the factor structure of the ces-d in a sample of rwandan genocide survivors
TRANSCRIPT
ORIGINAL PAPER
The factor structure of the CES-D in a sample of Rwandangenocide survivors
Justin J. Lacasse • Marie J. C. Forgeard •
Nuwan Jayawickreme • Eranda Jayawickreme
Received: 16 September 2012 / Accepted: 17 September 2013 / Published online: 31 October 2013
� Springer-Verlag Berlin Heidelberg 2013
Abstract
Background Past research suggests that culture shapes
the way psychopathology is experienced and expressed.
Standard psychiatric assessment instruments may therefore
not capture the same underlying constructs in different
contexts. The present study investigated the factor structure
of a standard depression scale in a sample of Rwandan
genocide survivors.
Methods One hundred ninety six Rwandan adults pro-
vided socio-demographic information and completed the
Center for Epidemiological Studies-Depression scale
(CES-D), one of the most widely used self-report instru-
ments assessing depressive symptoms, as part of a larger
study on well-being and mental health in Rwanda.
Results A two-factor solution provided the best fit for
these CES-D data. The first factor corresponded to general
depressive symptoms (including depressed affect, somatic
symptoms, and interpersonal concerns) and explained
37.20 % of the variance. The second factor included items
assessing positive affect and explained 8.68 % of the
variance.
Conclusions The two-factor solution found in the present
study deviates from the commonly reported four-factor
structure, but is consistent with studies showing that
depressed affect and somatic symptoms may not be expe-
rienced as distinct in certain non-Western and minority
cultural groups.
Keywords Depression � Psychopathology � Culture �Idioms of distress � Factor analysis
The factor structure of the CES-D in a sample
of Rwandan adults
Past research suggests that culture, defined as a ‘‘system of
meaning’’ shared by a particular group [1], affects the way
psychiatric disorders are experienced and expressed [2–4].
Culture plays an important role in determining how indi-
viduals interpret internal and external events, leading
Kleinman [5] to propose that ‘‘illness’’ should be under-
stood as ‘‘the personal, interpersonal, and cultural reactions
to disease’’. In keeping with this, many in the field have
stressed the importance of looking at the specific mecha-
nisms explaining how culture impacts the experience and
expression of psychopathology [2, 6–9].
In spite of this growing body of evidence, there appears
to be a widespread belief among public health and medical
researchers that assessment measures developed for use in
J. J. Lacasse and M. J. C. Forgeard contributed equally to this paper.
J. J. Lacasse
School of Osteopathic Medicine, University of Medicine
and Dentistry of New Jersey, Stratford, USA
Present Address:
J. J. Lacasse
Department of Psychiatry of Tufts Medical Center,
Tufts University School of Medicine, Boston, USA
M. J. C. Forgeard
Department of Psychology, University of Pennsylvania,
Philadelphia, USA
N. Jayawickreme
Department of Psychology, Manhattan College,
New York City, USA
E. Jayawickreme (&)
Department of Psychology, Wake Forest University,
P.O. Box 7778, Winston-Salem, NC 27109, USA
e-mail: [email protected]
123
Soc Psychiatry Psychiatr Epidemiol (2014) 49:459–465
DOI 10.1007/s00127-013-0766-z
Western samples are capable of meaningfully capturing
psychopathology as it is expressed and experienced in a
diverse range of populations around the world [10]. In a
comprehensive review of research on the health status of
refugee populations exposed to trauma, Hollifield et al.
[11] noted that a majority of studies utilized assessment
measures whose validity or reliability had not been inves-
tigated in the populations at hand.
One way to begin to address these issues is to investigate
the factor structure of standard measures of psychopa-
thology in various cultures. Factor structure can tell us not
only if an instrument reliably measures what it is supposed
to measure, but also if it measures the same underlying
constructs in distinct populations. Thus, by examining how
particular items group in different contexts, researchers can
develop a culturally informed understanding of the distinct
symptom patterns that emerge in particular groups. The
present study reports such an investigation, examining the
factor structure of the Center for Epidemiological Studies-
Depression Scale (CES-D) [12] in a sample of Rwandan
genocide survivors and comparing this factor structure with
the ones found in previous studies conducted with other
populations.
