cheung and Bagley, 1998; Lee et al., 1998). The procedures
usually involve translation of the original instrument by
bilingual translators, followed by blinded back-translation by
another translator and then examining Cronbach ␣ as an
indicator of internal consistency and/or factor analysis to
verify construct validity. However, Bhui et al. (2003) have
emphasized the benefits of using a bottom-up (or qualitative)
approach such as focus groups, pilot studies, and consultation
with community agencies to strengthen the face and content
validity of the adapted instrument. A number of researchers
have used qualitative data to explore the cultural meaning of
terms (e.g., Cheng et al., 2001; Ng-Tse, 2001). Thus, it
appears to be advantageous to integrate both top-down and
bottom-up approaches within cross-cultural validation of
measures (de Jong and Van Ommeren, 2002).
The present study aims to establish the cross-cultural
validity of the Chinese version of the DMI-10, allowing
future studies to pursue nuances of the Chinese experience of
© 2007 Lippincott Williams & Wilkins

Cross-Cultural Validation of DMI-10

depression. We employ both top-down and bottom-up approaches to explore the meanings of emotional expressions in
the Chinese context and establish the cross-cultural validity
of the new Chinese version, in comparison with the original
English version.

METHOD
Study 1
This study was part of a larger quantitative study
examining the impact of acculturation on depression among
Chinese living in Sydney. The larger study was conducted
from May 2003 to April 2004 at 11 general practices in
metropolitan Sydney, with the methodology detailed elsewhere (Parker et al., 2005). There were a total of 385 Chinese
subjects (answering questionnaires in their preferred Chinese
or English language). For the current test-retest reliability
study, a subsample of 28 bilingual subjects completed one
language version of the DMI-10 before consulting their
general practitioner and completed the alternative language
version after their consultation, approximately 15 to 20 minutes later. Bilingual subjects were those who self-reported
“competence in reading both Chinese and English” languages. Subjects selected their preferred language version for
completion first. The DMI-10 required the subject to rate the
agreement of each item in relation to how she or he has been
feeling over the last 2 to 3 days. The four possible options
were “not true” (0), “slightly true” (1), “moderately true” (2),
and “very true” (3). Thus, the total DMI-10 score (10 items)
ranged from a minimum of 0 to a maximum of 30. The order
of items in the DMI-10 for the Chinese version was different
from that of the English versions to reduce learning effects.

Study 2
Study 2 consisted of a series of qualitative focus groups
designed to explore the meaning of depressive experiences
and help seeking among the Chinese. After preliminary
analysis of data from study 1, it was deemed important to
examine the cultural nuances associated with the Chinese
emotional words and phrases in the translated version of the
DMI-10, in addition to the original research objectives. These
focus groups were conducted from April to November 2004
in Mandarin, Cantonese, or English, as language itself is not
a marker of culture. Participants responded to an advertisement posted in their local community centers seeking those
self-identified as Chinese and interested in a discussion on
mental health and depression. Informants from five focus
groups were asked to complete the DMI-10 in their preferred
language before discussion commenced. After a schedule of
semistructured questions for the focus group study had been
delivered, the participants were each given new forms of the
Chinese and English DMI-10. The facilitator (B. C.) then
invited participants to provide feedback on the specific meaning for the key words used to describe emotions embedded in
all items of the Chinese DMI-10, as well as the contexts in
which these words were used. This process usually lasted 10
to 30 minutes, and all discussions were audiotaped and transcribed. Transcripts of Chinese-speaking groups were later

21

The Journal of Nervous and Mental Disease • Volume 195, Number 1, January 2007

Chan et al.

translated into English. Based on these qualitative data, the
researchers prepared a revised version of Chinese DMI-10.

