Bertelsmann Stiftung (ed.) Immigrant Students Can Succeed

Bertelsmann Stiftung (ed.) Immigrant Students Can Succeed Lessons from around the Globe Carl Bertelsmann Prize 2008

Bertelsmann Stiftung (ed.) Immigrant Students Can Succeed Lessons from around the Globe Carl Bertelsmann Prize 2008

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available on the Internet at http://dnb.d-nb.de. 2010 ebook edition (PDF) 2008 Verlag Bertelsmann Stiftung, Gütersloh Responsible: Christal Morehouse Copy editor: Birte Pampel, München Production editor: Christiane Raffel Cover design: Nadine Humann Cover illustration: Dirk Eusterbrock Typesetting and printing: Hans Kock Buch- und Offsetdruck GmbH, Bielefeld ISBN 978-3-86793-251-6 www.bertelsmann-stiftung.org/publications

Contents Introduction... 7 Johannes Meier, Christal Morehouse An International Perspective on Student Achievement... 11 Gayle Christensen, Michael Segeritz Select Country Approaches to Education and Integration Classic Countries of Immigration... 37 Canada Diversity as an Expression of the Modern Canadian Education System... 37 Veronica Lacey United States of America Educating Children of Immigrants in the United States: A Call to Action... 46 Marcelo Suárez-Orozco, Carolyn Sattin Australia Building on Diversity... 60 Judith D. Chapman New Zealand Furthering that Spirit of Collaboration... 66 Cynthia White 5

Former Colonial Powers in Europe... 73 The United Kingdom Education as a Path to Integration... 73 Roisin L. Donachie The Netherlands Integration through Education... 79 Maurice Crul Countries of Post-War Labor Migration in Europe... 89 Switzerland International Benchmarking as a Motor for Reform... 89 Rosita Fibbi Norway Fostering a More Inclusive Society... 100 Barbro A. Bakken Sweden Multilingualism as a Key to Integration in Schools... 109 Monica Axelsson Integration through Education Promising Practices, Strategies and Initiatives in Ten Countries... 119 Hans J. Barth, Andreas Heimer, Iris Pfeiffer Contributors... 189 6

Introduction I have a dream that my four little children will one day live in a nation where they will not be judged by the color of their skin but by the content of their character. I have a dream today! Martin Luther King, Jr. Delivered 28 August 1963 Forty-five years ago, Martin Luther King delivered his famous I have a dream speech. Today, most societies still struggle to achieve his vision of equal opportunity, social cohesion and integration. Equal learning opportunities for children and young people are crucial for achieving this goal. In Germany, students ethnic origin and socioeconomic background continue to play a decisive role in school attainment. In OECD countries, students of minority ethnic origin and from underprivileged socioeconomic backgrounds on average perform more poorly than other students. This well-known fact is alarming. It suggests there are widespread and systemic flaws in schools ability to provide students with equal education opportunities across the OECD. This challenge requires a comprehensive and well-considered response from national governments. Especially in knowledge-based societies, educational attainment is a major determinant of an individual s further opportunities in life. Essentially, education policies can either reinforce or diminish inequalities in society. They can foster or hamper societal integration. If policies provide students with unequal learning opportunities, they can seriously limit an individual s long-term social mobility and deepen existing social cleavages. The consequences in terms of systematically wasting the human potential of groups of students will, therefore, be 7

more than just economic. Societal integration and social cohesion are dependent on equal opportunity and non-discrimination. Ultimately, this is a moral issue because the right to inclusive education is a human right. Therefore, fostering the integration of diverse populations through equal education opportunities must be a policy priority. In Germany, students unequal access to learning opportunities is particularly apparent. With regards to school attainment, fifteen yearold second generation students of immigrant origin lag more than two years behind native students. That is why the 2008 Carl Bertelsmann Prize has sought out innovative approaches to fair education in select OECD countries that promote the integration of children and youth of immigrant origin. This volume captures the main findings of a yearlong research process that spanned ten OECD countries. The research process incorporated expert input in the form of a workshop and entailed over 100 expert interviews as well as on-site investigations. This book presents new comparative research and analysis on policies and programs in these OECD countries, which provide students of immigrant origin with equal education opportunities. It aims to raise awareness about the urgent necessity to provide students with fair learning opportunities. The findings in this volume seek to inspire more enlightened education policy and practice worldwide, especially in countries where attainment gaps are large, such as in Germany. Chapter One begins with an analysis of international student attainment by Gayle Christensen and Michael Segeritz based on 2006 and prior PISA study results. The authors present an overview of how first and second generation students of immigrant origin perform as compared to non-immigrant origin students. The comparative analysis provides the reader with a snapshot of immigrant student performance and the role of socioeconomic status in immigrant achievement gaps. It finds that student achievement gaps have consistently correlated to the ethnic origin of students. This is the case even when accounting for students socioeconomic background. Chapter Two of this volume presents education policy strategies in nine OECD countries. The nine countries were selected based on the relatively high performance of their second generation students of immigrant origin as compared to Germany in the PISA studies. In traditional immigration countries, the diversity inherent in the student population is anchored in education policies. Veronica Lacey dem- 8

