Discrimination, mental problems and social adaptation in young refugees

Abstract

Background: Mental problems have been hypothesized to impede social adaptation and vice versa, and discrimination is assumed to interact with both. The available empirical documentation is, however, limited. The objective of this study is to contribute to a more comprehensive understanding of associations and pathways between discrimination, mental problems and social adaptation in young refugees. Methods: Structural Equation Modelling (SEM) was used for the analysis of cross-sectional data from interviews with 131 young Middle Eastern refugees residing in Denmark. Results: Participants reported a mean of 1.8 experiences of discrimination, and the prevalence of five indicators of positive social adaptation was 47–92%. Discrimination, mental problems and social adaptation were strongly mutually associated, without gender difference. Discrimination predicted internalizing behaviour. Improved social adaptation correlated negatively with discrimination and with externalizing and internalizing behaviour. Conclusion: Perceived discrimination among young refugees from the Middle East is associated with mental problems and social adaptation. Discrimination seems to provoke internalizing but not externalizing behaviour. The direction of other pathways is ambiguous, suggesting a certain amount of recursive interaction between mental health, discrimination and social adaptation.

Many^1,,2 but not all^3,,4 empirical studies have documented mental problems to be prevalent in refugee adolescents as compared to second generation immigrants or native adolescents. Most studies have documented a decreasing prevalence of symptoms over time in the immigrant country, but after many years, the prevalence may still be high.^5–7 Causal hypotheses have focused on experiencing and perceiving discrimination and the interactive role of this phenomenon in the social adaptation process demanded from refugees and immigrants.

Mental problems have been hypothesized to impede social adaptation and vice versa, and discrimination is assumed to interact with both.⁸ Discrimination involves cognitive appraisal of threat and is therefore expected to affect mental health similar to other stressors and to lead to increased identification with one's own group in contrast to the majority group thus impeding negatively on social adaptation.

Being subject to the experience of discrimination has been identified as a most important resettlement stressor, which may affect mental health. Thus, discrimination predicted anxiety and depression⁹ and psychological distress;¹⁰ it was associated with depression^11,,12 and posttraumatic stress disorder (PTSD),¹³ and led to increased ethnic identification impeding negatively on well-being¹⁴ in studies among various groups of adult refugees. In a study¹⁵ involving adolescent immigrants in Norway from 11 different ethnic origins, perceived discrimination correlated with both externalizing and internalizing behaviour, and discrimination was identified as the strongest predictor of mental problems among immigrant adolescents with Turkish background living in Norway and Sweden.¹⁶ Symptoms of depression and anxiety predicted perceived discrimination in a study of minority and immigrant adolescents of Armenian, Mexican and Vietnamese background in the USA.¹⁷ In a study of Vietnamese refugee adolescents in Finland,¹⁸ perceived discrimination had detrimental impact on school adjustment, whereas perceived parental support had a positive impact.

Mental problems, as indicated by a psychiatric diagnosis, have been found to predict poor social adaptation, as exemplified by short education and risk of being on social welfare,^19,,20 and mental symptoms were found to be associated with language proficiency, but not with family or peer relations or other indicators of adaptation, in a study of young Soviet Jewish refugees in the USA.²¹ Recent life stress and lack of proficiency in English predicted depression (but not PTSD) in a sample of young Cambodian refugees in Canada.²² Conversely, several studies have found mental problems to be un-associated with indicators of social adaptation, such as social competence and academic performance,^4,,6,23,24 work and studying, leisure activity, family relationships and economic independence,²⁵ functional impairment²⁶ and proficiency in English.²²

In spite of the public health importance of and the theoretical foundation for links and interactions between exposure to discrimination, mental problems and social adaptation, the amount of available empirical documentation is limited, and few studies have focused on young refugees. Focusing on young refugees from the Middle East, the aim of the present research was to identify associations between mental problems, discrimination and social adaptation among young refugees resettled in high-income countries. The main components in our hypothesis model were (figure 1A):

• Experiencing and perceiving discrimination causes:

  • Behaviour problems of internalizing and, particularly, of externalizing nature

  • Weakening of social adaptation

• Internalizing and externalizing behaviour problems cause weakening of social adaptation

• Weakening of social adaptation increases the risk of:

  • Experiencing and perceiving discrimination

  • Internalizing and externalizing behaviour problems.

