Migrants and educational achievement gaps

The socio-economic disadvantage of migrant children

Economists find a strong causal effect of parental education on child outcomes [2]. Students from socio-economically advantaged families perform better than students from disadvantaged backgrounds. As immigrants in most OECD countries have less education than their native-born counterparts, it is not surprising that performance gaps in immigrant youths’ reading, mathematics, and science achievement are explained largely by differences in parental socio-economic background and cultural status.

Ignoring all other relevant influences, Figure 1 depicts the high correlation between differences in parental socio-economic status and differences in reading literacy between immigrant and native students in selected OECD countries. In traditional immigrant-receiving countries that regulate immigration using a point- (skill-)based system (in particular Australia, Canada, and New Zealand), differences in the socio-economic background of parents between immigrants and native-born populations are small, or—in New Zealand’s case—even negative. Performance on the Programme for International Student Assessment (PISA) is similar for the two groups of students. By contrast, immigrant children do substantially worse than native-born children in European countries, where labor migrants account for most migration inflows and most migrants come from less advantaged socio-economic backgrounds. Considering that immigrant flows in the US are shaped by green-card regulation, the socio-economic profile of immigrants in the US looks surprisingly similar to that in Europe. It has to be noted, however, that US immigration is also influenced by illegal migration across the Mexican border and by waves of immigration amnesties.

Figure 1 suggests that immigrant students in the US perform better than predicted by the relatively high native–migrant gap in socio-economic background. This observation is in line with a study that focuses on second-generation immigrants using data from the 2003 and 2006 PISA [3]. The study suggests that traditional countries of immigration (Australia, Canada, New Zealand, and the US) might benefit from their long experience in absorbing new immigrants—for instance, by providing more transparent and less complex educational institutions than in European counties. Indeed, in European countries of labor migration (Austria, Belgium, Netherlands, and Switzerland), a large and highly significant performance gap of about 20–40 points on the PISA scale remained in both mathematics and reading tests after controlling for family background and language spoken in the home [3]. In general, the study concludes that the size of the persistent gap is highest in central and western Europe, followed by southern Europe. The Scandinavian countries of Denmark, Norway, and Sweden show only small gaps, and the traditional immigration countries show no mathematics or reading gap [3].

While these findings might suggest that traditional countries of immigration do better at absorbing new immigrants, it is also fair to say that the children of the high-skilled immigrants are a lighter burden on the education system than the children of migrants who are fleeing poverty, conflict, and political and religious persecution. The performance gap for this second group is large, and its causes—and thus solutions—are multidimensional.

Second-generation immigrants should gain from better integration into the local culture and schools as a result of having grown up in the host country. They benefit from well-settled immigrant networks and attendance at local kindergartens and primary schools. Indeed, conditional on socio-economic status and language spoken at home, being born in the host country is associated with higher PISA scores in all nine countries examined (four traditional countries of immigration, four major European immigration countries, and Finland), although the difference is statistically insignificant for New Zealand, Sweden, and the UK [4].

Low and devaluated cultural capital, books at home, and acculturation

The inequality between migrant and native-born students can also reflect a disadvantage originating in lower cultural capital among migrants, in the form of competencies, skills, knowledge, and values, and by objectified cultural capital, such as books. Inequality in cultural capital is related to differences in educational achievement not only between countries, but also within countries and schools [1], [3], [4]. A student’s cultural capital depends heavily on a child’s family. Moreover, because cultural capital is context-dependent and cannot be fully transferred between societies, valuable familial cultural capital will be devaluated after migration [5]. Language skills might be fully lost.

PISA evaluation surveys provide a unique opportunity to test the importance of cultural capital for educational achievement. There are large international differences in the objectified cultural capital of immigrant children across OECD countries, as measured by the number of books at home (Figure 2). There is also a strong positive correlation between books at home and academic performance: one additional book is associated with approximately one additional PISA point. Among a group of 17 OECD countries, the children of migrants in Italy, the Netherlands, and Spain have the smallest number of books at home, while the children of migrants in Australia, Canada, and New Zealand have more books and higher PISA scores. Again, migrant students in the US perform better than predicted by their embodied cultural capital. Another indicator of cultural capital, language spoken at home, also shows a statistically significant relationship to educational achievement as measured by the PISA.

Language gaps and related policies

Students take the PISA test at 15, an age that is critically important to future school and career opportunities in most countries. Language acquisition and literacy skills are essential to children’s educational achievement, and poor reading and language proficiency at this age can have lasting consequences.

International comparisons of PISA scores show that national differences between migrant and native-born students are very similar for reading, mathematics, and science proficiency. A 65-point difference on the 2009 PISA reading proficiency test separates a group of students most likely to perform poorly—first-generation migrant students with foreign-born parents and who speak a foreign language at home—from native students who speak the local language at home [1]. Even after adjusting for socio-economic status, the study finds a persistent gap of 41 points. Education achievement gaps are narrower for migrant youths who arrive at a younger age [6], [7].

