Printer Friendly

Validation of the multiple language versions of the Hopkins Symptom Checklist-37 for refugee adolescents.


During the last 20 years, there has been a substantial influx of immigrants and refugees to Europe (Eurostat, 2002). This has led to many more schools in Dutch and Belgian metropolitan areas providing education for children and adolescents who do not speak the language of the host country fluently. The transition from one country to another implies changes and difficulties such as the loss of social networks, changes in work status as well as encountering discrimination that can be very distressing (Vinokurov, Trickett, & Birman, 2002).

The traditional higher-order latent structure of internalizing (overcontrolling) and externalizing (undercontrolling) problems have for many years been a useful framework for emotional distress and maladaptive behaviors of children and adolescents (e.g., Achenbach & Edelbrock, 1978; Southam-Gerow & Kendall, 2002). In recent years, Krueger and colleagues (2001) confirmed the usefulness of this dichotomy in explaining the covariance among adult mental health and personality disorders. Moreover, Miller and colleagues (2003) have put forward an internalizing/externalizing model to explain the reactions of traumatic stress among adult combat veterans. The internalizing/ externalizing model seems to provide an adequate framework in which traumatic stress reactions and/or (comorbid) psychopathology can be understood.

The literature on the mental health of refugee adolescents depicts a high prevalence of psychosocial symptoms reported by refugee adolescents (Felsman, Leong, Johnson, & Felsman, 1990; Sack et al., 1993; Sourander, 1998; Smith, Perrin, Yule, Hacam, & Stuvland, 2002). The most frequently reported symptoms are somatic complaints, anxiety, depression, and (post)traumatic stress reactions. Unaccompanied refugee minors (URM) run an especially high risk for developing psychopathology due to separation from primary caregivers, exposure to sequential stressful events, limited educational opportunities, and conditions in asylum centers during a very vulnerable developmental period (Felsman et al., 1990; Sourander, 1998). High comorbidity has been documented between reactions to traumatic stress and other disorders such as depression (Sack et al., 1993) and anxiety (Warshaw et al., 1993). Significant adults in the lives of adolescents (i.e., caregivers, teachers) often report a lower prevalence of internalizing problems than do the adolescents themselves since they have difficulty determining the extent to which the adolescents suffer from psychological distress.

On the other hand, perceiving the disturbing nature of externalizing problems is not difficult. Adolescents with conduct problems have been found to be referred much sooner and more often to professional mental healthcare services than adolescents with internalizing problems (Wu et al., 1999). The literature on conduct problems of refugee adolescents is very limited. Allwood, Bell-Dolan, & Husain (2002) found a strong association between witnessing of organized violence and exhibiting aggressive behavior. Jensen and Shaw (1993) suggest that adolescents who have witnessed or taken part in a war are more likely to show delinquent or anti-social behavior. This opinion is, however, not supported in four studies which evaluated the delinquent and aggressive behaviors of refugee adolescents (Raboteg-Saric, Zuzul, & Kerestes, 1994; Mollica et al., 1997; Rousseau, Drapeau, & Corin, 1998; Sourander, 1998). Different authors (i.e., Pynoos & Nader, 1993) report that adolescents may temporarily show increased risk behavior following the witnessing/experiencing of a traumatic event. Moreover, several studies have found high levels of comorbidity between externalizing behavior and experiencing traumatic stress reactions among American adolescents (Deykin & Buka, 1997; Wozniak et al., 1999).

The "pathway" to professional mental healthcare for refugee adolescents has more barriers than for native adolescents in host countries (e.g., Howard & Hodes, 2000). There is sufficient evidence in the literature suggesting that young people in general who are in need of psychological support or treatment do not receive it (i.e., U.S. Department of Health and Human Services, 1999; Cuffe, Waller, Cuccaro, Pumareiga, & Garrison, 1995) or only when the symptoms have progressed and are perceived by significant adults in their lives (e.g., Wu et al., 1999). The psychological suffering of Unaccompanied Refugee Minors (URM) can go completely unnoticed due to the absence of parents or permanent caregivers, language difficulties, and living in minimally adult-supervised residential settings. Mental healthcare professionals in host countries are often hindered in acquiring accurate information concerning the mental health status of refugee adolescents due, in part, to language difficulties, lack of medical/psychological background information and reliable and valid translated diagnostic instruments.

Approaching refugee youth with long psychological questionnaires/interviews can be overwhelming (Barenbaum, Ruchkin, & Schwab-Stone, 2004). Brief, translated psychological instruments that measure reactions associated with traumatic stress could be of great assistance to mental healthcare professionals in the process of screening, diagnosing, and monitoring the mental health status of this specific population. An increasing number of studies have been conducted with refugee culturally homogeneous samples (Papageorgiu et al., 2000; Thabet & Vostanis, 1999) or samples from two different countries of origin (Rousseau & Drapeau, 1998). Smith and colleagues (2003) have validated the Revised-Impact of Events Scale with older children from Bosnia. However, the same instrument yielded less reliable results with older children and adolescents from Rwanda (Dyregrov, Gupya, Gjestad, & Mukanoheli, 2000) which clearly illustrates that a measure that has been validated for one immigrant or refugee population does not implicitly infer that it is valid and reliable for all refugee and immigrant populations.

Furthermore, a limited number of reliable and valid diagnostic instruments have established norms for immigrant/refugee adolescents to measure psychosocial distress and maladaptive behaviors. The Youth Self Report (Achenbach, 1991) and Strengths and Difficulties questionnaire (Goodman, 1997) have been used with refugee adolescents from certain countries to measures emotional and behavior problems (Fazel & Stein, 2002; Mollica et al., 1997). However, as far as it is known by the authors, these checklists as well as others used with refugee children (i.e., Smith et al., 2002) have not been validated for culturally diverse adolescent populations following the five dimensions of equivalence for cross-cultural validation of an instrument proposed by Flaherty et al. (1988).