TABLE 1. Bilingual Test-Retest Reliability for Each Item in
the DMI-10 and Total Scores in Study 1 and Study 3

Study 3
Study 3 consisted of a second bilingual test-retest study
with participants from another five focus groups (N ⫽ 26)
using this revised Chinese version. Each participant was
given the DMI-10 in their preferred language version. After
the scheduled questions for the main qualitative study had
been delivered, the participants were given the alternative
language version of the DMI-10, approximately 90 minutes
later. Instead of adopting the scoring scheme in the original
English version, a visual analogue strategy was used (a blank
line 70 mm in length with end points labeled as “not true” on
the left and “very true” on the right) to overcome issues of
scalar equivalence (Van de Vijver and Leung, 1997), and as
used by Eisenbruch (1990). For each item, the length from the
left end-point to the informant’s marking was measured and
converted to a raw score with a minimum of 0 and a
maximum of 1. Thus the total DMI-10 score (10 items) would
range from a minimum of 0 to a maximum of 10. As with
study 1, the inclusion criterion for bilingual subjects was
self-report reading competence in both languages, and the
order of items for the Chinese version was different from that
of the English versions to reduce learning effects. Ethics
approval for all three studies was obtained by the Human
Research Ethics Committee of the University of New South
Wales, Australia. Informed consent was obtained from participants in all three studies.

Study 1
Descriptive
Rating Option
(N ⴝ 28)

Original Item
1. Are you stewing over
things?
2. Do you feel more
vulnerable than usual?a
3. Are you being self-critical
and hard on yourself?
4. Are you feeling guilty
about things in your life?
5. Do you feel as if you have
lost your core and
essence?
6. Are you feeling depressed?
7. Do you feel less
worthwhile?a
8. Do you feel hopeless or
helpless?
9. Do you feel more distant
from other people?a
10. Do you find that nothing
seems to be able to cheer
you up?a
Overall correlation (sum of 10
items)

Study 3
Visual Analogue
Rating Option
(N ⴝ 26)

0.45*

0.47*

0.33

0.55**

0.61**

0.63**

0.54**

0.79***

0.45*

0.62**

0.74***
0.50**

0.59**
0.70***

0.41*

0.77***

0.28

0.49*

0.29

0.67***

0.73***

0.87***

*p ⬍ 0.05; **p ⬍ 0.01; ***p ⬍ 0.001.
a
These four items were revised prior to study 3.

RESULTS
Study 1

The mean age of informants was 31.5 years (SD ⫽ 9.0),
mean age at migration was 17.7 (SD ⫽ 9.24), and mean years
of residence was 13.8 years (SD ⫽ 7.8), with a female
preponderance of 57.1%.
Two strategies were used to examine bilingual testretest validity. First, Pearson correlation coefficients were
used to compare item scores and total score on the Chinese
and English versions for the 28 subjects. Second, mean total
scores for the Chinese and English versions were compared
using paired sample t tests. Table 1 presents the correlation
coefficients between English and Chinese versions for each
item and for total DMI-10 scores. For the total DMI-10 score,
the coefficient was large and significant (r ⫽ 0.73, p ⬍
0.001), while coefficients for individual items ranged from
0.28 to 0.74, and three items failed to reach significance
(Table 1). The mean total score on the Chinese version (M ⫽
9.7, SD ⫽ 7.2) was significantly higher (t ⫽ 2.81, p ⫽ 0.009)
than that for the English version (M ⫽ 7.1, SD ⫽ 6.1).

you stewing over things?” “Do you feel as if you have lost
your core and essence?” and “Are you feeling depressed?”
The other two items, “Are you being self-critical and hard on
yourself?” and “Do you feel hopeless or helpless?” contained
easily identifiable equivalent constructs in the Chinese language. Thus, no changes were made to these five items for the
second iteration of the DMI-10.
The item, “Do you find that nothing seems to be able to
cheer you up?” did not raise concerns for participants. However, in the original translation, the phrase gao-xı́ng (happy)
is very formal and often used in writing and less frequently in
daily conversation (especially in Cantonese; all the phonetic
transcriptions cited in this paper and their corresponding
Chinese characters have been listed in Figure 1). Given the
low initial correlation, the less formal phrase kai-xı́n was used
in the revised version in an effort to improve psycholinguistic
equivalence.

