The Current Study
The goal of this study is to test whether there is an immigrant paradox among adolescents in junior vocational education. Furthermore, we will test whether family obligations and school adjustment can predict adaptation in an immigrant adolescent and a national adolescent sample. Earlier, Steinberg (
1996) used school adjustment to explain positive adaptation outcomes, but did not include family obligations in his model. Fuligni used both family obligations and school adjustment to explain positive adaptation outcomes among immigrant adolescents, but only provided empirical evidence for the positive effects of family obligations on grades (Fuligni
1997,
1998; Fuligni et al.
1999). In this study the variables family obligations and school adjustment are used to predict a more comprehensive range of adaptation outcomes, namely school grades, self-esteem, psychological and behavioral problems. These analyses shall be conducted in a national and an immigrant adolescent sample.
The conclusion that an immigrant paradox exists is warranted if immigrant adolescents have a more positive pattern of adaptation than national adolescents, despite a lower socio-economic status. The first hypothesis is that the sample of immigrant adolescents in this study will have a lower socio-economic status than the national adolescents. In previous studies in the Netherlands it has already been reported that immigrant adolescents usually come from households with a lower socio-economic status than Dutch national adolescents (Central Bureau of Statistics
2008; Herweijer
2009). The second hypothesis is that immigrants perform better in terms of adaptation than nationals. We draw this hypothesis from immigrant paradox findings (e.g., Harker
2001). If the adaptation of immigrant adolescents in the Netherlands follows an immigrant paradox pattern these first two hypotheses should be supported.
The first two hypotheses focused on the possible existence of an immigrant paradox among immigrant adolescents in junior vocational education, whereas the third and fourth hypotheses will examine to what extent the adaptation of immigrant adolescents is related to a sense of family obligations and school adjustment. The third and fourth hypotheses are that immigrants score higher on family obligations and school adjustment than nationals (Fuligni
1997; Fuligni et al.
1999). Our final hypothesis is that in the immigrant adolescent sample school adjustment is a partial mediator in the relationships between family obligations and other aspects of adaptation. Partial mediation is expected as several scholars argue that family obligations have unique direct effects on adaptation and school adjustment (Fuligni
1998; Suarez-Orozco and Suarez-Orozco
1995; Zhou and Bankston
1998). The role of family obligations in national adolescents’ adaptation has not received much attention in earlier studies (however, see Berry et al.
2006). In this study we will test the relationship between family obligations and adaptation in both a sample of immigrant adolescents and a sample of national adolescents. Given the scarcity of earlier studies no specific hypotheses are specified with respect to the national sample. By testing these hypotheses, we may gain insight in whether a strong sense of family obligations and a positive school adjustment contribute to the explanation of immigrant paradox findings.
As adaptation variables we choose self-esteem, psychological problems, behavioral problems and grade point average. These variables have been included in a study concerning the immigrant paradox in Europe (Sam et al.
2008) and a study concerning family obligations (Fuligni et al.
1999). The measures previously used by Sam et al. were found to be structurally equivalent across cultures, which is important when making cross-cultural comparisons. By using similar variables, this study may serve as a further validation of previous results.
Method
Participants
A total of 277 Dutch national students and 175 non western immigrant students completed the questionnaires. In addition, nineteen western immigrants participated in the data collection. Since the number of western immigrants was too small to be meaningfully included in the analyses they were removed from the dataset. The students were spread out across eight schools and 27 classrooms. The mean age of the national students was 14.52 (SD = .97), and the mean age of the immigrant students was 14.64 (SD = 1.20). The national sample consisted of 150 boys and 126 girls, one respondent did not answer the question concerning gender. The immigrant sample consisted of 72 boys and 103 girls. Forty one had a Turkish cultural background (9.1%), 39 a Moroccan (8.6%), 16 a Surinamese (3.5%), 16 an Antillean (3.5%), and 63 immigrant adolescents (13.9%) had a different cultural background. They originated from Afghanistan, Algeria, Argentina, the Dominican Republic, the Philippines, Sri- Lanka, Somalia, Senegal, Columbia, Lebanon, Russia, Taiwan, Vietnam, Egypt, Cap-Verde, Haiti and the People’s Republic of China.