Aims of the Study

Modifications were made to one of the well-known instruments that has been used with refugees/non-western populations of adults over the last 15 years, the Hopkins-Symptom Checklist-25 (Lie, 2002; Mollica et al., 1987) to make it accessible to immigrant/refugee adolescents from a variety of cultures. The objective of this study was to provide preliminary information concerning the psychometric properties of the modified version of the HSCL-25 (Winokur, Winokur, Rickels, & Cox, 1984) the HSCL-37 for adolescents. Twelve externalizing items have been added to the 25 original items related to internalizing problems in order to measure externalizing problematic behavior in adolescents which may be trauma-related.


Context of the Validation Study

In the years preceding 2001, there was a dramatic increase in the number of URM living in the Netherlands, peaking at 15,000 in 2001. Because there was (and still is) a lack of research on the mental health and service utilization of URM, a national and longitudinal research project "Unaccompanied Refugee Minors and Dutch Mental Health Care Services" was started among unaccompanied refugee minors living in The Netherlands and their guardians, teachers, and professional mental healthcare providers. A secondary aim of the project was to validate and standardize screening instruments for this specific population group. It was also possible to administer the HSCL-37A in an independent research project conducted by the Department of Orthopedagogics, Ghent University, Belgium that examined whether being unaccompanied is a risk factor for refugee children and adolescents to develop emotional and behavioral problems.


A national, longitudinal study was carried out with URM living in the Netherlands. Approximately 4,000 URM were randomly selected from the Central Register of Nidos. Information about the study and permission waivers (available in translated versions) were sent to the guardians to discuss with the URM. Both the minor and his/her guardian gave written permission for the URM to participate. Roughly 2,300 URM permission waivers were returned; 57% wished to participate, 15% refused, 12% did not participate for a wide range of practical reasons, 9% were transferred to a different residential setting, and 7% turned out to be untraceable. A total of 920 URM were present for participation. The final sample was statistically representative (data not shown) in all of the main characteristics (age, gender, country of origin, type of residential setting) of the total URM population aged 12 to 18 years in 2002 in the Netherlands. The URM came from 48 countries, predominantly Angola (43%), Sierra Leone (10%), and China (8%). Two-thirds of the sample had lived in the Netherlands for a period of 18 months or less; 45% of the URM sample had received 5 years or less of formal education in their country of origin. A follow-up (63% of the original sample participated) was conducted one year after the first assessment. An interview regarding mental health care was individually administered. At least three research assistants administered the questionnaires during one hour to groups of 10 URM.

Dutch national sample (n = 1,059). Pupils from ten secondary and three tertiary trade schools throughout the Netherlands (schools had also taken part in the URM study) participated and functioned as a control group for the URM sample. Two weeks prior to administration of the instruments, informed consent letters were sent to the parents and adolescents asking for voluntary and anonymous participation (27 students abstained from participation). Assessment of the Dutch sample took approximately 15 minutes.

Belgian immigrant/refugee adolescents residing in The Netherlands (n = 1,294). A large study was carried out with non-Dutch speaking immigrant adolescents in Flanders (Belgium) during November 2002 to May 2003. The adolescents were from 111 countries, predominantly Morocco (14%), Ghana (11%), and Turkey (9%). All schools received standard informed consent letters (translated versions were available) asking parents and students for voluntary and anonymous participation. In 2002, there were 42 secondary schools in Flanders which provided education for recently immigrated adolescents. Thirty-four schools were randomly chosen to participate in the study of which none declined; 65% of the number of recently (less than 1 year) immigrated adolescents (immigrants and refugees) in Flanders between 13-18 years of age, participated in the study. Only 1 student abstained from participation who was present on the day of assessment. There was a continuous steam of new students during the year, which made it difficult to test the entire population. No attempt was made to test adolescents who were not present on assessment day. The assessment took place (1 hour) under supervision of two research assistants.

Belgian national sample (n = 617). A control group of Belgian adolescents participated between January 2003 and May 2003 for the Belgium immigrant/refugee study. From the six Flemish provinces, 17 secondary schools were randomly selected to participate in the study. All schools received standard informed consent letters asking parents and students for voluntary and anonymous participation. To assemble a well-balanced normative sample of the Flanders adolescent population, the same percentage of Belgian adolescents and immigrant/refugee adolescents per province took part in the study. In this way, there would not be an overrepresentation of Belgian adolescents living in urban or rural areas. Furthermore, the proportions for the different age and gender groups of the Belgian adolescents were carefully matched with those of the immigrant/refugee sample so that the two groups were similar on these variables. Finally, per province, the secondary schools chosen had students that were following all three educational track levels (trade, occupational, and preparatory for university). No Flemish student refrained from participating.


The HSCL-37A (Bean, Derluyn, Eurelings-Bontekoe, & Spinhoven 2004a) was modified to render the instrument multi-cultural and adolescent friendly. A 4-point rating scale in literal terms (not/never = 1, sometimes = 2, often = 3, always = 4) was used to indicate the severity of symptoms, feelings or behaviors. The literal terms of the Likert scale was improved by placing different colored balls increasing in size above the literal rating scale to clarify "quantity" of feelings. Second, items were (if needed) simplified to adapt the questionnaire to the (Dutch) language abilities of this population based on a vocabulary list developed for immigrant adolescents to the Netherlands, and third, the questionnaires were translated and presented in a bilingual form (Dutch-foreign language). It was necessary to have the questionnaires in this form because many of the refugee adolescents had limited written knowledge of their own language and learned the Dutch language quickly, allowing them to use both languages in order to better comprehend the item.