Study 2
Study 2 explored the cultural nuances of the items in
the DMI-10, revising them for a second test-retest study.
There were five items in the initial Chinese translation of the
DMI-10 that did not raise any concerns for the focus group
participants. Among these five items, three used a Chinese
idiomatic expression in the translation for the original: “Are

22

FIGURE 1.Chinese characters and corresponding phonetic
transcriptions (in alphabatical order of transcriptions).
© 2007 Lippincott Williams & Wilkins

The Journal of Nervous and Mental Disease • Volume 195, Number 1, January 2007

The remaining four items were discussed more extensively by the focus group participants. In relation to the item,
“Do you feel more vulnerable than usual?” participants with
different language backgrounds (Mandarin and Cantonese)
agreed that the translation cuı̀ ruò (vulnerable) was an appropriate choice. One Mandarin-speaking female said,
“Cuı̀ ruò (vulnerable) means break, easily breakable,
cuı̀ ruò (vulnerable) is to say when you face a pull back or a
blow, you fall ill, your mental state collapse or fall apart, this
is cuı̀ ruò (vulnerable).”
A Cantonese-speaking female talked about the word cuı̀
ruò (vulnerable) as related to one’s mental state in contrast to
a similar word that describes one’s temperament:
“I think your feeling is cuı̀ ruò (vulnerable) that is no
doubt. Cuı̀ ruò (vulnerable) is an adjective; cuò zhé (frustration)
is a noun, is different. Cuı̀ ruò (vulnerable), I think that is more
related to your mental state, ruăn ruò (weak) is something to
describe a person’s personality or temperament etc.”
However, traditionally the word cuı̀ ruò (vulnerable)
carries a strong feminine tone. To eliminate this gender bias
and confusion with the “physical sense of easily breakable”
(as raised by one informant), a decision was made to replace
the word cuı̀ ruò (vulnerable) with a phrase that spelled out
the meaning explicitly into the two components: g ăn-qı́ng
(literally meaning emotion) and shòu-shang-hài (literally
meaning to get hurt).
The item “Do you feel less worthwhile?” was viewed as
too severe, and the alternative phrase bù-zhong yòng (“no ⫹
use”) was described as being more familiar by focus group
participants. In the first translation, the Chinese phrase “no”
was combined with the phrase “worth” to convey the meaning “less worthwhile.” To illustrate, a young Cantonesespeaking female compared the phrase méi-you jià-zh (literally
means no ⫹ worth) and bù zhong yòng (no ⫹ use; mentioned
as a better alternative) in terms of the tone or intensity these
two words carry:
“Méi-you jià-zhı́ (literally means no ⫹ worth) is too
strong, bù zhong yòng (literally means no ⫹ use) is like just
venting out, it is lighter in tone.”
The phrase bù zhong yòng in Mandarin or mo mat yòng
in Cantonese is often observed in the spontaneous repertoire
of Mandarin or Cantonese speakers. Such familiarity transfers to a lower severity of psychological impact when responding to the word no ⫹ use.
In relation to the item, “Do you feel more distant from
other people?” the original phrase (shu-yuăn) contains the
concept of “distant,” but is usually used as a verb to denote
the act of distancing a relationship due to some valid reasons.
A quote from a female Mandarin-speaker demonstrated how
the real experience of not wanting to socialize fitted well in
the Chinese interpretations of the kind of behavior one would
expect when one’s mood was not good:
“I prefer not willing to socialize. Socialize means when
I am very happy, then I will tend to like to meet many people
from different places, and go out and have fun, i.e., I like to
socialize! When I am not happy, my mood is down then I will
turn down all the invitation, that means I don’t want to
socialize. shu-yuăn (keeping a distance) means regardless of
© 2007 Lippincott Williams & Wilkins