onstrates how diversity is an expression of modern education policy in Canada. Marcelo Suárez-Orozco and Carolyn Sattin detail the impact of migration on the US education system and the importance of incorporating the global context in education curricula. Judith Chapman explains how the Australian education system builds on a tradition of immigration and diversity that is rooted in the nation s history. Australian education policies are based on respecting diversity and the individual right to one s cultural traditions and beliefs. Cynthia White discusses education policies in New Zealand. There, the founding concepts of biculturalism have been a driving force for multicultural education policies that further a spirit of social cohesion. The second half of Chapter Two explores national approaches to integration through education in the UK and the Netherlands as well as in three countries with post-war labor migration: Switzerland, Norway and Sweden. Roisin Donachie outlines how the British education system has focused on measuring and closing achievement gaps in student attainment, especially among students of immigrant origin. Maurice Crul highlights the Dutch experience of systematically devoting government resources to schools with large populations of immigrant-origin students. He also provides an overview of effective programs that bridge and counterbalance inequalities in student attainment once they are detected. Rosita Fibbi investigates how international benchmarking of student attainment has impacted the Swiss education system and how Swiss education policies have evolved due to international benchmarking. Barbro Bakken analyses how the Norwegian government has taken a proactive approach to making education opportunities more equal in Norway with the goal of fostering a more inclusive society. Monica Axelsson explains the role of mother tongue tuition as a cornerstone of Swedish education policy. This policy provides mother tongue tuition as the right of a multilingual student and reinforces multilingualism not as a problem, but as an asset to Swedish society. Chapter Three of this volume concludes with a comparative analysis of promising programs in ten OECD countries that foster integration through education. Hans Barth, Andreas Heimer and Iris Pfeiffer provide a detailed assessment of such programs. International migration is a permanent fixture of life in the 21st century. In countries around the globe, student populations are becoming steadily more diverse. Yet education policy and practice have not 9

systematically developed the tools they need to harness the potential of this diversity in the vast majority of OECD countries. Education policy and practice must correct current disparities so as to provide equal learning opportunities for all students. Martin Luther King s dream is still inspiring us to achieve equal opportunity, social cohesion and integration, especially in the field of education. We would like to thank all the authors who contributed to this publication. This volume would not have been possible without the excellent work of this year s Carl Bertelsmann Prize team. The 2008 team was headed by Ulrich Kober and included Petra Rutkowsky, Christal Morehouse, Claudia Walther, Anja Hülsken, Christina Brickenkamp, Kristina Neumann, Orkan Kösemen and Anke Knopp. Dr. Johannes Meier Executive Board Member Bertelsmann Stiftung Christal Morehouse Project Manager Bertelsmann Stiftung 10

An International Perspective on Student Achievement Gayle Christensen, Michael Segeritz Schools are key institutions that can foster or hinder societal integration. As immigration becomes a norm in our global world, schools are increasingly challenged to provide equal education opportunities for diverse student bodies. Education systems cannot expect their student bodies to adapt to teaching methods of the past; education policy and practice must accurately assess and meet the needs of their student bodies. Success of immigrant children in school plays an important role in supporting social and economic integration by providing young people with the opportunity to learn vital skills, the language and the culture of the host country. However, immigrant students in many countries are not as successful in school as their native peers, indicating that special attention must be paid to ensure that these students have the support they need to succeed (Stanat and Christensen 2006). This brief summary paper examines immigrant student achievement across 16 immigrant-receiving countries in 2003 and 2006. It also looks at how socioeconomic status is associated with both student achievement and immigrant status. Immigrant Students and the Program for International Student Assessment (PISA) The OECD s Program for International Student Assessment (PISA) has been conducted every three years since 2000. The advantage of using the PISA data to examine immigrant student achievement is that it allows for international comparisons of student performance. In 2006, 57 countries participated. The focus of the assessment that year was science. Reading and mathematics abilities were also tested. The PISA assessment has a literacy focus, meaning that students are 11