The HSCL-37A, SLE, and RATS questionnaires were all translated into the most prevalent languages of URM in the Netherlands: Albanian, Amharic, Arabic, Baldini, Chinese, Dari, Dutch, English, Farsi, French, German, Mongolian, Portuguese, Russian, Servo-Croation, Soerani, Somalil, Spanish and Turkish. No written back-translations were done in this study. All written forward translations were done by professional translators. Every translation was controlled for grammatical and idiomatic errors on two different occasions by two different translators. The translated questionnaires were reviewed orally with professional interpreters who where regularly involved in treatment sessions of traumatized adult refugees to control the quality of the translations, to ensure that the original meaning was conveyed in the items, and to attempt to achieve semantic equivalence of the HSCL-37A. All of the instruments were tested in a pilot study. If an adolescent filled in a bilingual version of the instruments (Dutch/foreign language), the foreign language of the questionnaire was recorded. If an adolescent completed the Dutch language version only, Dutch was recorded as the language of the questionnaire(s).

The internalizing scale of the HSCL-37A can be divided into ten anxiety questions (items 1, 2, 5, 9, 12, 16, 19, 22, 26, 29) and fifteen depression questions (items 6, 10, 13, 15, 17, 20, 23, 24, 27, 30, 31, 32, 33, 35, 36). Twelve items were added to measure externalizing behavior which may be trauma-related. The items bullies, steal things, intentionally hurting someone, starts fights, destroying others' property correspond with five criteria from the diagnosis for a Conduct Disorder. The item easily angered and argues often correspond with two criteria from the diagnosis for an Oppositional Defiant Disorder according to the DSM-IV (American Psychiatric Association, 1994). The other five items are related to substance abuse (use of alcohol on the weekend, use of alcohol through the week, smoking cigarettes, use of sedatives, and use of drugs). The scale for externalizing behavior (items 3, 7, 11, 14, 18, 21, 34, 3, 8, 25, 28, 37; min = 12, max = 48) can be used to attain a total score for externalizing behavior. The total score of the HSCL-37A consists of all 37 items (min = 37, max. = 148). Percentile scores and severity classifications are available in the user's manual (Bean et al., 2004a).

The Stressful Live Events questionnaire (SLE) (Bean, Derluyn, Eurelings-Bontekoe, Broekaert, & Spinhoven, 2006; Bean et al., 2004b) consists of 12 dichotomous (yes/no) questions and an open question on the occurrence of stressful life events of relevance for adolescent refugee minors (e.g., "Have you ever experienced a war or an armed military conflict going on around you in your country of birth?" or "Has someone ever hit, kicked, shot at or tried in some other way to physically hurt you?"). Experiencing a traumatic event is the first criterion of the A cluster of the SDM-IV for PTSD (APA, 1994). The overall average total score of 6.5 of the SLE has been validated in 5 independent studies (Bean et al., 2004b).

The Reactions of Adolescents to Traumatic Stress (RATS) (Bean et al., 2006; Bean et al., 2004c) is a self-report questionnaire developed to assess posttraumatic stress reactions defined in the DSM-IV (APA, 1994) with culturally diverse adolescents. The twenty-two item scale can be divided into three subscales: intrusion (six items), avoidance (nine items), and hyper-arousal (seven items) which correspond to the 17 criteria in a PTSD diagnosis. Internal reliability for the URM sample for the total score, intrusion, numbing/avoidance, and hyperarousal was correspondingly, .88, .85, .69, and .73. Twelve-month test-retest reliability was for total score .61 (p < .001). Using a confirmatory factor analysis, the three-factor structure was verified in the URM sample with a loss of only 3% of the explained variance. Similar results were confirmed in the other 3 samples.

The self-report version for 11- to 16-year olds of the Strengths and Difficulties Questionnaire (SDQ) (Goodman, 1997) is a screening questionnaire that measures twenty-five attributes divided into five subscales: emotional symptoms, conduct problems, inattention-hyperactivity, peer problems, and pro-social behavior. Research shows that the SDQ has an acceptable reliability and validity (Goodman, 2001). The SDQ was also available in the languages of the immigrant/ refugee adolescents in Belgium. In this study, the internal reliability (Cronbach's alpha) of the total score of the multiple language versions of the SDQ ranged from .62-.79, with an average value of .63. Average sub-scales reliability was low-to unacceptable .68, emotion symptoms -.42, peer problems for the total population.

Indicators of Psychopathology

The criteria "referral" and "utilization of MHC" have been documented as being important in the evaluation of psychopathology in children and adolescents (i.e., Verhulst & Van der Ende, 1997). For this reason, (a) self-reported need for mental health care (MHC); (b) need for professional MHC for the URM, evaluated by the legal guardian; (c) need for professional MHC for the URM, evaluated by the teacher; (d) self-reported utilization of MHC by URM; and (e) referral to MHC services by a legal guardian were utilized as external criteria of psychopathology. The URM were individually interviewed in Dutch about their needs and mental health use. They were also able to read the questions in one of the languages mentioned previously. Guardians and teachers received short questionnaires on need for professional MHC and referral to MHC services by URM which they completed and returned by mail.

A strong, significant, and positive relationship should exist between the HSCL-37A total and the SDQ total scores because these two scales measure the same construct. There should also be a strong association between the HSCL-37A internalizing scale and the RATS because as reported earlier, high comorbidity has been found between PTSD on the one hand and general anxiety/depression on the other. The correlation between the externalizing score (measuring trauma-associated acting out behaviour) of the HSCL-37A and RATS scores should be present but weak. The total SLE score should be positively related to the total score of the HSCL-37A and subscales, since trauma is related to psychopathology (Allwood et al., 2002; Tiet et al., 1998).


Ethical approval for both Belgian studies was given by the Ethics Committee of the Faculty of Psychology and Educational Sciences, Ghent University and by the Medical Ethics Committee of the Leiden University Medical Center, Leiden University to conduct the Dutch URM study.

Testing of the Belgian and Dutch normative samples took place in small groups (10-25 young people) during school time. The URM were assessed at schools, if possible. Approximately 20% of the URM were not tested at schools. URM were also assessed (in groups of 10) at the regional offices of Nidos, reception centers for refugees, and residential settings. Demographic information on the URM in the Netherlands was supplied by the Nidos Foundation (legal guardian of all of the URM living in the Netherlands). The rest took part anonymously and answered written questions that provided their demographic characteristics.