Instruments
A survey consisting of several scales was administered to the students. The survey began with questions about demographics, namely age, gender, the birthplace of both parents of the respondent, the birthplace of the respondent and the respondents’ religion.
Socio-economic Status
Socio-economic status was measured with the family affluence scale (Currie et al.
1997). A sample item of this scale is “How many computers does your family own”. Since the scale has different response categories for the separate items Cronbach’s alpha could not be computed. Higher scores are indicative of a higher socioeconomic status. The family affluence scale has been found to be a valid indicator of adolescents’ socio-economic status (Boyce et al.
2006).
Grade Point Average
To get an indication of students’ grade point average we asked students to report their last report grades for Dutch language, English language and math. The mean of these three grades was used in the analyses. The grade point average can vary between one through ten, in which one is the lowest and ten is the highest possible score.
Family Obligations
The scale used to measure family obligations was developed by Georgas et al. (
1996). It consists of fourteen items which are answered on a five point scale ranging from “completely disagree” up to “completely agree”. Sample items are “Children should obey their parents” and “A child has to take care of its parents when they need help”. A higher total score indicates a stronger sense of family obligations. The Cronbachs’ alphas were .73 for the immigrant sample and .66 for the national sample.
School Adjustment
The school adjustment questionnaire was adapted from the ICSEY-study (Berry et al.
2006). It consists of six items which are answered on a five point scale ranging from “completely disagree” up to “completely agree”. Sample items are “I would like to quit school completely” and “I dislike going to school in the morning”. Five items were reverse coded so that a higher total score is indicative of a better school adjustment. In the ICSEY study this scale was found to have a unifactorial factor structure and to be structurally equivalent across different ethnic groups. The Cronbachs’ alphas were .63 for the immigrant sample and .70 for the national sample.
Self-esteem
To measure self-esteem we used the Rosenberg’s (
1965) self-esteem inventory. It consists of ten items which are answered on a five point scale ranging from “completely disagree” up to “completely agree”. A sample item is “On the whole I am satisfied with myself”. A higher score indicated higher self-esteem. The Cronbachs’ alphas were .83 for the immigrant sample and .89 for the national sample.
Psychological Problems
The psychological problems scale was taken from the ICSEY-study (Berry et al.
2006) and consisted of 15 items answered on a five-point scale ranging from “never” up to “very often”. A sample item is “I feel restless”. A higher score was indicative of more psychological problems. In the ICSEY study this scale was found to have a unifactorial factor structure and to be structurally equivalent across different ethnic groups. The Cronbachs’ alphas were .92 for the immigrant sample and .90 for the national sample.
Behavioral Problems
The behavioral problems questionnaire was an adaptation of Olweus’ antisocial behavior scale (Olweus
1989,
1994). The original scale has been shown to have a good construct validity and reliability (Bendixen and Olweus
1999). The Cronbachs’ alphas were .85 for the immigrant sample and .80 for the national sample. The scale consisted of ten items. The items were scored on a five point scale ranging from “never” up to “more than 3 times during the past 12 months”. A sample item of this questionnaire is: “had a serious fight with a teacher”.
Procedure
Forty-eight junior vocational schools in the Netherlands with an ethnically diverse student body were invited to participate in a survey about multiculturalism and adaptation. Schools were first contacted via telephone. Only schools in the urbanized western part of the Netherlands were contacted, as these schools have ethnically diverse student populations. We followed a convenience sampling approach since there is no register available of schools and tracks and the ethnic composition of their student body. When schools showed an interest in the study, an appointment was made to explain the study in more detail. This led to seven schools participating in the study. Prior to the data collection, the teachers were informed about the goal of the study and letters of informed consent were sent to the students’ parents. If parents did not want their children to participate in the study, they could contact the researcher. No parents made use of this possibility. Strict anonymity was promised for the schools, the teachers and the students. The questionnaires were administered in the classroom during school hours under the supervision of a teacher and a researcher. Students were informed that participation in the study was entirely voluntary. No students decided to withdraw from the study. All questionnaires were written in Dutch. The research assistants received no indication from the teachers or the students that the Dutch fluency of any of the immigrant students was insufficient to understand the questionnaire. This was not to be expected as following education in junior vocational education requires good command of the Dutch language. Prior to the admission of the questionnaires the teachers received instructions.