Methods

Participants

The study group comprises 131 young refugees from the Middle East [76 girls and 55 boys; mean age 15.3 years (11–23)] from 67 families. They were assessed in 2000–01 as part of a follow-up study of 311 children, who from 1 February 1992 till 30 April 1993 were consecutively registered in Denmark as asylum seekers from Middle Eastern countries accompanied by at least one parent.²⁷ The response rate in the original study was 90.4%. In 2000, 190 youths from a total of 90 families had been granted permission to stay. Of these one had died, six had emigrated and one was unknown to the Danish registry, leaving 182 for study. The parents of 37 youths declined to participate, making the contact to these youths impossible, and another nine youths refused to participate even if the parents had accepted. Thus, 136 youths accepted to participate, but in five cases data turned out to be too incomplete for analysis. Finally, 72% participated out of the 182 youths left for study. Non-participants did not differ significantly from participants as concerns nationality, ethnicity or parents’ religion. However, more girls participated (79.2% girls vs 64.0% boys, P < 0.05), and younger age predicted participation [odds ratio(OR) 0.90 per year, P < 0.04].

Data collection

Parents and youths were interviewed separately by the use of structured questionnaires basically focusing on family structure; parents’ and youths’ social situation and health; the youths’ exposure to violence, while in Denmark and his/her possible loss of and separation from parents. Interviews were conducted by a Danish visiting nurse. All youths were interviewed in Danish, their parents with the assistance of a professional interpreter whenever necessary.

Indicators of social adaptation and discrimination

Social adaptation in the youths was indicated by the amount of age-relevant participation in social life in Denmark. Measurable indicators were developed by the use of an ethnographic approach,²⁸ including a focus group interview with five 18–21-year-old male refugees from the Middle East, and qualitative interviews with two 17-year-old female refugees. None of these youths were members of the study population itself. Social adaptation was decomposed into two qualitative different parts: actual participation (behaviour) and prerequisites for participation, components that were identified in previous research among minorities in Denmark.^28,,29 Four specific measurable indicators were developed within each area: participation—‘attending school or work’, ‘participation in organized leisure activities’, ‘has been with friends the last month’ and ‘criminal behaviour’, all dichotomous variables; prerequisites for participation—‘spoken and written Danish proficiency’ measured on a 0–3 scale; ‘number of Danish friends’ and ‘number of schools attended’. The variable ‘number of schools attended’ originates in the experience of young refugees of being forced to change school several times due to the Danish refugee placement policy and in this way experiencing broken networks and change in academic subjects and learning systems. The parents’ (the father or, in case the father was not in Denmark, the mother) social adaptation was indicated by three items identified as the most basic prerequisites for being self-supporting and thus adapted to life in Denmark:³⁰ ‘employment’ (having any job); ‘education’ (having an education that is recognized in Denmark); and ‘Danish language skills’ (self-evaluated on a 0–3 scale).

Also, indicators of discrimination were identified based on the focus group interview and the individual qualitative interviews. Discrimination was indicated by five specific experiences of being rejected or harassed due to being a foreigner, any time since arrival in Denmark (table 1).

Mental health scales

The mental health of the youths was assessed by use of part of the Achenbach System,^31,,32 which has been widely used in transcultural research.³³ In accordance with the recommendation in the Danish manual and standardization,^34,,35 young refugees up to the age of 16 were assessed using the Youths Self Report (YSR), while young refugees aged 17 and above were assessed by use of the Young Adult Self Report (YASR). The official Danish translation of the YSR was applied, and the YASR was translated into Danish for this investigation in co-operation with the Danish distribution centre following the recommended procedure with two independent translators.

Referring to the past 6 months, the scales consisted of closed questions rated 0–2 and focusing on a wide range of behaviours and emotions. For each scale, an externalizing and an internalizing score was computed based on general as well as age-specific behaviour.^31,,32,36 Only items included in both of the two original scales (18 and 17 items for the externalizing and internalizing scales, respectively) were applied. The item ‘steals’ was excluded from the externalizing scale in order to avoid redundancy in the analysis. Cronbach's Alpha for the externalizing scale with the shared items was 0.84 and for the internalizing scale 0.87.

Statistical analysis

The statistical analysis applied χ²-test for 2 × 2 tables (without Yates’ correction), Fisher's exact test, Cronbach's Alpha and linear regression.³⁷ Structural equation models (SEM) specifying the relationship between variables were estimated using the statistical software packages Mplus, version 3.0, and using polychoric correlation coefficients to measure the degree of association between categorical items.³⁸ SEM is a comprehensive statistical approach to analysis of relations among observed and latent (unobservable) variables and represents a combination of path analysis with confirmatory factor analytic procedures.³⁹

Model fit was estimated using χ²-statistic. Supplementary indicators of model fit used were comparative fit index (CFI) and the root mean square error of approximation (RMSEA), the latter with values less than 0.05 as indicative of good model fit.³⁸ The indicators of the latent variables (social adaptation and discrimination) were entered into the analysis one by one and fitted independently, where after they were adjusted for covariates before being combined in a comprehensive model. The externalizing and internalizing scales were entered into the analysis as such. Models were adjusted for age, gender, nationality, ethnicity, religion and social adaptation of parents.