Cross-Cultural Validation of DMI-10

whether I am happy or not, if I don’t like that person, I will
try to keep a distance from him, not wanting to befriend him.”
Though the phrase not willing to socialize might seem
very different from the literal meaning of the original English
word feeling distant, to the Chinese, it matched the intended
meaning of the construct in original DMI-10 more closely.
In relation to the item, “Are you feeling guilty about
things in your life?” bilingual Chinese responded differently
to the word guilty in Chinese and English. For a bilingual
participant, the word guilty in English stimulated feelings of
emotional distance from people, but the same word in Chinese, nèi-jiù, triggered the guilty feelings of not fulfilling the
family obligation:
“. . . the word guilty, you use it quite frequently to
describe, right, how you feel about your work and stuff like
that . . . But in Chinese, when you read this nèi-jiù, you
actually use it to describe more serious offenses like you
know my child . . . (feel guilty towards your child) . . . the
Chinese one is more intense but then the English one . . .”
This participant reported that the Chinese phrase guilty
was able to trigger more intense emotion as compared with
the English word. Thus, the translation of this item was not
modified, but caution was noted about the inherent difference
in the construct of Chinese and English emotional words.

Study 3
Following the modification of the four items described
in study 2, the second test-retest study was conducted. The
mean age of informants (N ⫽ 26) in the second test-retest
reliability measure was 33.7 years (SD ⫽ 10.0), mean age at
migration was 22.0 years (SD ⫽ 9.3), mean years of residence was 11.0 years (SD ⫽ 7.0), and there was a female
preponderance of 61.5%. Comparing the informants in study
3 with those in study 1, there were no significant differences
between the mean age, mean age at migration, and mean
length of residence between the groups.
Paired t tests were used to determine cross-cultural
validity of items and mean total scores for the Chinese and
English versions. Table 1 also presents the results for study 3
(in comparison with study 1 results). The test-retest coefficients for the total DMI-10 were r ⫽ 0.87 (p ⫽ 0.001), with
all individual item correlations reaching significance and
ranging from 0.47 to 0.77. Study 3 findings showed an
improvement when compared with results of study 1. Three
items that failed to achieve significance in study 1 reached
significance in study 3. The mean total scores of the Chinese
and English DMI-10 were 2.4 (SD ⫽ 1.8) and 2.0 (SD ⫽ 1.8),
respectively, but were still significantly different (t ⫽ 2.44;
p ⬍ 0.01), with Chinese language scores again exceeding
English language scores.

DISCUSSION
The aim of the present study was to establish the
cross-cultural validity of the Chinese version of the DMI-10,
a self-report measure of cognitive-based depression symptoms. This study comprised both top-down and bottom-up
methods, with a series of focus groups following the initial
bilingual test-retest study to explore the psycholinguistic
equivalence of constructs and a second test-retest study of the

23

Chan et al.

The Journal of Nervous and Mental Disease • Volume 195, Number 1, January 2007

iterated version. The second test-retest study (study 3) demonstrated an excellent test-retest coefficient, and therefore
established the cross-cultural validity of the DMI-10 Chinese
version.
While the first test-retest study achieved a satisfactory
overall level of cross-cultural validity (0.73), several items
were problematic. Our focus group strategy explored the
meaning of the items, with revision of four to respect cultural
nuances. It was of interest to find that three items in the first
translation that used a Chinese idiomatic expression matched
closely with the emotional construct in the original version.
Idioms conveys well-defined meanings based on experiences
that most Chinese can relate to, and other researchers have
emphasized the importance of including idioms in self-report
measures (e.g., Phan et al., 2004). As a consequence of focus
group discussions, revisions were made to four items where
more culturally appropriate words and phases were identified.
The second strategy to improve the test-retest coefficients was the use of a visual analogue scale to overcome
cultural differences in measuring the abstract concepts in
different cultures. Of importance, significant improvements
in test-retest coefficients were observed from study 1 to study
3 in three of the six items with no revision. Improvement in
the correlation coefficients between the two test-retest studies
may also reflect the use of the visual analogue strategy, and
it is not possible to determine to what extent improvement in
the four revised items was due only to changes in translation.
Regardless of the relative contribution of these two factors to
the improved test-retest coefficient, the revised Chinese
DMI-10 shows a strong correlation with the English version,
thus arguing for its validity.
The findings of this study have implications for crosscultural research into depression in Chinese communities.
There is a pressing need to find a valid depression screening
tool for Chinese living in Western regions as international
migration has increased dramatically in recent decades. With
the robust methodology used in the present study to validate
the DMI-10 Chinese version, the measure has the potential to
be used as an efficient screening tool for depression among
Chinese. However, it would be important for future validation
studies to examine its use in other Chinese communities and
with clinical samples of depressed patients. The fact that the
DMI-10 focuses on cognitive symptoms rather than on physical symptoms allows interpretation of the results without the
complication of issues surrounding somatization.
There are important study implications for the methodology of future cross-cultural studies on clinical depression
and other psychopathology. Exploring meanings through
qualitative research is a crucial step to test for the existence of
overlapping concepts in emotional constructs within different
cultures. There are obvious advantages in integrating topdown and bottom-up approaches. For example, the first set of
quantitative test-retest reliability results was able to identify
any mismatch in the English and Chinese expressions. During
the first translation, the phrase shu-yuăn, which carries the
literal meaning of keeping a distance, was used, but this was
revised following focus group discussion. Overall, the narratives gathered in the qualitative study provide an opportunity to