Results
To analyze whether nationals scored higher on the SES than immigrants, a t-test was computed. The t-test indicated that there was a significant difference between immigrants and nationals, with immigrants (M = 2.41, SD = .46) scoring significantly lower than nationals (M = 2.69, SD = .42), t(453) = (6.487), p < .05, with a medium effect size, Cohen’s D = .64. This indicates that immigrant adolescents score substantially lower than nationals on the SES scale, although they were all in the same schools and classes.
To test for an immigrant paradox a MANOVA was computed. The immigrants were compared to the nationals on the variables grade point average, self-esteem, behavioral problems and psychological problems. As the number of boys and girls differed substantially for both the immigrant and the national adolescent samples, gender was added as a variable in the design. The means and standard deviations are reported in Table
1. The multivariate test indicated that there were no significant differences between the national and immigrant adolescent samples [Wilks’ lambda
F(3, 438) = .1.297,
p > .05,
η
2 = .01]. Subsequent ANOVAs indicated that immigrant adolescents did not experience more behavioral problems [
F(1, 440) = .893,
p > .05,
η
2 = .00] or psychological problems [
F(1, 440) = 2.082,
p > .05,
η
2 = .01], and did not differ in terms of self-esteem [
F(1, 440) = 1.512,
p > .05,
η
2 = .00], when compared with the nationals. There was a multivariate effect for gender [Wilks’ lambda
F(3, 438) = .24.944,
p < .05,
η
2 = .15]. Subsequent ANOVAs indicated that girls experienced fewer behavioral problems [
F(1, 440) = 12.689,
p < .05,
η
2 = .03], more psychological problems [
F(1, 440) = 48.625,
p < .05,
η
2 = .01], and a lower self-esteem [
F(1, 440) = 21.055,
p < .05,
η
2 = .05], than boys did. There was a multivariate interaction effect between gender and nationality on adaptation [Wilks’ lambda [
F(3, 438) = 6.254,
p < .05,
η
2 = .04]. Subsequent ANOVAs indicated that there were no interaction effects for behavioral problems [
F(1, 440) = 2.568,
p > .05,
η
2 = .01] or psychological problems [
F(1, 440) = .010,
p > .05,
η
2 = .00], but that there was a significant interaction term for self-esteem [
F(1, 440) = 12.351,
p < .05,
η
2 = .03]. Immigrant boys scored lower than national boys in terms of self-esteem whereas immigrant girls scored higher than national girls in terms of self-esteem.
Table 1
Means and standard deviations (between brackets) for the variables school grades, self-esteem, psychological problems and behavioral problems
School grades | 7.08 (.97) | 6.91 (.97) | 6.88 (.84) | 6.81 (.91) |
Self-esteem | 3.85 (.71) | 3.79 (.85) | 4.05 (.76) | 3.43 (.83) |
Psychological problems | 1.97 (.67) | 2.44 (.79) | 2.05 (.60) | 2.52 (.66) |
Behavioral problems | 2.14 (1.01) | 1.72 (.70) | 1.90 (.70) | 1.80 (.73) |
As the immigrant adolescents scored lower on socio-economic status than the nationals and socio-economic status may have effects on adaptation, we reran the MANOVA described above and added socio-economic status as a covariate. There was a significant multivariate effect for nationality [Wilks’ lambda F(4, 394) = 2.727, p < .05, η
2 = .03] and a significant multivariate interaction effect between gender and nationality [Wilks’ lambda F(4, 394) = 2.727, p < .05, η
2 = .03]. A series of follow up ANCOVAs indicated that immigrants scored significantly higher grades than nationals [F(1, 397) = 4.343, p < .05, η
2 = .01]. A marginally significant effect indicated that immigrant adolescents experienced fewer behavioral problems than nationals [F(1, 397) = 1.673, p < .10, η
2 = .01]. An interaction effect indicated that immigrant adolescent girls had a higher self-esteem than national adolescent girls [F(1, 397) = 7.129, p < .05, η
2 = .03].