The study was approved by the Regional Committee for Ethics in Medical Science and the Danish Data Protection Agency.

Results

Background

Ninety-seven (74%) of the youths were of Iraqi nationality, 17 (13%) were Iranian and 12 (9.2%) Stateless Palestinian, while three were from Syria and two from Lebanon. Fifty-four (41.2%) were of Kurdish and 14 (10.7%) of Palestinian ethnicity. One hundred and two (77.9%) of the youths were Muslims, 14 (10.7%) were Christians, 12 (9.2%) belonged to other denominations and three did not have any religious affiliation. Both parents lived in Denmark in the case of 113 (86.3%) youths, only the mother in the case of 13 youths (9.9%) and only the father in the case of four youths (3.1%), while one youth had no parents in Denmark. Most youths (109, 83%) were still living at home with at least one of their parents; seven were married and one had divorced. Eighty-one (62%) youths still attended primary school, 17 (13%) attended high school and 15 (11%) attended further education.

Discrimination

Participants had a mean of 1.8 experiences of discrimination (21% zero experiences, 4% five experiences). There were no differences between nationalities, ethnicities or religious affiliations concerning specific experiences. The strongest predictor for any indicator of discrimination in the bivariate analysis was male gender, which predicted being refused at a discotheque (OR 20.9, P < 0.005) (table 1). Higher age was also related to being refused at discotheques, as well as by friends or ignored by a teacher. More youths from families with an unemployed father (or mother) experienced being ignored by a teacher or being teased, and more youths from families, where the father (or mother) did not have an education recognized in Denmark, experienced derogatory remarks.

Five indicators of discrimination were fitted in a confirmatory factor analysis resulting in a good fit with the introduction of one correlation between ‘derogatory remarks’ and ‘teased’ (χ² = 1.168, P = 0.76; RMSEA 0.000).

Social adaptation

The prevalence of the five positive indicators of social adaptation was 47–92%, while 15% had committed illegal acts. The participants had a mean of 1.5 Danish friends and had attended a mean of 3.6 schools during their stay in Denmark.

Attending school or work (OR 0.8, P < 0.05), having Danish writing skills (OR 0.9, P < 0.05) and participating in organized leisure activities (OR 0.8, P < 0.005) all depended on age in the bivariate analysis. Fewer youths of Iraqi nationality (OR 0.4, P < 0.05) and Muslim religious denomination (OR 0.4, P < 0.05) had committed illegal acts, youths of Iranian nationality had significantly more Danish friends than others (B 1.1, P < 0.05), and youths of Kurdish ethnicity had significantly poorer Danish reading skills (OR 0.2, P < 0.005). Indicators of parents’ social situation were not associated with the children's social adaptation.

The eight indicators of social adaptation were fitted in a confirmatory factor analysis, first on its own and then including covariates. Omitting one indicator (‘been with friends the last months’) and introducing terms of interaction (reading ‘with’ writing skills; organized leisure activities ‘with’ commitment of illegal acts) resulted in an acceptable fit (χ² = 9.4, P = 0.307; RMSEA = 0.037). Among covariates, only age was significantly associated with social adaptation, so that age independently both predicted the latent variable and the indicator ‘number of schools attended’, resulting in an improved model fit (χ² = 12.53, P = 0.325; RMSEA = 0.033).

Mental health

The mean externalizing and internalizing behaviour scores in the combined self-report scale were 8.5 (out of 18) and 9.9 (out of 17), respectively. In bivariate linear regressions only age predicted internalizing behaviour significantly (B 0.4, P < 0.05). Nationality, ethnicity, religion and parents’ social situation were not associated with mental problems in the youths.

A comprehensive model

Indicators of discrimination, social adaptation and the externalizing and internalizing behaviour scales were fitted in a SEM to test the existence of pathways and their directions. As indicated by partial correlation coefficients, discrimination, social adaptation and internalizing behaviour were strongly associated, without any gender difference (table 2, figure 1B). Discrimination predicted internalizing but not externalizing behaviour, and was correlated with social adaptation. Externalizing behaviour correlated with internalizing behaviour, and both correlated with social adaptation. Social adaptation correlated with discrimination as well as with internalizing and externalizing behaviour. Parental social adaptation was not associated with any of the other variables. Male gender, Iraqi nationality and Kurdish ethnicity predicted externalizing behaviour as age predicted discrimination and social adaptation. Model diagnostics indicated an acceptable fit (χ² = 57.78, DF = 46, P = 0.11, CFI = 0.96, RMSEA = 0.044).