24

understand the emotions associated with the Chinese phrases
and improve the translation of the items. The qualitative component in the study enabled exploration of meanings in cultural
contexts which Good (1992) has described as the importance of
addressing the “lived experience” to reconstruct the concept of
the phenomenology of clinical depression.
Bilingual speakers express emotions differently depending on the language used and the context of its use. This
was demonstrated by the narratives of a bilingual informant,
who explained how guilty in English elicited feelings surrounding her work environment and were more related to
breaking of rules, while the same word in Chinese tapped into
feelings connected with family obligations. Koven (2004)
described a single case study of a French-Portuguese bilingual speaker whom expressed emotions differently in narratives of personal experience in her two languages. The author
argues that it is not the mere difference of the structures of the
two languages but the repertoire of the role to which this
speaker has access in each language that is relevant. As our
bilingual informants obtained a significantly higher score on
the Chinese than English versions— even in the second iteration of the measure—this may be due to the cultural difference in language usage, which is an inherent difficulty in
cross-cultural research. Chan (1991) administered the original and Chinese version of the Beck Depression Inventory to
a large sample of bilingual subjects. While he reported
overall similar mean scores for both Chinese and English
version, he observed some differences in item scores that
were judged due to intrinsic language differences. Aside from
the inherent language differences, the difference in scores
could also be due to the small sample size (in both study 1
and study 3), and therefore, replication of this study using a
larger sample size is required.
There were several limitations to our studies. The first
pertains to the nonuniform source of research subjects across
the three studies. The bilingual subjects in study 1 were
recruited from a medical clinic located in a relatively affluent
area, whereas the focus group participants in studies 2 and 3
were recruited from community centers with a wide range of
socioeconomic backgrounds. As we included no formal assessment of the subjects’ relative reading comprehension in
either English or Chinese, they may not have had comparable
bilingual skills.
The second limitation relates to the fact that only one
version of Chinese language was used on a diverse group of
informants speaking many different dialects. In this current
study, the native dialects of the focus group participants were
mainly Cantonese, Mandarin, and Shanghainese. The printed
form of the Chinese DMI-10 in all studies was Traditional
Chinese. While it is not always practical to print forms in
every dialect, managing linguistic diversity remains one of
the central challenges faced by cross-cultural researchers.
Future studies may explore cultural issues in more detail.
Finally, as the current study only focused on bilingual
test-retest coefficients for an indication of the goodness of fit
between the Chinese translated version and the original English DMI-10 using nonclinical samples, further cross-cultural validation studies of this measure are required. Valida© 2007 Lippincott Williams & Wilkins

The Journal of Nervous and Mental Disease • Volume 195, Number 1, January 2007

tion studies using the DMI-10 Chinese version with a clinical
sample should compare results with clinician judgment, establish concurrent validity with other standardized depression
measures for the Chinese population, and use different Chinese samples to support fully the psychometric properties of
the DMI-10 Chinese version.