An ANOVA was computed to analyze whether immigrants scored higher on the family obligations scale than nationals. To account for the unbalanced number of boys and girls, gender was added in the design. The ANOVA yielded a significant difference between immigrants and nationals, with immigrants (
M = 3.06, SD = .60) embracing family obligations more strongly than nationals (
M = 2.57, SD = .50), [
F(1, 441) = 96.737,
p < .05,
η
2 = .18]. This indicates that immigrant adolescents embrace family obligations much more strongly than western adolescents. The means and standard deviations are reported in Table
1. There was no significant effect of gender [
F(1, 441) = .870,
p > .05,
η
2 = .00], and there was no significant interaction effect [
F(1, 441) = 1.160,
p > .05,
η
2 = .00].
To analyze whether immigrants scored higher on the school adjustment scale than nationals, an ANOVA was computed. To account for the unbalanced number of boys and girls, gender was added in the design. The ANOVA indicated that there was a significant difference between immigrants and nationals, with immigrants (
M = 3.61, SD = .80) scoring significantly higher than nationals (
M = 3.35, SD = .84), [
F(1, 446) = 7.350,
p < .05,
η
2 = .02]. This indicates that immigrant adolescents had a better school adjustment than the national adolescents. The means and standard deviations are reported in Table
1. There was no significant effect of gender [
F(1, 446) = .870,
p > .05,
η
2 = .00], and there was no significant interaction effect [
F(1, 446) = 1.160,
p > .05,
η
2 = .00].
To test the relationships between family obligations, school adjustment and adaptation, we ran mediation analyses in the national and immigrant samples. To test for mediation, we used the stepwise method described by Baron and Kenny (
1986). The correlations between the variables are reported in Table
2. Positive relationships between family obligations and school adjustment were found in both the national and the immigrant adolescent sample. In the immigrant sample partial mediation was found for the variables behavioral problems and self-esteem. These partial mediations indicated that family obligations and school adjustment both had unique contributions to a higher self-esteem and fewer behavioral problems. Furthermore it was found that family obligations were related to a higher grade point average, but this relationship was not mediated by school adjustment. The results of the regression analyses in the immigrant sample are summarized in Table
3.
Table 2
Pearson correlation coefficients for the variables in this study
Immigrant |
SES | | −.004 | .042 | .139* | −.069 | .132* |
Family obligations | −.107 | | .148* | .036 | −.140* | −.202** |
School adjustment | −.116 | .208** | | .253** | −.372** | −.453** |
GPA | .078 | .200* | .142 | | −.104 | −.189* |
Psychological problems | −.032 | −.026 | −.378** | −.096 | | .193** |
Behavioral problems | .191* | −.291** | −.374** | −.035 | .042 | |
Self-esteem | .053 | .227** | .230** | .141 | −.375** | .045 |
Table 3
Results of the regression analyses for the immigrant adolescent sample
Step 1 | Fam.obl. | GPA | 6.276(1,150) | .315 | .126 | .200* | .04* | |
Fam.obl. | Psych. | .111(1,168) | −.033 | .099 | −.026 | .00 | |
Fam.obl. | Beh. prob. | 15.346(1,166) | −.407 | .104 | −.291** | .09** | |
Fam.obl. | Self-est | 9.201(1,170) | .294 | .097 | .227 | .05** | |
Step 2 | Fam.obl. | School adj | 7.708(1,170) | .278 | .100 | .208** | .04** | |
Step 3 | Fam.obl. | GPA | 3.968(2,149) | .278 | .129 | .177* | .05* | .04 |
School adj | | | .123 | .096 | .105 | | .01 |
Fam.obl. | Beh. prob. | 19.100(2,165) | −.319 | .100 | −.227** | .19** | .09 |
School adj | | | −.346 | .075 | −.328** | | .10 |
Fam.obl. | Self-est | 7.567(2,169) | .246 | .098 | .189* | .08** | .05 |
School adj | | | .175 | .073 | .180* | | .03 |
In the national sample, partial mediation was found for the variable behavioral problems. This partial mediation indicated that family obligations and school adjustment both had unique contributions to fewer behavioral problems. A full mediation was found between family obligations, school adjustment and psychological problems. This means that the relationship between a stronger sense of family obligations and fewer psychological problems can be fully explained through school adjustment in the national sample. The results of the regression analyses in the national sample are summarized in Table
4.