The Center for Epidemiological Studies-Depression
scale (CES-D)
The CES-D is one of the most widely used self-report
instruments assessing depressive symptoms. This 20-item
scale was first validated in three primarily White Ameri-
can samples [12]. In the original validation study, a
principal components analysis (PCA) revealed four factors
together explaining 48 % of the variance in the data.
These included (1) depressed affect (e.g., sadness, crying);
(2) positive affect (e.g., hope, enjoyment); (3) somatic and
retarded activity (e.g., appetite problems, problems ‘‘get-
ting going’’); and (4) interpersonal concerns (e.g., per-
ceiving others as unfriendly). This initial analysis
presented a number of methodological problems. First, the
use of PCA (instead of factor analysis) is not recom-
mended for a scale with as few as 20 items as it can lead
to inflated factor loadings [13]. Second, this analysis was
conducted using orthogonal varimax rotation, which
assumes uncorrelated factors. This assumption is prob-
lematic given that depressive symptoms likely covary, at
least to some degree (for example, individuals who pres-
ent depressed affect often also report little positive affect).
Finally, the interpersonal problem factors only included
two items, and a minimum of three items is generally
recommended to ensure reliability across samples [14]. In
light of these limitations, it is unclear whether the four-
factor solution provided the best fitting solution for these
original data.
A meta-analysis of 28 studies using exploratory factor
analysis (EFA) or PCA provided support for the four-factor
solution [15] although it remains unclear whether the ori-
ginal studies suffered from the same limitations noted
above. This solution was compared to a two-factor (general
depression and positive affect) and a three-factor (general
depression, positive affect, and somatic symptoms) solu-
tion. Despite support for the four-factor structure,
researchers have noted that the CES-D is in practice treated
as a unidimensional scale, as it yields only one score [16].
A study using confirmatory factor analysis again provided
support for the four-factor solution, and found that an
acceptable one-factor solution could only be produced by
eliminating the four items assessing positive affect from
the scale [16]. Thus, some authors have recommended the
use of a shortened scale to produce a meaningful and valid
total score [16, 17].
Factor structure of the CES-D in ethnically diverse
samples
Several studies conducted with ethnically diverse samples
have replicated the scale’s original four-factor solution.
These studies were conducted among Hispanic individuals
in the United States [18–21], African Americans [21],
Korean immigrants in Canada [22], and Colombian adults
[23]. However, other studies have found different versions
of the four-factor structure. For example, in a sample of
older African Americans, the four factors corresponded to
depressive/somatic symptoms (combined), positive affect,
interpersonal problems, and an ambiguous fourth factor
termed ‘‘social well-being’’ [24]. The same first three
factors emerged in a sample of Asian-Americans (includ-
ing Chinese, Korean, Japanese, and Filipino Americans), in
addition to a fourth factor termed pessimism [25]. Finally,
in a sample of Brazilian college students, the fourth factor
corresponded to a hybrid and somewhat ambiguous
depressive/somatic symptoms factor, while the other fac-
tors corresponded to positive affect, depressive, and
somatic symptoms [26].
Other similar studies have evidenced a three-factor
structure. Studies conducted with samples of Mexican-
Americans, Cuban Americans and Puerto Ricans [27, 28],
Chinese immigrants [29, 30], depressed Spanish-speaking
internet users from around the world [31], female Arab
medical students from the United Arab Emirates [32], and
Greek psychiatric inpatients [33] have produced a three-
factor solution by combining depressive and somatic
symptoms into one single factor. In samples of American-
Indian [34, 35] and Chinese [36] participants, the emerging
three factors resulted from the omission of the interper-
sonal factor (resulting in a depressed affect factor, a
somatic/retardation factor, and positive affect factor).
460 Soc Psychiatry Psychiatr Epidemiol (2014) 49:459–465
123
Finally, some studies have found evidence for a two-
factor solution. In studies conducted with Filipino Ameri-
can adolescents [37] and ethnic Armenians living in Leba-
non [38], the two factors included a general depression
factor and a positive affect factor. In a sample of Hong
Kong adolescents, the two factors corresponded to depres-
sive/somatic symptoms and interpersonal problems [39].
The present study
The present study investigated the factor structure of the
CES-D in a sample of Rwandan adults having experienced
the 1994 genocide. The Rwandan genocide is widely
considered to constitute one of the worst atrocities of the
20th century, with more than 800,000 mostly Tutsi civil-
ians brutally slaughtered [40]. Examining the meaning of
depression in this sample is important given the high levels
of depressive symptomatology and associated functional
impairment in this population [41]. Given results of prior
studies, we expected that a two-, three-, or four-factor
structure would emerge.