A comparison of average gaps in PISA 2009 reading achievement scores of first-generation migrant students who speak the local language at home and those who speak their native language substantiates the importance of language deficits [1]. The largest gaps in reading scores between the two groups are found in Switzerland (71 PISA points) and New Zealand (62 points); the OECD average gap is 35 points. Gaps are very small in the Netherlands (12), the UK (13), and Canada (19). Also remarkable is the close to average educational achievement of migrants in Germany, which still had very high language-related PISA score gaps in 2000 [4]. By 2009, the score of the average migrant student in Germany had improved by 32 points and was no longer substantially different from the average OECD value.

Interpreting PISA gaps requires careful analysis of the composition of immigration flows [4]. For instance, in the UK, most migrants come from English-speaking countries, and the rest mostly come from European countries. As English is the common foreign language taught in European schools, most foreign parents and their children living in the UK already speak English. The situation is similar in Canada, which selects for high-skilled migrants, who are more likely to be bilingual. The picture is more complex in the Netherlands, which has a long history of immigration, and where migration policy has shifted from very liberal to more restrictive. Until the mid-1990s, there was generous support for immigrants’ integration, including providing instruction in the native language in primary schools. Immigration and integration policies became stricter after that, and language tests are now a condition of entry, although preservation of native cultures remains an important policy concern. The high proportion of immigrants from former Dutch colonies also affects immigrant integration in the Netherlands. Education systems in these former Dutch colonies are based on the Dutch system, which suggests less devaluation of immigrants’ cultural capital in the Netherlands than elsewhere [6].

Concentration

A high concentration or social segregation of immigrants can result from intentional selectivity mechanisms of the education system, such as tracking. However, even comprehensive education systems cannot avoid concentration that results from residential segregation. Immigrants may profit from such concentration because of potential efficiency gains in targeting the needs of students in more homogeneous classes. However, more concentration means missed opportunities for immigrant children from disadvantaged family backgrounds to benefit from social interaction in school with high-ability students from advantaged socio-economic backgrounds.

Empirical evidence shows that concentration in itself need not have negative consequences, provided that all schools are of the same quality [8]. However, ethnic segregation as a consequence of residential sorting may have deleterious effects on schools, and thus on students attending these disadvantaged schools. Such adverse neighborhood effects can result in larger differences in reading scores than those between students with high-educated mothers and those with low-educated mothers. Thus, not even having a highly educated mother can compensate migrant students for the penalty associated with attendance at a disadvantaged school. School disadvantage effects are more likely in countries such as France, Germany, and Italy than in Scandinavian countries [1].

Early selection by ability

The literature on the pros and cons of student tracking is rich and not restricted to concerns about migrant children. However, the trade-off between equal opportunities and efficiency associated with tracking is particularly important for children of disadvantaged migrant households. In most countries, education remains uniform and comprehensive through the lower secondary school level; selection does not generally occur until age 15 or 16. However, a few countries (such as Austria, Germany, Hungary, and Slovakia) track students based on ability as early as age 10.

Education researchers argue that early differentiation by skill level has a negative impact on the school performance of children who enter school with language and social deficits, a high proportion of whom come from immigrant families [10]. PISA results confirm that school systems that track students at an early age tend to show a stronger impact of socio-economic background on learning outcomes. This relationship implies increasing inequality between native-born and migrant students, whose socio-economic disadvantage is intensified by their language problems [1].

Empirical studies confirm this negative (though not always significant) effect of age at first selection by ability on average migrant outcomes. This effect persists even after carefully controlling for numerous variables that capture the effect of socio-economic background, including age at arrival, household structure, educational resources at home, and class composition, as well as national institutional factors such as school starting age [7].

Beneficial or adverse peer effects?

The combination of concentration of migrant students in schools and early tracking might lead to increasing inequality between disadvantaged groups of migrant students and more privileged native-born students. In testing the strength of peer relations within and between groups of migrant and native-born students, studies have shown that in countries of early selection in schools by ability, such as Austria and Germany, influential social interaction takes place mainly within groups, with interaction between groups being less pronounced [8]. This finding suggests that early tracking has the potential to magnify existing inequalities based on parents’ socio-economic background and to transmit them to the next generation.

Only a few studies are able to analyze the effect of reducing residential concentration. The 
Moving to Opportunity for Fair Housing Demonstration program, sponsored by the US Department of Housing and Urban Development in the 1990s, is one the most ambitious randomized controlled social experiments ever conducted. More than 4,600 low-income families living in high-poverty inner-city public housing were selected through a lottery to receive housing vouchers allowing them to move to private-market housing in less distressed communities. This program provides a rare opportunity to evaluate “clean” evidence on neighborhood effects by looking at the effects of changing families’ neighborhood environment. Families that received rental housing vouchers with no restrictions on area of residence (the treatment group) could be compared with families of similar socio-economic background that applied for the lottery but that did not win a voucher (the control group).

Many families who participated in the program are members of racial and ethnic minority groups, and thus are similar in some of their disadvantages and social isolation to immigrant families. While the program has had some positive effects on health outcomes, education and labor market performance were no better among families that moved than among families that did not [11]. A possible explanation is that the disadvantaged families that moved to more prosperous areas moved into neighborhoods that were also racially or ethnically segregated, so their exposure to high-quality schools might have been limited. The literature surveying the results of this program also reports that most families participating in the program were socially isolated in their new neighborhoods, thus restraining any influence of peer effects from contact with families from a higher socio-economic background.