Table 4
Results of the regression analyses for the national adolescent sample
Step 1 | Fam. | GPA | .323(1,251) | .064 | .112 | .036 | .00 | |
Fam. | Psych. | 5.425(1,272) | –.187 | .080 | −.140* | .02* | |
Fam. | Beh. prob. | 11.559(1,272) | −.300 | .088 | −.202** | .04** | |
Fam. | Self | .008(1,272) | .009 | .104 | .005 | .00 | |
Step 2 | Fam. | School | 6.113(1,272) | .254 | .103 | .148* | .02* | |
Step 3 | Fam. | Psych | 23.279(2,271) | −.115 | .076 | −.086 | .15** | .02 |
School | | | −.281 | .044 | −.360** | | .14 |
Fam. | Beh. prob. | 39.067(2,271) | −.205 | .080 | −.138* | .22** | .04 |
School | | | −.375 | .047 | −.433** | | .18 |
Discussion
The aim of this study was to test for an immigrant paradox among immigrant adolescents in the Netherlands and to analyze the relationships between family obligations, school adjustment and adaptation in an immigrant and national adolescent sample. Based on immigrant paradox findings (e.g., Hayes-Bautista
2004; Garcia Coll
2005) the immigrant adolescents were expected to score better on measures of adaptation than national adolescents. However, the results provided no clear support for an immigrant paradox. When we reran the analysis with socio-economic status as a covariate, it was found that immigrant adolescents had higher grades and there was a marginally significant effect for behavioral problems. Immigrant girls had a significantly higher self-esteem than national adolescent girls. This pattern of adaptation is not as spectacular as the immigrant paradox findings in American and Canadian studies (Beiser et al.
2002; Harker
2001), but it is similar to previous European results (Sam et al.
2008). A particular aspect of the immigrant paradox is that immigrants tend to lose their benefit in terms of adaptation when they become more assimilated in the national society (Sam et al.
2008), eventually resulting in an adaptation pattern that is the opposite of an immigrant paradox. This phenomenon is well documented in the US (Jasso et al.
2004; Rumbaut
1999; Suarez-Orozco and Suarez-Orozco
2002), but may not materialize in the Netherlands. In the Netherlands, there is a more extensive and accessible social welfare system than in the United States (Alesina et al.
2001; Gustafsson and Stafford
1994). The accessibility of health care and welfare in the Netherlands may ensure that all ethnic groups and social classes are doing relatively well, and as such not allow for large differences in levels of adaptation.
More importantly, the immigrant adolescents in this study did not score worse in terms of adaptation than their national contemporaries, despite a strong disadvantage in terms of socio-economic status. As there are many developmental risks associated with poor socio-economic conditions (Bradley and Corwyn
2002; McLoyd
1998), it is remarkable that immigrant adolescents show as good an adaptation as national adolescents. Immigrant adolescents had a stronger sense of family obligations and a better school adjustment than national adolescents. This is in line with previous studies (Fuligni
1997; Fuligni et al.
1999; Steinberg
1996). In this study, a sample of non-western immigrant adolescents was used. They are more likely than national adolescents to embrace collectivistic values. As such, they may place a higher importance on helping, obeying and spending time with family members. Because immigrant adolescents see the sacrifices their families made to come to the new country for their better futures, and because they see school as a possibility to help their families in the future, they are usually well motivated to perform well in school (Fuligni
1998; Zhou and Bankston
1998).
For the immigrant adolescent sample, we hypothesized that the relationships between family obligations and mean grades, behavioral problems, psychological problems and self-esteem were partially mediated by school adjustment. We formulated this hypothesis based on literature that states that a strong sense of family obligations is related to a better school adjustment and a more positive pattern of adaptation (Fuligni
1997; Fuligni et al.
1999. Zhou and Bankston
1998), and to literature that states that school adjustment is related to a better pattern of adaptation (Bryant et al.