Data Analysis

Descriptive statistics were used to provide summary descriptions of the sociodemographic characteristics of the sample. Confirmatory factor analyses, per language version, were calculated using the Multiple Group Method (MGM) procedure of the Simultaneous Components Analysis (SCA) (Kiers, 1990) to verify the factorial validity of the HSCL-37A (all cases with missing data were removed). MGM is closely related to the rotation of component weights to perfect congruence and the cross-validation of components weights (Ten Berge, 1986). SCA is based on the same set of weights for the variables in all populations enabling conclusions to be drawn on the common components found across the samples. It is not a formal statistical test, such as the Maximum Likelihood estimation method; however, this is not a serious objection because the null hypothesis of a factor model based on a small number of factors is invariably false as has been known since Browne (1969, p. 385). Failure to reject it merely means that the sample size has been too small (see McCrae, Zonderman, Costa, Bond, & Paunonen, 1996 for a discussion). Internal consistency of the total scale and subscales of the HSCL-37A was calculated with Cronbach's alpha. Test-retest reliability was calculated for a twelve-month interval for the URM sample only (n = 519). Pearson's product-moment correlations (two-tailed) were used to study the association between total and subscale scores of the HSCL-37A and the scores on the remaining questionnaires. Differences between groups were determined by using ANOVA's and effect sizes Cohen's d (Cohen, 1988). A maximum of 10% of missing items was allowed in order to still be able to extrapolate the total or subscale scores of all scales. (See Table 1.)


The factor structure of the HSCL-37A was tested with the Simultaneous Components Analysis (SCA). The scale consisting of Internalizing items was established based on the results of a previous factor analysis on a large item pool and the opinions of several experienced clinicians (Derogatis et al., 1974). The externalizing items made up the second scale. For the total sample, a principal component analysis (PCA) was used with Varimax rotation (oblique) to simple structure which allowed for correlation between the two factors (Kiers, 1990) which yielded a model that explained 33.1% of the total variance. The SCA-MGM analysis based on the two a priori factors showed that the multiple group components explained 32.7% of the variance, implying a small acceptable discrepancy of only .4%.

Separate MGM analyses were conducted on the Portuguese, French, Chinese, English, Arabic, Dutch, and Russian language versions. The amount of variance that was lost in enforcing the a priori factor structure in comparison to the results of an explorative PCA in the separate language versions was very limited, ranging from 2.2% in the Chinese version to .4% in the Dutch version. Due to the limited number of completed questionnaires (n < 100) in Badini, Servo-Croatian, Albanese, Turkish, Soerani, Dari, Farsi, Amharic, Somali, and Mongolian, no individual MGMs could be conducted for these languages. The two-factor model is confirmed in all the separate MGM analyses per language. (See Table 2.)

Internal Consistency

The internal consistency (Cronbach's alpha) of the HSCL-37A indicates a high degree of homogeneity among items comprising the total and subscales in the separate language versions. The internal consistency of the total scale of the HSCL-37A in the total sample was .90 and of the individual language versions ranged from .95 to .84. This is an exceptionally high alpha, despite the high degree of heterogeneity in the samples. The alphas for the subscales and apart language versions can be found in Table 2. It should be noted in general that the alpha for the internalizing sub-scale were higher than those for the externalizing sub-scale.

Temporal Stability

The test-retest scores are utilized to provide an indication of scale stability and consistency over time. The coefficients for a one-year period show that the HSCL-37A scales are reasonably stable (r > .50) over time in measuring internalizing and externalizing behavior (see Table 2) not deviating from findings of other studies with the same time interval (see Cheng & Nicholas, 1998 for a discussion).

Content validity is a measure of the relevance of the items with regard to that behavior which they aim to measure. The HSCL-37A claims to measure internalizing (anxiety and depression symptoms) and externalizing behavior. The choice of items to measure anxiety and depression was based on the expertise of clinicians with experience in the treatment of patients with anxiety and depression (Derogatis et al., 1974). All items of the HSCL-37A correspond with the DSM-IV criteria for anxiety, depression, and behavior symptoms. The 12 externalizing items correspond with the five criteria of conduct disorder and the two criteria of the oppositional-defiant disorder, as defined in the DSM-IV (APA, 1994). The content validity of the HSCL-37A is good. (See Table 3.)

Construct validity is a measure of the relationship between the instrument and variables that, on theoretical grounds, are expected to correlate with the measured variable. Three processes are used to establish construct validity: (1) convergent validity; high correlations between a particular scale and others that in theory measure the same construct, (2) discriminant validity; low associations between the scale under study and other measures that should theoretically not be related, and (3) factorial validity; supports the theory-based grouping of items when a particular construct is complex. Table 3 shows the intercorrelations (two-tailed) between the HSCL-37A total and subscale scores, the RATS total score, and the SLE total score for the URM sample and native Dutch sample. In Table 3, the intercorrelations are presented between the SDQ total score (the SDQ administered only in the Belgian studies), the RATS total score, and the SLE total score for the immigrant/refugee sample and native Belgian sample.

As hypothesized, the HSCL-37A total scores and internalizing scale scores show significant and positive correlations with the RATS total scores, SLE total scores, and SDQ total scores. The significant and positive relationship between the externalizing scale scores and the other scale scores is weaker, but still present. The relationship between the total, internalizing, and externalizing scores on the HSCL-37A and the total number of experienced events on the SLE is significant and positive. These findings are applicable to all samples.