asked to apply their knowledge and understanding to real-life situations. The assessments in each domain are scaled such that for the baseline year (e.g., for science the baseline year is 2006), the average student score in OECD countries is 500 points with a standard deviation of 100 points. The same scale is then applied for the following cycles, which makes it possible to compare different years on the same scale. This paper examines performance in all three domains. The student background questionnaires included as part of the PISA allow for comparisons between three sub-groups: first-generation students (foreign-born students with foreign-born parents) second-generation students (students born in the country being assessed, with two foreign-born parents) native students (students with at least one parent born in the country of the assessment) 1 The term immigrant students is used to refer to both first-generation and second-generation students. Only a portion of the countries participating in the PISA have a large enough fraction of immigrant students in the sample to yield reliable estimates of their achievement levels. In this paper, we include countries with a minimum of three percent in each of the immigrant subgroups (first-generation and second-generation). Country samples also had to include a minimum of 100 first-generation and 100 second-generation students. Countries included in this paper are Australia, Austria, Belgium, Canada, Denmark, France, Germany, Israel, Luxembourg, Netherlands, New Zealand, Norway, Sweden, Switzerland, the United Kingdom and the United States. 2 The percentage of first-generation students in these countries ranges from three percent in Norway to nearly 20 percent in Luxembourg. The range is similar for second-generation students from three percent in Norway to almost 17 percent in Luxembourg (see Table 1). Across 1 Students born in the assessed country who had one foreign-born parent were included in the native or non-immigrant category as previous research has indicated that these individuals perform at similar levels as non-immigrant students. 2 We use data for 2003 and 2006 where available. Data from 2003 is not available for Israel or the United Kingdom. Reading data for 2006 is not available for the United States. 12

these countries, there are substantial variations in countries of origin, socioeconomic status and language background. PISA of course has its limitations for conducting analysis of student performance, including the fact that the data are not longitudinal, making it impossible to examine student performance across consistent time periods. In addition, it is only possible in a very limited way to examine the role of the country of origin regarding immigrant students performance, as many countries do not include this question in their questionnaires. The sample size in most countries would be too small to obtain reliable comparisons. Immigrant Student Achievement in Mathematics, Reading and Science This section examines performance differences between immigrant and non-immigrant students across the three subject domains tested (science, reading and mathematics). For reading and math, we are able to explore the performance differences in 2003 and 2006 to see if these differences remain similar across a period of several years. 3 2006 science assessment results Science is the focus domain for 2006 and this section first examines the performance differences among the subgroups in science. Firstgeneration students may face the greatest challenges in adjusting to a new school system as they directly experience the challenges of migration. The results indicate that in most countries first-generation students perform at significantly lower levels than their native peers (see Figure 1). 4 In fact, first-generation immigrant students in OECD countries lag on average more than 50 points behind their peers, which is roughly 3 The science scale for 2006 cannot be linked with previous assessments. However, it will be possible to link science results in future PISA assessments (see OECD 2007). 4 Significant differences throughout the paper are calculated at p < 0.05. For multiple comparisons, the Bonferroni correction is applied. 13

Figure 1: Science performance differences between native and immigrant students PISA 2006 Australia Austria Belgium Canada Denmark France Germany Israel Luxembourg Netherlands New Zealand Norway Sweden Switzerland United Kingdom United States 92 89 80 93 85 89 67 93 77 66 67 79 68 58 60 78 69 95 57 48 48 41 43 2 3 12 22 17 28 10 26 6 120 100 80 60 40 20 0 20 Second-generation students First-generation students Source: OECD PISA 2006 database, Table 2. equivalent to one and a half years of schooling when considering the OECD average difference in performance between school years (OECD 2007). However, there is substantial variation across countries. In the majority of the countries (11) included in this paper, first-generation students score more than 50 points less than non-immigrant students (Austria, Belgium, Denmark, France, Germany, Luxembourg, the Netherlands, Norway, Sweden, Switzerland and the United States). In contrast, there are no significant differences between first-genera- 14