2003; Ludden and Eccles
2007; Steinberg
1996; Vazsonyi and Pickering
2003). The correlations indicated that in the immigrant adolescent sample family obligations were related to better grades, a better school adjustment, higher self-esteem, and fewer behavioral problems. The relationships between family obligations and self-esteem and family obligations and behavioral problems were partially mediated by school adjustment. These partial mediation effects indicate that part of the effect of family obligations on self-esteem and behavioral problems can be explained through school adjustment, but there is also a direct effect of family obligations on behavioral problems and self-esteem. Thus, the hypotheses were confirmed for two of the four measures of adaptation. In the national adolescent sample the correlations table indicated that a stronger sense of family obligations was related to a better school adjustment, fewer psychological problems and fewer behavioral problems. The relationship between family obligations and psychological problems was fully mediated by school adjustment and the relationship between family obligations and behavioral problems was partially mediated by school adjustment. The full mediation between family obligations and psychological problems indicates that the effect of family obligations on psychological problems can be completely explained by the better school adjustment.
In the national adolescent sample, the positive relationships that were found between family obligations and school adjustment and behavioral problems may be explained by the same process as in the immigrant adolescent sample. Those adolescents, national or immigrant, who feel a strong sense of obligation to their families may take more of an effort to do well in school to make their families proud and may be more hesitant to engage in delinquent acts to avoid bringing shame to their families. But what explains the differences between these two groups, why are family obligations related to self-esteem and mean grades in the immigrant but not in the national sample? Even more than national adolescents, immigrant adolescents may feel that they have a debt to pay to their parents who migrated so that they could have a better future. As such, approval of family members may be even more important, and those adolescents who are obedient to their parents may feel a sense of fulfillment translating in better self-esteem whereas those immigrant adolescents that disobey parents may feel more ashamed, which may translate into a lower self-esteem. The lower socio-economic status that immigrant adolescents likely grew up in may explain the relationship between family obligations and grades. Immigrant adolescents may feel a stronger need to help their families overcome comparatively poor living standards, and getting good grades may qualify them for a better education and a better job later (Freeburg and Stein
1996).
Limitations and Prospects
This study has several limitations. The reliabilities of the family obligations and school adaptation scales are on the low side. By using the variables self-esteem, behavioral problems, psychological problems and grades as dependent variables in the design, we aimed to provide a better understanding of the immigrant paradox. Given the inconclusive pattern of results in this study—full mediation, partial mediation and no mediation were found, other models need to be considered in future studies. Self-esteem was used as a dependent variable but it can be argued that self-esteem causes a better school adjustment (Rosenberg et al.
1995), while a reciprocal relationship between self-esteem and school achievement has also been suggested (Rosenberg et al.
1989). School adjustment has in this study been used as a predictor for behavioral and psychological problems, but it may be that behavioral and psychological problems lower a students’ motivation to do well in school. While this study provides evidence that family obligations and school adjustment are related to better adaptation outcomes in immigrant and national adolescents, the specific relationship between the variables remains unclear. New studies with longitudinal designs are needed to clear this up. Nonetheless, the results from this study that immigrant adolescents score higher on school adjustment and family obligations than national adolescents, and that family obligations and school adjustment are related to a higher self-esteem and fewer behavioral problems, suggest that family obligations and school adjustment may indeed be important factors in explaining the immigrant paradox.
An interesting finding is that national adolescents also seem to benefit from a strong sense of family obligations. However, they score substantially lower on family obligations than immigrant adolescents. This may provide an interesting opportunity for nationals, as learning from the immigrant community to increase the sense of family obligations may help improve national adolescents’ adaptation. Acculturation is a two way street as prolonged contact between cultural groups can change the culture of both groups. Ideally, this would entail positive changes for both groups. However, in the Netherlands there is a growing demand for immigrant groups to forsake their ethnic cultures and completely assimilate in mainstream Dutch culture (Arends-Toth and Van de Vijver
2003; Van Oudenhoven et al.
1998). This is unfortunate as forcing immigrants to completely abandon their ethnic culture does not only deprive immigrants from the means to ensure a positive adaptation but also robs nationals from valuable acculturation opportunities.