Method
Participants
Participants were 200 adults (97 males, 103 females) from
five districts in three provinces—the districts of Bugesera
in the East Province (n = 51), Kamonyi in the South
Province (n = 48), and Gasabo (n = 56), Kicukiro
(n = 19), and Nyarugenge (n = 26) in the Kigali Province
of Rwanda. Participants were 29.40 years old on average
(SD = 10.17 ranging from 18 to 75). Although recom-
mendations regarding sample size for factor analytic
studies vary [14, 42], we recruited 200 participants based
on past research indicating that 10 participants per item is
an acceptable ratio [43–46].
Procedures and materials
Participants were invited to take part in a study about well-
being and mental health by a team of 12 Rwandan research
assistants who visited a total of five genocide survivor vil-
lages (one from each district). Participation was voluntary,
and all participants provided consent before taking part in
the study. The total time to complete the battery of measures
for this study was around 75 min. While individuals were
not compensated individually for their participation (fol-
lowing the advice of the Rwandan Ethics Commission) a
token of appreciation was provided to each village at the
completion of data collection. Participants first received
general instructions in how to respond to the questionnaire
items, and were provided with examples of how to select
appropriate responses. All administered measures were in
Kinyarwanda. These measures had been translated and
back-translated by two bilingual Rwanda translators with
prior experience in translating mental health measures.
The CES-D [12], the 20-item self-report measure
described above, was administered as part of this assess-
ment. Only socio-demographic data (as reported above)
and CES-D results are included in the present study. CES-
D instructions ask respondents to indicate the degree to
which they have experienced various symptoms of
depression during the past week on a 4-point scale (ranging
from 0 to 3). The response scale includes the following
fixed category choices: ‘‘0 = rarely or none of the time
(\1 day),’’ ‘‘1 = Some or a little of the time (1–2 days),’’
‘‘2 = Occasionally or a moderate amount of time
(3–4 days),’’ and ‘‘3 = Most or all of the time (5–7 days)’’.
Following standard practice, four positively worded items
measuring positive affect (‘‘I felt that I was just as good as
other people,’’ ‘‘I felt hopeful about the future,’’ ‘‘I was
happy,’’ and ‘‘I enjoyed life’’) were reverse scored.
Responses on the CES-D range between 0 and 60 points.
Results
Data were analyzed using SAS 9.3. Four participants who
had more than 50 % of datapoints missing were excluded
from the analysis (i.e., 4 participants who did not provide
any answers, and 1 participant who only answered 8 out of
20 CES-D items). The final sample included 196 partici-
pants (89 % of participants had no missing data, 6 % had
one data point missing, 4 % had two data points missing,
0.50 % had three data points missing, and 0.50 % had four
data points missing). Missing data points (0.89 % of all
data) were imputed using the Markov-Chain Montecarlo
method to create five datasets with no missing data. As all
of the missing data imputed was non-monotone, no other
imputation method was needed. The resulting five datasets
were merged by calculating the mean of the five imputed
values for all missing data points.
Participants’ mean score on the CES-D was 22.02 points
(SD = 10.65, min = 0, max = 47). 69.90 % of partici-
pants in this sample had a total score exceeding 16, a cutoff
used in past research to identify individuals experiencing
marked levels of psychological distress [47]. Given the
current paucity of research on the use of the CES-D as a
screening tool in non-Western populations, it is not clear
whether this cutoff is meaningful in the present sample,
and further research should therefore seek to establish
culture-specific cutoffs. These descriptive statistics never-
theless suggest that the present sample displayed high
levels of depressive symptoms on average.
Soc Psychiatry Psychiatr Epidemiol (2014) 49:459–465 461
123
Initial tests on these data indicated that they were suit-
able for EFA. The Kaiser–Meyer–Olkin measure of sam-
pling adequacy was 0.88, above the recommended value of
0.6 [48, 49], and Bartlett’s test of sphericity was signifi-
cant, v2(190) = 1426.40, p \ 0.001.