tion students and their native peers in Australia, Israel and New Zealand (see Figure 1). One might expect very different results for second-generation students, as they have experienced all of their schooling in the same school system as native students. However, in all of the case countries, except Australia, second-generation students perform at significantly lower levels than non-immigrant students. In many countries, the science performance differences are smaller than for first-generation students. As a matter of fact, in Canada, Sweden and Switzerland, secondgeneration students perform significantly better than first generation-students, although in Switzerland and Sweden the performance gap between immigrant and native students remains large (69 and 48 points respectively (OECD 2007). In eight countries (Austria, Belgium, Denmark, Germany, Luxembourg, the Netherlands, Norway and Switzerland), the performance difference between second-generation and native students is greater than 50 points (roughly the OECD average equivalent of about one and a half years of schooling) on the science scale 5. This should be of particular concern, given that these students have spent all of their school years in the same country. 2006 reading and math assessment results The achievement gaps for immigrant and non-immigrant students in reading and mathematics in 2006 are similar to those found in science in 2006 (see Figures 2 and 3). With regard to reading, first- and second-generation students lag significantly behind their native peers in all of the assessed countries, except Australia and Israel and second-generation students in Canada, New Zealand and the United Kingdom. As with science, the achievement gaps between second-generation students and native students tend to be smaller than those for firstgeneration students. Again, the achievement gap is still quite large 5 There are differences across countries in the average difference of performance across school years. 38 points is the average difference across OECD countries. 15

Figure 2: Reading performance differences between native and immigrant students PISA 2003 PISA 2006 Australia 4 12 7 1 Austria 73 77 79 48 Belgium 117 84 81 102 Canada 19 10 19 0 Denmark 57 42 64 79 France 79 48 36 45 Germany 96 86 83 70 Israel 7 1 Luxembourg 69 47 61 69 Netherlands 50 61 61 65 New Zealand 22 25 7 19 Norway 59 68 63 42 Sweden 89 20 68 29 Switzerland 93 53 85 48 United Kingdom 44 7 United States 50 22 120 100 80 60 40 20 0 20 120 100 80 60 40 20 0 20 Second-generation students First-generation students Source: OECD PISA 2003 & PISA 2006 database, Table 3a & 3b. for second-generation students in some countries more than 50 points (roughly the OECD average equivalent of about one and a half years of schooling) in Austria, Belgium, Denmark, Germany, Luxembourg and the Netherlands. The results for mathematics are similar, with significant achievement differences between the two immigrant subgroups and native students in all of the countries except Australia (where immigrant students actually perform better than native students), Canada, Israel and New Zealand. 16

As in reading, a number of European countries (Austria, Belgium, Denmark, Germany, the Netherlands and Switzerland) exhibit performance differences of more than 50 points on the mathematics scale between second-generation and native students. One might expect that the reading differences would be even greater than for the other subjects, however, the literacy focus of all of the PISA assessments may help to explain why the results are similar across the domains. For reading and mathematics, it is possible to compare performance across the 2003 and 2006 PISA assessment data (see Figures 2 and 3). What is most striking for both of these subject areas is the similarity of the results across years. In general, the achievement differences between immigrant and non-immigrant students are within 10 points of each other, and there are only a few cases where the performance differences change by more than 20 points across years. For reading, the achievement differences between first-generation and native students decreased by at least 20 points from 2003 to 2006 in three countries Austria, France and Sweden and grew by at least 20 points in Denmark. For secondgeneration students, there is not a single country which exhibits such a large change in the performance difference (see Figure 2). When examining the mathematics results from 2003 to 2006, there is also only a small number of countries where the immigrant achievement gap changes by more than 20 points. From 2003 to 2006, the performance difference between first-generation and native students decreased at least 20 points in the Netherlands and Sweden. This is also the case for the performance difference between secondgeneration and native students in New Zealand. In contrast, the performance gap between second-generation and native students in Austria is more than 20 points larger in 2006 as compared to 2003 (see Figure 3). Even in the cases where there are seemingly large changes from 2003 to 2006, the results need to be interpreted with caution. The measurement error tends to be higher when conducting analysis among subgroups, which means that these changes are unlikely to be statistically significant. Furthermore, given the cross-sectional nature of the data, changes may be a result of cohort differences among the samples. As such data is measured across a longer time period, clearer patterns may emerge. The limitations, however, do not diminish the most 17