Discussion

Significant associations were found between the latent variables ‘discrimination’ and ‘social adaptation’ and between social adaptation and externalizing and internalizing behaviour. Discrimination was a predictor for internalizing but not for externalizing behaviour as hypothesized, and externalizing behaviour thus does not seem to have been provoked by discrimination. Social adaptation seems to be related to discrimination as well as to the two aspects of mental health, while discrimination is related to social adaptation directly or through its effect on internalizing behaviour.

In the present study, assessment of discrimination was based on the young participants’ own perception of other people's behaviour. It is possible that socially well-adjusted people do not take the same act towards them as discriminatory as do people, who are less well adjusted. On the other hand, experiencing acts of discrimination may also affect people's perception of the host-country at large and result in the victim turning away from the persons whom they identify as the assaulting group (the host-population), and strengthen their identification and association with their own ethnic group,⁸ thus affecting social adaptation negatively. This in turn may enhance the perception of being discriminated against and may thus result in a vicious circle. This is in accordance with the results of a qualitative study of Ethiopian asylum-seekers in the UK,⁴⁰ which concluded that problems associated with migration as well as with adaptation to the English culture and social systems and negative settlement experiences including discrimination all had impact on the health of the refugees. In biographical interviews with refugees in 15 member states of the EU, one of the most fundamental barriers to integration was identified as racism at both the personal and the institutional levels.⁴¹ In a review of integration research from the Nordic countries, discrimination and prejudice were identified as most important barriers to the integration of ethnic minorities.⁴² Based on bureaucratic procedures, inflexibility and enforced dependency, the welfare systems of some European countries may support a development of the apprehension of the refugees as less able, more dependent and belonging to the socially excluded, which all in all may have a negative impact on social adaptation.

Victims of discrimination can react in different ways, e.g. by withdrawal (internalizing) or by aggression towards externally blamed agents (externalizing).⁸ The findings of the present study indicate that internalizing reactions are associated with discrimination, while externalizing behaviour seems to follow other pathways, although both forms of reaction are negatively associated with social adaptation. This appears to be in accordance with studies which have found discrimination to be associated with anxiety, depression and other indicators of internalizing reactions.^9–11 One hypothesis is that young refugees are reluctant to show aggression, since that could confirm the negative representation of them in societies with increasing levels of xenophobia. While internalizing behaviour did not seem to be gender specific, externalizing behaviour was associated with male gender. In a study of adolescent Cambodian refugees in Canada,⁶ boys tended to manifest fewer internalizing symptoms in early adolescence, while male gender was a risk factor for internalizing behaviour in mid-adolescence.

Our study group originated from the Middle East, and cultural differences may contribute to explain disagreement with findings in previously published research. Thus, studies demonstrating lack of association between mental problems and social adaptation have been carried out among refugees from the South East (Cambodians and Vietnamese).^4,,5,25 Different cultural perceptions of mental problems may influence the extent to which problems are reported and thus blur associations.⁴

Limitations

Non-participation amounted to 28%, and the age and gender differences between participants and non-participants may have influenced the findings e.g. the lack of association between externalizing behaviour and discrimination. The results apply to the specific context of Middle Eastern refugee youths in exile in a high-income country and cannot, without further investigations, be generalized to other refugee groups.

SEM appears to be a relevant methodology when analysing concurrent associations between several different latent variables, but results should be interpreted with caution. Even if the model fitted the data well, some of the associations were rather weak. Parental social adaptation did not seem to be related to neither social adaptation nor mental health of the youths, but this might be related to the rather narrow indicators used for measuring this variable, and further studies are necessary in order to draw conclusions. Furthermore, the sample size in this study was relatively small. Small samples tend to favour simpler models over more complex models with more parameters due to loss of precision when more estimates are included in the model.⁴³ Moreover, the sample included siblings resulting in decreased independence of some of the observations. Finally, causality can only to a limited extent be concluded based on cross-sectional data, and further studies are necessary to understand the directional influences between latent variables. The results of the present study, however, provide a basis for formulating more specific hypotheses for further study.

Conclusion

Perceived discrimination among young refugees from the Middle East is associated with mental problems and weakening of social adaptation. Discrimination seems to provoke internalizing but not externalizing behaviour. The direction of other pathways is ambiguous, suggesting a certain amount of recursive interaction between discrimination, mental problems and social adaptation.

Acknowledgements

The study was supported by the BG-Fund, the Egmont Foundation, the Henrik Henriksens Fund and the Research Fund of The Danish Medical Association/The Hojmosegaard Grant. Structured interviews were collected by visiting nurse Berit Haahr Rindorf.

Conflicts of interest: None declared.

References