Given the polytomous nature of the data, a smoothed
polychoric correlation matrix was created in MicroFact 2.0
[50] for use in EFA (using Principal Factor Analysis). The
number of factors to include in the model was estimated
using Velicer’s minimum average partialling (MAP) Test,
which recommended a two-factor solution. Given mixed
results in past research, we tested a one-, two-, three- and
four-factor solutions using oblique promax rotation, as
factors extracted from the CES-D should theoretically be
correlated. For all factor solutions, promax rotation was
conducted using k = 2, 3, and 4 [51]. For each factor
solution, the k value which maximized the hyperplane
count was retained.
The quality of each factor solution was judged by
examining: (1) the amount of variance explained by each
factor, (2) the proportion of items loading on a factor, (3)
hyperplane counts (in percentages), (4) the reliability of
extracted factors, and finally (5) the meaningfulness of
extracted factors. According to these criteria, we retained
the two-factor solution.
In this solution, the first factor explained 37.20 % of the
variance and the second factor explained 8.68 % of the
variance. For the promax rotation, k was set at 2 (hyper-
plane count = 25 %). MicroFact 2.0 also reported two
goodness-of-fit tests (the Goodness of Fit Index and the
Root Mean Squared Residual). Both of these statistics
indicated that the two-factor solution was a good fit for the
data (GFI = 0.98, RMSR = 0.06).
The correlation between the two factors was r = 0.34.
Fifteen items loaded on the first factor (Cronbach’s
a = 0.90). This factor appeared to assess a wide array of
depressive symptoms (including cognitive, somatic, and
interpersonal symptoms). The remaining five items loaded
on the second factor (Cronbach’s a = 0.57), which mainly
included items assessing positive affect (reverse-coded
prior to data analyses). One item loading on this factor was,
however, negatively correlated with others and decreased
its reliability (‘‘I felt everything I did was an effort’’). The
second factor reached an adequate level of reliability by
excluding this item (Cronbach’s a = 0.76).1
The one-, three-, and four-factor solutions were not
retained for the following reasons. The one-factor solution
explained less of the variance (the unique factor explained
36.95 % of the variance) and only included 18 items out of
20 (instead of 19 for the two-factor solution). This was
nevertheless the most plausible solution following the two-
factor solution. The three- and four-factor solutions yielded
a factor with only two items (Items 5 and 7) and insuffi-
cient reliability (Cronbach’s a = 0.34).
Discussion
The present study found that a two-factor solution best
fitted data obtained from a sample of 196 Rwandan geno-
cide survivors on the CES-D. The amount of variance
explained by these two factors (46 %) was similar to the
amount of variance explained by the four factors found in
Radloff’s initial validation study of the CES-D (48 %)
[12]. The two factors in this study corresponded to a gen-
eral depression factor (including depressed affect, somatic/
retardation symptoms, and interpersonal problems) as well
as a positive affect factor (Table 1). The findings of our
study are limited by the somewhat modest sample size used
in this study. These results nevertheless replicate the
findings of previous studies conducted in samples of Fili-
pino American adolescents [37] and ethnic Armenians
living in Lebanon [38]. They confirm that affective and
somatic symptoms are not distinct in this population, as
seen in other cultural contexts [52].
From our results, it remains unclear whether the positive
affect factor truly corresponds to a different construct or
merely constitutes a method factor given that these items
were positively worded, a concern echoed by other
researchers [15]. These items have precluded researchers
from finding a reliable one-factor solution in the CES-D,
which is problematic in light of the fact that this instrument
produces a single score and is therefore in practice treated
as a unidimensional scale [16]. If these four items indeed
form a method factor, then a one-factor solution might best
represent the structure of depression in the Rwandan
context.
Results found regarding one specific item, ‘‘I felt that
everything was an effort,’’ also provide an interesting
example of the need to carefully investigate the specific
meanings of items in different cultural contexts. In this
study, it appeared that this item might have been inter-
preted as positive by participants since it loaded on the
positive affect factor (although not highly enough to be
included in our final solution). This is an important prob-
lem to be considered when using instruments with new
populations.
1 We also reverse-coded this item to examine the resulting factor
reliability, which was improved (Cronbach’s a = 0.71). However,
this item still displayed a lower corrected item-total correlation
(r = 0.19) than all four other items (all rs = 0.51-0.63). Further-
more, it remained unclear whether the content of this item was
meaningfully related to the content of the four positive affect items.
Given these limitations, we decided to exclude this item from the final
solution.