Figure 3: Mathematics performance differences between native and immigrant students PISA 2003 PISA 2006 Australia 5 2 12 11 Austria 56 63 81 65 Belgium 92 109 112 84 Canada 7 6 8 7 Denmark 70 65 80 63 France 72 48 47 62 Germany 93 71 78 65 Israel 10 9 Luxembourg 31 45 46 55 Netherlands 79 59 66 58 New Zealand 32 5 13 6 Norway 61 39 49 58 Sweden 92 34 64 42 Switzerland 89 59 88 62 United Kingdom 25 25 United States 22 36 23 37 120 100 80 60 40 20 0 20 120 100 80 60 40 20 0 20 Second-generation students First-generation students Source: OECD PISA 2003 & PISA 2006 database, Table 4a & 4b. noteworthy finding across subject domains and years, namely that there are large and significant performance differences among immigrant and non-immigrant students in the majority of countries. Of particular concern, especially for policy makers, should be the fact that second-generation immigrant students in many countries continue to lag significantly behind their native peers despite spending all of their schooling in the receiving country. 18

Socioeconomic Status and Immigrant Student Achievement Decades of educational research have shown the strong relationship between socioeconomic background and student achievement. The PISA data also shows that students with more advantaged backgrounds generally score higher on the PISA (OECD 2007). However, the strength of the association between socioeconomic status (SES) and student performance varies across countries. In fact, there are some countries that have higher than average performance with more similar outcomes for students from advantaged and disadvantaged backgrounds. Figure 4 shows this relationship between students backgrounds and performance by plotting average performance in science on the vertical axis with the percentage of variance in performance explained by the PISA SES index on the horizontal axis. 6 This index has been standardized for each country with a mean of 0 and a standard deviation of 1. This figure provides a general sense of the equity of learning opportunities as measured by achievement. The countries in the upper left quadrant, Australia, Canada and the United Kingdom exhibit both high levels of student achievement in science and a lower than average association between SES and performance. These countries indicate that it is possible to simultaneously achieve high levels of performance and greater equity across social backgrounds. The majority of countries, however, does not reside in the desirable quadrant of high performance and relatively higher equity. Some countries, including Austria, Belgium, Germany, the Netherlands, New Zealand and Switzerland, exhibit higher than average levels of science performance, yet they also have a stronger than average association between SES and performance, indicating that there are particularly pronounced differences between advantaged and disadvantaged students. Other countries perform below average and 6 The components comprising the PISA index of economic, social and cultural status include home possessions, the index of highest occupational status of parents, the index of highest educational level of the parents and national item parameters. The scale construction was done through Item Response Theory (IRT) scaling, with item parameters estimated first for common items separately for each country. The sum of the common items parameters was constrained to zero for each country. The item parameters were anchored and the remaining items were added. Each country was scaled separately (from OECD 2007: 333). 19

Figure 4: Student performance in science and the association with socioeconomic background Science performance 550 540 530 520 510 500 490 480 470 460 450 5 Canada Australia New Zealand Netherlands United Kingdom A ustria Germany Belgium Sweden Switzerland France Norway Denmark United States Luxembourg Israel 10 15 20 25 R-squared Note: Country average performance on the science scale and explained variance in science performance by ESCS. Source: OECD PISA 2006 database, Table 5. have a strong association with SES (France, Luxembourg and the United States) or they perform below average and have a lower than average association with SES (Denmark, Israel, Norway and Sweden). Ideally, the countries in the upper left quadrant may provide some models for the possibility of higher performance and equity. What does the association between student performance and socioeconomic background mean for immigrant students? Immigrant students vary not only in terms of their countries of origin, but also with regards to the the human and social capital that is available to them and the socioeconomic status they achieve once they settle in the host country. In the majority of case countries, immigrant students are from families with significantly lower levels on the socioeconomic status index as compared to native students (see Table 6). First-generation students are from families with significantly lower levels of SES in every country, except Australia, Canada, New Zealand (where first-generation students have higher levels on the SES index) and the United Kingdom. The significant differences range from 0.45 in Israel to 1.06 in Germany. Second-generation students in every country, except Australia, exhibit significantly lower 20