The mean scores of girls are expected to be significantly higher than that of boys. Girls reported significantly higher internalizing (F(1,3646) = 74.96, p < .001, d = .29) and externalizing mean scores (F (1,3718) = 25.03, p < .001, d = .17) than did boys). There are contradictory findings in the literature concerning age and emotional distress. Age, in this study, seemed to play a small role with respect to total mean scores, with older adolescents ([less than or equal to] 14, 15 years) for internalizing (F(3,3597) = 30.50, p < .001, d = .39.23) and externalizing problems (F(3,3668) = 16.96, p < .001, d =.31-.10). Because of the numerous risk factors overshadowing the lives of URM, it was expected that URM would score significantly higher than immigrant/refugee and native adolescents living with at least one parent. This expectation is partly confirmed. URM reported significantly higher internalizing mean scores (F(2,3661) = 269.24, p < .001, d = .87-.78) on the HSCL-37A than the immigrants/refugees and natives, but significantly lower externalizing mean scores than native adolescents (F(2,3733) = 273.37, p < .001, d = .72). (See Table 4.)

Criterion validity refers to the association between the instrument with some form of external or outside criterion that is supposed to measure the same construct. Criterion validity can be further divided into two types: predictive and concurrent. Only the concurrent validity of the HSCL-37A was addressed in this study. Five indicators of psychopathology were utilized as external criteria (see Indicators of Psychopathology description in Questionnaires section). The results shown in Table 4 suggest that the HSCL-37A discriminated well, consistently, and significantly between URM that report having a need for psychosocial help and URM that did not report having a need for psychosocial help.


The results indicate that the HSCL-37A is a psychometrically sound screening measure of internalizing and externalizing problems experienced by a heterogeneous population of refugee adolescents. The data were collected among four independent samples from across the Netherlands and Flemish Belgium. The psychometric properties of the HSCL-37A demonstrate invariance of factor structure in a heterogeneous sample, strong reliability, and good validity which is remarkable considering the diversity of the populations.

The layout modifications (bilingual and visual/literal rating scale) of the instrument made the HSCL-37A comprehensible for adolescents from a variety of cultures. In a small number of individual cases, lengthy explanations of the meaning/nuances of the items were necessary, especially with "almost" illiterate adolescents (1 per group of 25 refugees/immigrants). It is not clear if errors in understanding the question might not be visible in the data. Only the Spanish version of the HSCL-37A had obvious less internal consistency on all subscales. This could be due to the fact that the translation was in European Spanish and adolescents came from South American countries which speak a different dialect of Spanish. European Spanish was used because of the wide differences in dialects in all Spanish speaking countries (the same holds for American English and European French which did not show lower reliability levels).

The two factors showed strong reliability and good validity considering the diversity of the sample populations (for example, adolescents from 35 different countries completed in the French version, adolescents from 57 countries completed the English version, and adolescents from 20 different countries completed the Arabic version). The preliminary validity findings suggest that the HSCL-37A can discriminate consistently and significantly between refugee adolescents that do need to utilize MHC services and those who do not.

The brevity of the HSCL-37A takes into account the importance of not overburdening apprehensive adolescents and allows for quick, repeated measurements to assist in determining initial and enduring refugee adolescent symptomatology. When the HSCL-37A is used with the SLE and RATS questionnaires, a preliminary assessment of the global mental health of refugee adolescents can be reliably assessed. In all settings, one must be aware that the instrument may trigger emotional distress. Therefore, adequate crisis and/or follow-up MHC should be arranged prior to administration to protect the integrity of the adolescents. The HSCL-37A is not meant to be used alone as a diagnostic instrument for internalizing distress or behavioral problems. Clinical observations and additional assessment are important in establishing a valid diagnosis and making treatment recommendations.

There were several methodological challenges of this study. Written back-translations of the language versions were not done, deviating from standard protocol which can be seen as a limitation of the study. Back-translation is the method used to verify semantic equivalence of translated measures (see Mallinckrodt & Wang, 2004). However, a back-translation does not guarantee that the content equivalence of the translated instrument has been established (Flaherty et al., 1988). Great effort was made to ensure the content equivalence of the items of the HSCL-37A for different cultures.

The number of instruments used were kept to a minimum for a number of reasons: (a) the short attention span of the refugee adolescents, (b) the time needed to explain and administer the three instruments was about 15 minutes of the total testing time, (c) the substantial amount of time and effort taken by the refugee adolescents to complete only three questionnaires, and (d) the ethical issues related to the administration of long instruments which might induce emotional distress in severely traumatized individuals. Additional measures would have enhanced the quality of the study and would have been useful in determining the divergent validity of the HSCL-37A which will need to be evaluated in future studies.

The stability (test-retest) of the HSCL-37A was calculated over a longer interval (12 months) than the usual 8 weeks resulting in a lower temporal stability than is desired. However, it could be expected after one year that many changes (developmental, stressful life events, transfers, change in residential status, and therapeutic interventions) would have taken place in the lives of URM which could have led to even lower stability levels.

Because no standardized diagnostic interview was utilized in this study, the sensitivity and specificity of the HSCL-37A could not be evaluated. Preferably, a standardized diagnostic interview is used in combination with questionnaires to determine the presence and severity of psychopathology. However, "referral" of children and adolescents to psychiatric services has been used as a "golden standard" instead of a diagnostic interview (e.g., Nolan et al., 1996). It was not feasible in the URM study to administer a diagnostic interview for the reasons noted and that there is no validated psychiatric diagnostic interview available in all the languages of the adolescents who took part in this study. Furthermore, the use of diagnostic interviews itself invokes a host of methodological issues such as classifying culture-specific disorders and ensuring "the semantic and psycholinguistic equivalence of psychiatric symptoms across cultures" (Cheng, 2001). Even so, the preliminary validity findings suggest that the HSCL-37A can discriminate between adolescents that do and do not need mental health services.

Self-report questionnaires such as the HSCL-37A yield less diagnostic information than extensive structured interviews and therefore should be used only to indicate clinically elevated levels of internalizing and externalizing problems and not to diagnosis anxiety, depression or conduct disorder. Additional information should be collected regarding the mental health of the adolescent from the viewpoint of significant adults (caregivers/teachers) in the environment of the adolescent--information that is crucial in assessing the degree of impairment in daily functioning and the severity of the symptoms of adolescents.


Allwood, M. A., Bell-Dolan, D., & Hussain, S. A. (2002). Children's trauma and adjustment reactions to violent and non-violent war experiences. Journal of the American Academy of Child and Adolescent Psychiatry, 41, 450-457.