462 Soc Psychiatry Psychiatr Epidemiol (2014) 49:459–465
123
This finding points to the crucial need to investigate
local expressions of distress, as instruments developed in
one context may not capture all of the symptoms consid-
ered important in another context [53, 54]. Future research
must use ethnographic methods to investigate the rela-
tionship between social/cultural factors and psychological
processes, find out what specific words and phrases people
use to describe these experiences, and finally develop
questionnaires that use these idioms and are validated using
standard methods [2, 11, 55, 56].
In the context of Rwanda, Bolton [57] has for example
delineated specific local syndromes and symptoms that do
not completely overlap with the DSM-IV-TR [58] diag-
nostic criteria for depression. These local syndromes
resembling depression include guhahamuka (best
translated as ‘‘mental trauma’’), which includes common
idioms of depression, as well as culturally specific idioms
such as ‘‘feeling like you have a cloud inside’’. Agahinda
(best translated as ‘‘deep sadness or grief’’) provides
another example of a local syndrome again including
common idioms of depression as well as culture-specific
idioms such as ‘‘burying one’s cheek in one’s palm’’. Thus,
combining standard instruments with culturally informed
ones may be important when conducting psychiatric epi-
demiological and intervention work in ethnically diverse
populations [59].
Conclusion
Given that culture provides a system of meaning that
influences the experience and expression of psychological
distress, researchers run the risk of measuring constructs
that are ‘‘experientially meaningless’’ [60] if they do not
investigate the reliability and validity of the instruments
they use in the particular populations they are studying.
Furthermore, future research should supplement the use of
validated standard instruments with locally meaningful
instruments incorporating local idioms of distress, as this
would help improve the assessment of psychiatric disorders
in populations at risk (such as Rwandan genocide survi-
vors), and in turn the evaluation of the effectiveness of
intervention efforts in such populations [10].
Acknowledgments We are grateful to the Positive Psychology
Center and the Department of Psychology at the University of
Pennsylvania for providing the funding for this project. Additional
funding was generously provided by Eva Kedar, Ph. D. We thank
Virgile Uzabumugabo and his research team (Beza Gisele, Igena
Clarisse, Kankindi Antoinette, Mugisha Norbert, Mushimiyimana
Delphine, Mwiseneza Sophie, Nsengiyumva Joselyne, Runyurangabo
Philbert, Twajamahoro Contstantin, Ufitemariya Janviere, Usabimana
Hawa, and Uwanyiligira Honorine) for collecting the data for this
project. We also thank Paul Di Stefano and Richard Bisa for assisting
with the translation process and Eli Tsukayama for advice on data
analysis.
Conflict of interest The authors declare that they have no conflict
of interest.
References
1. Rohner RP (1984) Toward a conception of culture for cross-
cultural psychology. J Cross Cult Psychol 15:111–138. doi:10.
1177/0022002184015002002
2. Lopez SR, Guarnaccia PJ (2000) Cultural psychopathology:
uncovering the social world of mental illness. Annu Rev Psychol
51:571–598. doi:10.1146/annurev.psych.51.1.571
3. Deisenhammer EA, Coban-Basaran M, Mantar A et al (2012)
Ethnic and migrational impact on the clinical manifestation of
depression. Soc Psychiatry Psychiatr Epidemiol 47:1121–1129.
doi:10.1007/s00127-011-0417-1
Table 1 Factor loadings for the two-factor solution using promax
rotation (k = 2)
Items Construct Depressive
symptoms
Positive
affect
1. I was bothered by things that
usually do not bother me
SR 0.68 –
2. I did not feel like eating; my
appetite was poor
SR 0.59 –
3. I felt that I could not shake off
the blues even with help from
my family and friends
DA 0.73 –
4. I felt that I was just as good as
other people (R)
PA – 0.63
5. I had trouble keeping my mind
on what I was doing
SR 0.58 –
6. I felt depressed DA 0.64 –
7. I felt that everything I did was
an effort. EXCLUDED
SR – (-0.40)
8. I felt hopeful about the future
(R)
PA – 0.68
9. I thought my life had been a
failure
DA 0.69 –
10. I felt fearful DA 0.67 –
11. My sleep was restless SR 0.62 –
12. I was happy (R) PA – 0.80
13. I talked less than usual SR 0.49 –
14. I felt lonely DA 0.67 –
15. People were unfriendly IP 0.64 –
16. I enjoyed life (R) PA – 0.64
17. I had crying spells DA 0.65 –
18. I felt sad DA 0.66 –
19. I felt that people disliked me IP 0.79 –
20. I could not get ‘‘going’’ SR 0.62 –
Loadings inferior to 0.395 are not included. Constructs refer to the
factors identified by Radloff [12]: DA depressed affect, SR somatic/
retardation, PA positive affect, and IP interpersonal. Items with the
mention (R) were reverse-coded prior to statistical analyses
Soc Psychiatry Psychiatr Epidemiol (2014) 49:459–465 463
123
4. Bener A, Ghuloum S, Abou-Saleh MT (2012) Prevalence,
symptom patterns and comorbidity of anxiety and depressive
disorders in primary care in Qatar. Soc Psychiatry Psychiatr
Epidemiol 47:439–446. doi:10.1007/s00127-011-0349-9
5. Kleinman A (1977) Depression, somatization and the ‘‘new cross-
cultural psychiatry’’. Soc Sci Med 11:3–9. doi:10.1016/0037-
7856(77)90138-x
6. Batniji R, Van Ommeren M, Saraceno B (2006) Mental and
social health in disasters: relating qualitative social science
research and the sphere standard. Soc Sci Med 62:1853–1864.
doi:10.1016/j.socscimed.2005.08.050
7. Marsella AJ (2003) Cultural aspects of depressive experience and
disorders. Online Read Psychol Cult Unit 10
8. Nichter M (1981) Idioms of distress: alternatives in the expres-
sion of psychosocial distress: a case study from South India. Cult
Med Psychiatry 5:379–408
9. Nichter M (2010) Idioms of distress revisited. Cult Med Psy-
chiatry 34:401–416
10. Bass JK, Bolton PA, Murray LK (2007) Do not forget culture
when studying mental health. Lancet 370:918–919. doi:10.1016/
S0140-6736(07)61426-3
11. Hollifield M (2002) Accurate measurement in cultural psychiatry:
will we pay the costs? Transcult Psychiatry 39:419–421. doi:10.
1177/136346150203900401
12. Radloff LS (1977) The CES-D scale. Appl Psychol Meas
1:385–401. doi:10.1177/014662167700100306
13. Snook SC, Gorsuch RL (1989) Component analysis versus
common factor analysis: a Monte Carlo study. Psychol Bull
106:148–154. doi:10.1037/0033-2909.106.1.148
14. Velicer WF, Fava JL (1998) Affects of variable and subject
sampling on factor pattern recovery. Psychol Methods
3:231–251. doi:10.1037/1082-989x.3.2.231
15. Shafer AB (2006) Meta-analysis of the factor structures of four
depression questionnaires: beck, CES-D, Hamilton, and Zung.
J Clin Psychol 62:123–146. doi:10.1002/jclp.20213
16. Edwards MC, Cheavens JS, Heiy JE et al (2010) A reexamination
of the factor structure of the Center for Epidemiologic Studies
Depression scale: is a one-factor model plausible? Psychol Assess
22:711–715. doi:10.1037/a0019917
17. Schroevers M, Sanderman R, van-Sonderen E et al (2000) The
evaluation of the Center for Epidemiologic Studies Depression
(CES-D) scale: depressed and positive affect in cancer patients
and healthy reference subjects. Qual Life Res 9:1015–1029.
doi:10.1023/a:1016673003237
18. Crockett L, Randall BA, Shen Y et al (2005) Measurement
equivalence of the Center for Epidemiological Studies Depres-
sion scale for latino and anglo adolescents: a national study.
J Consult Clin Psychol 73:47–58. doi:10.1037/0022-006x.73.1.47
19. Golding JM, Aneshensel CS (1989) Factor structure of the Center
for Epidemiologic Studies Depression scale among mexican
Americans and non-hispanic whites. Psychol Assess 1:163–168.
doi:10.1037/1040-3590.1.3.163
20. Posner SF, Stewart AL, Marın G et al (2001) Factor variability of
the Center for Epidemiological Studies Depression scale (CES-D)
among urban Latinos. Ethn Health 6:137–144. doi:10.1080/
13557850120068469
21. Roberts RE (1980) Reliability of the CES-D scale in different
ethnic contexts. Psychiatry Res 2:125–134. doi:10.1016/0165-
1781(80)90069-4
22. Noh S, Kaspar V, Xinyin C (1998) Measuring depression in
Korean immigrants: assessing validity of the translated korean
version of CES-D scale. Cross Cult Res 32:358–377. doi:10.