CONCLUSION
This study established the validity of the DMI-10 Chinese version with excellent test-retest reliability found on the
second iteration of the questionnaire. This is the first step
taken to establish psycholinguistic equivalence of the Chinese
version and should be followed by further validation studies
using clinical samples in different regions. The Chinese
DMI-10 has the potential to be used as an efficient screening
tool for depression both for Chinese in Australia and in Chinese
regions. Cross-cultural researchers need to acknowledge the
limitations in translating abstract cognitive concepts into different languages and cultures. A combined quantitative and qualitative approach, as used in the present study, is a robust way to
explore meanings and contexts as well as identify subtle cultural
nuances in emotional words across cultures.

ACKNOWLEDGMENTS
We thank the NHMRC (Program Grant 222708) and
Centre for Mental Health for funding support, A/Professor
Jan Ritchie for assisting with the interpretation of qualitative
data, and the Medical Centre and Community Centres facilitating this study.
REFERENCES
American Psychiatric Association (1994) Diagnostic and Statistical Manual
of Mental Disorders (4th ed). Washington (DC): American Psychiatric
Association.
Bhui K, Mohamud S, Warfa N, Craig TJ, Stansfeld SA (2003) Cultural
adaptation of mental health measures: Improving the quality of clinical
practice and research (editorial). Br J Psychiatry. 183:184 –186.
Chan DW (1991) The Beck Depression Inventory: What difference does the
Chinese version make? Psychol Assess. 3:616 – 622.
Chen IG, Roberts RE, Aday LA (1998) Ethnicity and adolescent depression:
The case of Chinese Americans. J Nerv Ment Dis. 186:623– 630.
Cheng ATA, Tien AY, Chang CJ, Brugha TS, Cooper JE, Lee CS, Compton
W, Liu CY, Yu WY, Chen HM (2001) Cross-cultural implementation of
a Chinese version of the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) in Taiwan. Br J Psychiatry. 178:567–572.
Cheung C, Bagley C (1998) Validating an American scale in Hong Kong:
The Center for Epidemiological Studies Depression Scale (CES-D). J Psychol. 132:169 –187.
Cheung FM (1995) Facts and myths about somatization among the Chinese.
In TY Lin, WS Tseng, EK Yeh (Eds), Chinese Societies and Mental
Health (pp 156 –166). Oxford: Oxford University Press.
de Jong JTVM, Van Ommeren M (2002) Toward a culture-informed epidemiology: Combining qualitative and quantitative research in transcultural
contexts. Transcult Psychiatry. 39:422– 433.
Eisenbruch M (1990) Classification of natural and supernatural causes of
mental distress: Development of a Mental Distress Explanatory Model
Questionnaire. J Nerv Ment Dis. 178:712–719.
Good B (1992) Culture and psychopathology: Directions for psychiatric
anthropology. In T Schwartz, MW White, CA Lutz (Eds), New Directions

© 2007 Lippincott Williams & Wilkins

Cross-Cultural Validation of DMI-10

in Psychological Anthropology (pp 181–205). Cambridge: Cambridge
University Press.
Hwu HG, Chang IH, Yeh EK, Chang CJ, Yeh LL (1996) Major depressive
disorder in Taiwan defined by the Chinese Diagnostic Interview Schedule.
J Nerv Ment Dis. 184:497–502.
Kirmayer L, Minas H (2004) The future of cultural psychiatry: An international oerspective. Can J Psychiatry. 45:438 – 446.
Kirmayer LJ, Robbins JM, Dworkind M, Yaffe MJ (1993) Somatization and
the recognition of depression and anxiety in primary care. Am J Psychiatry. 150:734 –741.
Kleinman A (1986) Social Origins of Disease and Distress: Depression,
Neurasthenia and Pain in Modern China. New Haven (CT): Yale University Press.
Koven M (2004) Getting “emotional” in two languages: Bilinguals’ verbal
performance of affect in narratives of personal experience. Text. 24:471–
515.
Lee DTS, Yip SK, Chiu HFK, Leung TYS, Chan KPM, Chau IOL, Leung
HCM, Chung TKH (1998) Detecting postnatal depression in Chinese
women: Validation of the Chinese version of the Edinburgh Postnatal
Depression Scale. Br J Psychiatry. 172:433– 437.
Lee S (1998) Estranged bodies, simulated harmony and misplaced cultures:
Neurasthenia in contemporary Chinese society. Psychosom Med. 60:448 –
457.
Levy RI (1984) Emotion, knowing and culture. In RA Shweder, RAL Vine
(Eds), Culture Theory: Essays on Mind, Self and Emotion (pp 214 –237).
Cambridge: Cambridge University Press.
Lin TY (1985) Mental disorders and psychiatry in Chinese culture: characteristic features and major issues. In WS Tseng, YH Wu (Eds), Chinese
Culture and Mental Health (pp 369 –394). Orlando (FL): Academic Press.
Lutz C, White GM (1986) The anthropology of emotions. Annu Rev Anthropol. 15:405– 436.
Ng-Tse TM (2001) Development of a questionnaire to assess the impact of
mental illness on Chinese women. Hong Kong J Psychiatry. 11:9 –16.
Parker G, Gladstone G, Chee KT (2001a) Depression in the planet’s largest
ethnic group: The Chinese. Am J Psychiatry. 158:857– 864.
Parker G, Hilton T, Bains J, Hadzi-Palovic D (2001b) Screening for depression in the medically ill: The suggested utility of a cognitive-based
approach. Aust N Z J Psychiatry. 35:474 – 480.
Parker G, Hilton T, Bains J, Hadzi-Palovic D (2002) Cognitive-based
measures screening for depression in the medically ill: The DMI-10 and
the DMI-18. Acta Psychiatr Scand. 105:419 – 426.
Parker G, Hilton T, Hadzi-Palovic D, Irvine P (2003) Clinical and personality correlates of a new measure of depression: A general practice study.
Aust N Z J Psychiatry. 37:104 –109.
Parker G, Gladstone G (2004) Capacity of the 10-item Depression in the
Medically Ill screening measure to detect depression “caseness” in psychiatric out-patients. Psychiatry Res. 127:283–287.
Parker G, Chan B, Tully L, Eisenbruch M (2005) Depression in the Chinese:
The impact of acculturation. Psychol Med. 35:1475–1483.
Phan T, Steel Z, Silove D (2004) An ethnographically derived measure of
anxiety, depression and somatization: The Phan Vietnamese Psychiatric
Scale. Transcult Psychiatry. 41:200 –232.
Poortinga YH (1989) Equivalence of cross-cultural data: An overview of
basic issues. Int J Psychol. 24:737–756.
Rosaldo R (1983) Grief and the headhunters’ rage. In E Bruner (Ed), Play,
Text and Story (pp 178 –195). Washington (DC): American Ethnological
Society.
Van de Vijver F, Leung K (1997) Methods and Data Analysis for CrossCultural Research. Thousand Oaks: Sage.
Westermeyer J, Janca A (1997) Language, culture and psychopathology:
Conceptual and methodological issues. Transcult Psychiatry. 34:291–311.
Yeung A, Howarth S, Chan R, Sonawalla S, Nierenberg AA, Fava M, (2002)
Use of the Chinese version of the Beck Depression Inventory for screening
depression in primary care. J Nerv Ment Dis. 190:94 –99.
Zhang WX, Shen YC, Li SR (1998) Epidemiological investigation on mental
disorders in 7 areas of China (in Chinese). Chin J Psychiatry. 31:69 –71.
Zheng Y, Liang W, Goa L, Zhang G, Wong C (1988) Applicability of the
Chinese Beck Depression Inventory. Compr Psychiatry. 29:484 – 489.

25