American Psychiatry Association (APA). (1994). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington, D.C.: American Psychiatric Association.

Barenbaum, J., Ruckin, V., & Schwab-Stone, M. (2004). The psychological aspects of children exposed to war: Practice and policy initiatives. Journal of Child Psychology and Psychiatry, 45, 41-62.

Bean, T., Eurelings-Bontekoe, E. H. M., Derluyn, I., & Spinhoven, P. (2004a). HSCL-37A User's Manual. Retrieved from the Web site of Centrum '45, September 1, 2004:

Bean, T., Eurelings-Bontekoe, E. H. M., Derluyn, I., & Spinhoven, P. (2004b). SLE User's Manual. Retrieved from the Web site of Centrum '45, September 1, 2004:

Bean, T., Eurelings-Bontekoe, E. H. M., Derluyn, I., & Spinhoven, P. (2004c). RATS User's Manual. Retrieved from the Website of Centrum '45, September 1, 2004: http:/

Bean, T., Derluyn, I., Eurelings-Bontekoe, E. H. M., Broekaert, E., & Spinhoven, P. (2006). Validation of the multiple language versions of the reactions of adolescents to traumatic stress questionnaires. Journal of Traumatic Stress, 19, 141-255.

Browne, M. W. (1969). Fitting the factor analysis model. Psychometrika, 34, 375-394.

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Erlbaum.

Cheng, A. T. (2001). Case definition and culture: Are people all the same? British Journal of Psychiatry, 179, 1-3.

Cheng, S. T., & Nicholas, H. P. (1998). Stability of symptom measures: Implications for stress research. Social Behavior and Personality, 26, 245-258.

Derogatis, L. R., Lipman, R., Rikels, K., Ulenhoth, E., & Covi, L. (1974). The Hopkins Symptom Checklist (HSCL): A self-report symptom inventory. Behavioral Science, 19, 1-15.

Dyregrov, A., Gupta, L., Gjestad, R., & Mukanaheli, E. (2000). Trauma exposure and psychological reactions to genocide among Rwandan children. Journal of Traumatic Stress, 13, 3-21.

European Commission. Eurostat (2002). European social statistics (p. 73). Migration Luxemborg.

Fazel, M., & Stein, A. (2003). Mental health of refugee children: Comparative study. British Medical Journal, 27, 134.

Felsman, J., Leong, F., Johnson, M., & Felsman, I. (1990). Estimates of psychological distress among Vietnamese refugees: Adolescents, unaccompanied minors, and young adults. Social Science & Medicine, 31, 1251-1256.

Flaherty, J. A., Gaviria, F. M., Pathak, D., Mitchell, T., Wintrob, R., & Richman, J. A. (1988). Developing instruments for cross-cultural psychiatric research. Journal of Nervous and Mental Disease, 176, 257-263.

Howard, M., & Hodes, M. (2000). Psychopathology, adversity, and service utilization of young refugees. Journal of the American Academy of Child & Adolescents Psychiatry, 39, 368-377.

Goodman, R. (1997). The Strengths and Difficulties Questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38, 581-586.

Goodman, R. (2001). Psychometric properties of the strengths and difficulties questionnaire (SDQ). Journal of the American Academy of Child & Adolescent Psychiatry, 40, 1337-1345.

Guttman, L. (1952). Multiple group methods for common-factor analysis: Their basis, computation and interpretation. Psychometrika, 17, 209-222.

Kiers, H. A. L. (1990). User's Manual SCA: A program for Simultaneous Components Analysis of variables measured in two or more populations. Groningen: ProGAMMA.

Krueger, R. F., McGue, M., & Iacono, W. G. (2001). The higher-order structure of common DSM mental disorders: Internalization, externalization, and their connections to personality. Personality and Individual Differences, 30, 1245-1259.

Lie, B. (2002). A 3-year follow-up study of psychosocial functioning and general symptoms in settled refugees. Acta Psychiatrica Scandinavica, 106, 415-425.

Mallinckrodt, B., & Wang, C. C. (2004). Quantitative methods for verifying semantic equivalence of translated research instruments: A Chinese version of the Experiences in Close Relationships Scale. Journal of Counseling Psychology, 51, 368-379.

McCrae, R. R., Zonderman, A. B., Costa, P. T., Bond, M. H., & Paunonen, S. V. (1996). Evaluating replicability of factors in the NEO personality inventory: Confirmatory factor: analysis versus Procrustes Rotation. Journal of Personality and Social Psychology, 70, 552-566.

Miller, M. W., Greif, J. L., & Smith, A. A. (2003). Multidimensional Personality Questionnaire profiles of veterans with traumatic combat exposure: Internalizing and externalizing subtypes. Psychological Assessment, 15, 205-215.

Mollica, R. F., Wyshak, G., de Marneffe, D., Khuon, F., & Lavelle, J. (1987). Indochinese versions of the Hopkins Symptom Checklist-25: A screening instrument for the psychiatric care of refugees. American Journal of Psychiatry, 144, 497-500.

Mollica, R., Poole, C., Son, L., & Murray, C. (1997). Effects of war trauma on Cambodian refugee adolescents' functional health and mental health status. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 1099-1107.

Nolan, T. M., Bond, L., Adler, R., Littlefield, L., Birleson, P., & Marriage, K. et al. (1996). Child Behavior Checklist classification of behavior disorder. Journal of Paediatric Child Health, 32, 405-411.

Nunnally, J. C. (1978). Psychometric Theory (2nd ed.). New York: McGraw-Hill.

Papageorgiou, V., Frangou-Garunovic, A., Iordanidou, R., Yule, W., Smith, P., & Vostanis, P. (2000). War trauma and psychopathology in Bosnian refugee children. European Child & Adolescent Psychiatry, 9, 84-90.