1177/106939719803200403
23. Campo-Arias A, Dıaz-Martinez LA, Rueda-Jaimes GE et al
(2007) Psychometric properties of the CES-D scale among
Colombian adults from the general population. Rev Colomb
Psiquiat 36:664–674
24. Long Foley K, Reed PS, Mutran EJ et al (2002) Measurement
adequacy of the CES-D among a sample of older African
Americans. Psychiatry Res 109:61–69. doi:10.1016/s0165-
1781(01)00360-2
25. Kuo WH (1984) Prevalence of depression among Asian-Ameri-
cans. J Nerv Ment Dis 172:449–457. doi:10.1097/00005053-
198408000-00002
26. DaSilveira DX, Jorge MR (2002) Reliability and factor structure
of the Brazilian version of the Center for Epidemiologic Studies-
Depression. Psychol Rep 91:865–874. doi:10.2466/pr0.2002.91.
3.865
27. Guarnaccia PJ, Angel R, Worobey JL (1989) The factor structure
of the CES-D in the Hispanic health and nutrition examination
survey: the influences of ethnicity, gender and language. Soc Sci
Med 29:85–94. doi:10.1016/0277-9536(89)90131-7
28. Stroup–Benham CA, Lawrence RH, Trevifio FM (1992) CES-D
factor structure among Mexican American and Puerto Rican
women from single- and couple-headed households. Hisp J Behav
Sci 14:310–326. doi:10.1177/07399863920143002
29. Gupta R, Yick A (2001) Validation of CES-D scale for older
Chinese immigrants. J Ment Health Aging 7:257–272
30. Ying Y (1988) Depressive symptomatology among Chinese-
Americans as measured by the CES-D. J Clin Psychol
44:739–746. doi:10.1002/1097-4679(198809)44:5\739:aid-
jclp2270440512[3.0.co;2-0
31. Leykin Y, Torres LD, Aguilera A et al (2011) Factor structure of
the CES-D in a sample of Spanish- and English-speaking smokers
on the Internet. Psychiatry Res 185:269–274. doi:10.1016/j.
psychres.2010.04.056
32. Ghubash R, Daradkeh TK, Al Naseri KS et al (2000) The per-
formance of the Center for Epidemiologic Study Depression scale
(CES-D) in an Arab female community. Int J Soc Psychiatry
46:241–249. doi:10.1177/002076400004600402
33. Fountoulakis K, Iacovides A, Kleanthous S et al (2001) Reliability,
validity and psychometric properties of the Greek translation of the
Center for Epidemiological Studies-Depression (CES-D) scale.
BMC Psychiatry 1:1–10. doi:10.1186/1471-244x-1-3
34. Manson SM, Ackerson LM, Dick RW et al (1990) Depressive
symptoms among American Indian adolescents: psychometric
characteristics of the Center for Epidemiologic Studies Depres-
sion scale (CES-D). Psychol Assess 2:231–237. doi:10.1037/
1040-3590.2.3.231
35. Somervell PD, Beals J, Kinzie JD et al (1992) Use of the CES-D
in an American Indian village. Cult Med Psychiatry 16:503–517.
doi:10.1007/bf00053590
36. Yen S, Robins CJ, Lin N (2000) A cross-cultural comparison of
depressive symptom manifestation: China and the United States.
J Consult Clin Psychol 68:993–999. doi:10.1037/0022-006x.68.6.
993
37. Edman JL, Danko GP, Andrade N et al (1999) Factor structure of
the CES-D (Center for Epidemiologic Studies Depression scale)
among Filipino-American adolescents. Soc Psychiatry Psychiatr
Epidemiol 34:211–215. doi:10.1007/s001270050135
38. Kazarian SS (2009) Validation of the Armenian Center for Epi-
demiological Studies Depression scale (CES-D) among ethnic
Armenians in Lebanon. Int J Soc Psychiatry 55:442–448. doi:10.
1177/0020764008100548
39. Lee SW, Stewart SM, Byrne BM et al (2008) Factor structure of
the Center for Epidemiological Studies Depression scale in Hong
Kong adolescents. J Pers Assess 90:175–184. doi:10.1080/
00223890701845385
40. DesForges A (1999) Leave none to tell the story: genocide in
Rwanda. New York
464 Soc Psychiatry Psychiatr Epidemiol (2014) 49:459–465
123
41. Bolton P, Neugebauer R, Ndogoni L (2002) Prevalence of
depression in rural Rwanda based on symptom and functional
criteria. J Nerv Ment Dis 190:631–637. doi:10.1097/00005053-
200209000-00009
42. MacCallum RC, Widaman KF, Zhang S, Hong S (1999) Sample
size in factor analysis. Psychol Methods 4:84–99. doi:10.1037/
1082-989X.4.1.84
43. Bryant FB, Yarnold PR (1995) Principal-components analysis
and exploratory and confirmatory factor analysis. In: Grimm LG,
Yarnold PR (eds) Reading and understanding multivariate sta-
tistics. American Psychological Association, Washington DC,
pp 99–136
44. Everitt BS (1975) Multivariate analysis: the need for data, and
other problems. Br J Psychiatry 126:237–240. doi:10.1192/bjp.
126.3.237
45. Kunce JT, Cook DW, Miller DE (1975) Random variables and
correlational overkill. Educ Psychol Meas 35:529–534. doi:10.
1177/001316447503500301
46. Nunnally J (1978) Psychometric theory. McGraw-Hill, New York
47. Eaton W, Muntaner C, Smith C et al (2004) Center for Epide-
miologic Studies Depression scale: review and revision (CESD
and CESDR). The use of psychological testing for treatment
planning and outcomes assessment. Erlbaum, Mahwah, NJ
48. Hutcheson, GD, Sofroniou, N (1999) The multivariate social
scientist: introductory statistics using generalized linear models.
SAGE Publications Limited
49. Kaiser HF (1974) An index of factorial simplicity. Psychometrika
39:31–36. doi:10.1007/BF02291575
50. Waller NG (2000) MicroFACT 2.0: a microcomputer factor
analysis program for ordered polytomous data and mainframe
size problems. Assessment Systems Corporation, St Paul, MN
51. Tataryn DJ, Wood JM, Gorsuch RL (1999) Setting the value of k
in promax: a Monte Carlo study. Educ Psychol Meas 59:384–391.
doi:10.1177/00131649921969938
52. Parker G, Cheah YC, Roy K (2001) Do the Chinese somatize
depression? a cross-cultural study. Soc Psychiatry Psychiatr Ep-
idemiol 36:287–293. doi:10.1007/s001270170046
53. Jayawickreme N, Jayawickreme E, Atanasov PD et al (2012) Are
culturally specific measures of trauma-related anxiety and
depression needed? The case of Sri Lanka. Psychol Assess
24:781–800. doi:10.1037/a0027564
54. Zandi T, Havenaar JM, Limburg–Okken AG et al (2008) The need
for culture sensitive diagnostic procedures. Soc Psychiatry Psy-
chiatr Epidemiol 43:244–250. doi:10.1007/s00127-007-0290-0
55. Bolton P, Bass J, Betancourt T et al (2007) Interventions for
depression symptoms among adolescent survivors of war and
displacement in northern Uganda: a randomized controlled trial.
JAMA 298:519–527. doi:10.1001/jama.298.5.519
56. Bolton P, Bass J, Neugebauer R et al (2003) Group interpersonal
psychotherapy for depression in rural Uganda: a randomized
controlled trial. JAMA 289:3117–3124. doi:10.1001/jama.289.
23.3117
57. Bolton P (2001) Local perceptions of the mental health effects of
the Rwandan genocide. J Nerv Ment Dis 189:243–248
58. Association AP (2000) Diagnostic and Statistical Manual of
Mental Disorders Fourth, Revised edn. American Psychiatric
Association, Washington DC
59. Bolton P, Tang AM (2002) An alternative approach to cross-
cultural function assessment. Soc Psychiatry Psychiatr Epidemiol
37:537–543. doi:10.1007/s00127-002-0580-5
60. Kleinman A (2004) Culture and depression. N Engl J Med
351:951–953. doi:10.1056/NEJMp048078
Soc Psychiatry Psychiatr Epidemiol (2014) 49:459–465 465
123