Pynoos, R. S., & Nader, K. (1993). Issues in the treatment of posttraumatic stress in children and adolescents. In J. O. Wilson & B. Raphael (Eds.), International handbook of traumatic stress syndrome (pp. 535-549). New York: Plenum.

Raboteg-Saric, Z., Zuzul, M., & Kerestes, G. (1994). War and children's prosocial behavior. European Journal of Personality, 8, 201-212.

Rousseau, C., Drapeau, A., & Corin, E. (1998). Risk and protective factors in Central American and Southeast Asian refugee children. Journal of Refugee Studies, 11, 20-37.

Sack, W., Clarke, G., Him, C., Dickason, D., Goff, B., & Lanham, K. et al. (1993). A 6-year follow-up study of Cambodian refugee adolescents traumatized as children. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 431-437.

Smith, P., Perrin, S., Yule, W., Hacam, B., & Stuvland, R. (2002). War exposure among children from Bosnia Hercegovina: Psychological adjustment in a community sample. Journal of Traumatic Stress, 15, 147-156.

Smith, P., Perrin, S., Dyregrov, A., & Yule W. (2003). Principal components analysis of the impact of event scale with children in war. Personality and Individual Differences, 34, 315-322.

Southam-Gerow, M. A., & Kendall, P. C. (2002). Emotion regulation and understanding: Implications for child psychopthology and therapy. Clinical Psychology Review, 22, 189-222.

Sourander, A. (1998). Behavior problems and traumatic events of Unaccompanied Refugee Minors. Child Abuse & Neglect, 22, 719-727.

Ten Berge, J. M. F. (1986). Rotation to prefect congruence and the cross-validation of component weights across populations. Multivariate Behavioral Research, 21, 41-64.

Tiet, Q. Q., Bird, H., Davies, M., Hoven, C., Cohen, P., & Jensen, P. S. et al. (1998). Adverse life events and resilience. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 1191-1200.

Verhulst, F. C., & Van der Ende, J. (1997). Factors associated with child mental health service use in the community. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 901-909.

Vinokurov, A., Trickett, E. J., & Birman, D. (2002). Acculturative hassles and immigrant adolescents: A life domain assessment for Soviet Jewish refugees. Journal of Social Psychology, 142, 425-445.

Warshaw, M. G., Fierman, E., Pratt, L., Hunt, M., Yonkers, K. A., & Massion, A. O. (1993). Quality of life and dissociation in anxiety disorder patients with histories of trauma or PTSD. American Journal of Psychiatry, 150, 1512-1516.

Winokur, A., Winokur, D. F., Rickles, K., & Cox, D. (1984). Symptoms of emotional stress in family planning service: Stability over a four-week period. British Journal of Psychiatry, 144, 395-399.

Wu, P., Hoven, C. H., Bird, H. R., Moore, R. E., Cohen, P., & Alegria, M. et al. (1999). Depressive and disruptive disorders and mental health service utilization in children and adolescents. Journal of the American Academy of Child & Adolescents Psychiatry, 38, 1081-1092.

The Dutch studies were supported by Achmea Victim and Society Foundation and the Health Research Development Counsel (ZON-Mw). The Belgian studies were supported by a PhD bursary of the Special Research Fund of Ghent University.

Ilse Derluyn, Department of Orthopedagogics, Ghent University, Ghent, Belgium.

Elisabeth Eurelings-Bontekoe, Department of Health and Clinical Psychology, Leiden University, Leiden, The Netherlands.

Eric Broekaert, Department of Orthopedagogics, Ghent University, Ghent, Belgium.

Philip Spinhoven, Department of Health and Clinical Psychology, Department of Psychiatry, Leiden University Medical Center, Leiden, The Netherlands.

Reprint requests should be sent to Tammy Bean, Postdoctoral Fellow, Clinical Child and Adolescent Psychology, University of Pittsburgh Medical Center, Western Psychiatric Institute and Clinic, Room 546 Bellefield Towers, 3811 O'Hara Street, Pittsburgh, PA 15221. E-mail:
Table 1
Summary of Sample Characteristics
 Gender **

 N Boys Girls

Total sample adolescents 3890 59.3 40.7
Dutch URM 920 72.8 27.2
Belgian immigrant/refugee (I/R) 1294 53.9 46.1
Dutch 1059 56.8 43.2
Belgian 617 54.6 45.4

 Age in years

 Mean SD Range

Total sample adolescents 15.72 1.74 8-26
Dutch URM 15.68 1.49 8-20
Belgian immigrant/refugee (I/R) 15.41 1.88 10-26
Dutch 15.72 1.54 13-21
Belgian 1,646 1.92 13-21

 Group **

 Natives I/R URM

Total sample adolescents 40.7 30.7 29
Dutch URM 0 0 100
Belgian immigrant/refugee (I/R) 1.0 89.1 9.9
Dutch 90.1 9.9 0.0
Belgian 97.9 2.1 0.0

 Type of caregiver **

 Parental Other

Total sample adolescents 70 30
Dutch URM 0 100
Belgian immigrant/refugee (I/R) 84.3 15.7
Dutch 97.3 2.7
Belgian 97.6 2.4

Note. ** percentages

Table 2
Summary of Confirmatory Factor Analyses and Reliability Analyses
per language version

 Two factor

Language n EV LV [alpha]

Total sample 3019 33.1% .4% .90
Dutch 1640 31.4% .4% .88
Portuguese 326 29.5% .6% .90
English 215 32.6% .8% .91
French 163 33.0% .5% .91
Arabic 127 34.9% 1.3% .92
Turkish 118 NA NA .92
Russian 111 43.6% 0.7% .95
Chinese 95 47% 2.2% .92
Spanish 47 NA NA .84
Farsi NA NA NA
Albanese NA NA NA
Croatian NA NA NA
Dari 23 NA NA .91
Amharic 18 NA NA .91
Somali NA NA NA
German NA NA NA
Mongolian 11 NA NA .86

 Total scale

Language [] [] [r.sub.ab*]

Total sample .18 .05-.59 .63
Dutch .17 .13-.55
Portuguese .18 .01-.62
English .20 .17-.63
French .21 .05-.63
Arabic .22 .04-.70
Turkish .24 .00-.70
Russian .29 .01-.80
Chinese .24 .13-.69
Spanish .13 .04-.62
Farsi NA NA
Albanese NA NA
Croatian NA NA
Dari .21 .10-.74
Amharic .20 .04-.70
Somali NA NA
German NA NA
Mongolian .14 .34-.90


Language n [alpha] []

Total sample 3126 .92 .30
Dutch 1670 .90 .27
Portuguese 342 .91 .28
English 230 .91 .29
French 166 .91 .30
Arabic 141 .92 .31
Turkish 123 .92 .33
Russian 119 .95 .46
Chinese 96 .93 .37
Spanish 47 .78 .12
Farsi 44 .89 .25
Albanese 29 .88 .23
Croatian 22 .93 .33
Dari 24 .90 .28
Amharic 21 .92 .32
Somali 16 .94 .37
German 17 .91 .30
Mongolian 11 .86 .17


Language [] [r.sub.ab*]

Total sample .36-.66 0.64
Dutch .30-.63
Portuguese .22-.64
English .19-.68
French .34-.68
Arabic .23-.75
Turkish .33-.70
Russian .41-.79
Chinese .43-.71
Spanish .05-.51
Farsi .17-.67
Albanese .07-.75
Croatian -.14-.77
Dari .11-.68
Amharic .37-.70
Somali -.03-.85
German .09-.77
Mongolian .34-.89


Language n [alpha] []

Total sample 3524 .75 .20
Dutch 1771 .75 .22
Portuguese 374 .62 .12
English 298 .72 .18
French 220 .71 .20
Arabic 184 .67 .16
Turkish 151 .66 .16
Russian 137 .58 .12
Chinese 106 .74 24
Spanish 53 .76 25
Farsi NA NA NA
Albanese NA NA NA
Croatian NA NA NA
Dari 30 .67 .17
Amharic 26 .61 .19
Somali NA NA NA
German NA NA NA
Mongolian 12 .60 .18


Language [] [r.sub.ab*]

Total sample .13-.45 .53
Dutch .15-.50
Portuguese .10-.42
English .14-.51
French .24-.57
Arabic .16-.50
Turkish .15-.50
Russian .11-.42
Chinese .24-.59
Spanish .21-.69
Farsi NA
Albanese NA
Croatian NA

Dari .07-.61
Amharic .02-.73
Somali NA
German NA
Mongolian -.28-.76

Note, EV = Explained Variance with PCA. LV = Loss of Explained
Variance with MGM. [alpha] = Alpha coefficient.
[r.sub.[??]] = Mean inter-item correlation. [] = Range
item-total correlations. [r.sub.ab*] = Test- re-test reliability
calculated for 12 month interval for URM sample only, n x519. NA =
not shin to analyze, more than one item with zero variance.

Table 3
Measure intercorrelations

 Total RATS Total SLE Total SDQ

URM (n) r (n) r (n) r

Total HSCL-37A (771) .74 (819) .39
Internalizing (761) .79 (812) .41
Externalizing (780) .32 (835) .12
Dutch natives
Total HSCL-37A (1058) .75 (1057) .48
Internalizing (1058) .76 (1057) .36
Externalizing (1058) .23 (1057) .39

Total HSCL-37A (870) .66 (1167) .38 (1117) .65
Internalizing (854) .68 (1149) .38 (1101) .64
Externalizing (886) .33 (1192) .22 (1141) .43
Belgian natives
Total HSCL-37A (596) .67 (615) .38 (612) .70
Internalizing (596) .67 (614) .30 (611) .64
Externalizing (597) .31 (616) .34 (613). .42

Note. All correlations are significant at the .001 level and

Table 4
External criteria influencing HSCL-37A Internalizing and Externalizing


 n M SD t(df) p d

URM: Need for MHC *
Need for MHC 438 54.33 12.75 8.84 (592) <.001 .83
No need for MHC 156 44.03 11.79

Guardian: Need for MHC *
Need for MHC 95 57.47 12.07 6.03 (492) <.001 .69
No need for MHC 399 49.05 12.28

Teacher: Need for MHC *
Need for MHC 115 54.63 14.22 4.06 (401) <.001 .45
No need for MHC 288 48.62 13.09

URM: MHC Utilization *
Utilization of MHC 96 54.79 14.14 2.68 (682) <.01 .30
No utilization of MHC 588 50.98 12.70

Referral: MHC by guardian *
Referred to MHC 59 59.28 13.54 5.73 (496) <.001 .80
Not referred MHC 439 49.53 12.09


 n M SD t(df) p d

URM: Need for MHC *
Need for MHC 448 15.63 3.22 1.92 (605) ns .18
No need for MHC 159 15.06 3.09

Guardian: Need for MHC *
Need for MHC 97 16.52 3.69 3.38 (506) <.001 .38
No need for MHC 411 15.28 3.13

Teacher: Need for MHC *
Need for MHC 120 15.96 2.93 3.17 (413) <.01 .34
No need for MHC 295 14.94 2.99

URM: MHC Utilization *
Utilization of MHC 95 15.97 3.09 1.40 (697) ns .16
No utilization of MHC 604 15.47 3.22

Referral: MHC by guardian *
Referred to MHC 59 16.56 2.97 2.79 (511) <.01 .39
Not referred MHC 454 15.36 314

Note, MHC = Mental Health Care. * = Only available for the Dutch URM
COPYRIGHT 2007 Libra Publishers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2007 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Bean, Tammy; Derluyn, Ilse; Eurelings-Bontekoe, Elisabeth; Broekaert, Eric; Spinhoven, Philip
Geographic Code:4EUBL
Date:Mar 22, 2007
Previous Article:Rural adolescent aggression and parental emotional support.
Next Article:Nexus Aliquis: in pursuit of efficacy, resilience, and full potential.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters