Understanding School Segregation: Patterns, Causes and Consequences of Spatial Inequalities in Education 9781350033511, 9781350033542, 9781350033528

Citation preview

Understanding School Segregation

Also available from Bloomsbury Education and Disadvantaged Children and Young People, edited by Mitsuko Matsumoto Education in South America, edited by Simon Schwartzman Education, Poverty, Malnutrition and Famine, edited by Lorraine Pe Symaco Governance of Educational Trajectories in Europe, edited by Andreas Walther, Marcelo Parreira do Amaral, Morena Cuconato and Roger Dale

Understanding School Segregation Patterns, Causes and Consequences of Spatial Inequalities in Education Edited by Xavier Bonal and Cristián Bellei

BLOOMSBURY ACADEMIC Bloomsbury Publishing Plc 50 Bedford Square, London, WC1B 3DP, UK 1385 Broadway, New York, NY 10018, USA BLOOMSBURY, BLOOMSBURY ACADEMIC and the Diana logo are trademarks of Bloomsbury Publishing Plc First published in Great Britain 2019 Paperback first published 2020 Copyright © Xavier Bonal, Cristián Bellei and Contributors, 2019 Xavier Bonal, Cristián Bellei and Contributors have asserted their right under the Copyright, Designs and Patents Act, 1988, to be identified as Authors of this work. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage or retrieval system, without prior permission in writing from the publishers. Bloomsbury Publishing Plc does not have any control over, or responsibility for, any third-party websites referred to or in this book. All internet addresses given in this book were correct at the time of going to press. The author and publisher regret any inconvenience caused if addresses have changed or sites have ceased to exist, but can accept no responsibility for any such changes. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress. ISBN: HB: 978-1-3500-3351-1 PB: 978-1-3501-5969-3 ePDF: 978-1-3500-3352-8 eBook: 978-1-3500-3353-5 Typeset by RefineCatch Limited, Bungay, Suffolk To find out more about our authors and books visit www.bloomsbury.com and sign up for our newsletters.

Contents Notes on Contributors 1

Introduction: The Renaissance of School Segregation in a Context of Globalization  Xavier Bonal and Cristián Bellei

vii

1

Part I  Patterns of School Segregation and Social Inequalities 2

3

4

School Segregation in France: The Role of Public Policies and Stakeholder Strategies  Georges Felouzis, Barbara Fouquet-Chauprade and Samuel Charmillot

29

Structural and Systemic Dimensions of School Segregation in French-­speaking Belgium  Vincent Dupriez, Samir Barbana and Marie Verhoeven

45

Patterns of School Segregation in Brazil: Inequalities and Education Policy  Tiago Lisboa Bartholo and Marcio da Costa

65

Part II  School Segregation and Student Performance 5 6 7

Refining Measures of Poverty and Their Impact on Student Progress in England  Stephen Gorard and Nadia Siddiqui

85

An Evaluation of the Intensity and Impacts of Socio-­economic School Segregation in Argentina  Natalia Krüger

103

A Synthesis of Social Science Research on the Effects of Ethnic, Racial and Socio-­economic Composition of Schools in the United States  Roslyn Arlin Mickelson

123

Part III  Market Dynamics and School Segregation 8

School Segregation in the Free School Choice Context of Dutch Cities  Willem R. Boterman

155

Contents

vi 9

The Ungoverned Education Market and the Deepening of Socio-­economic School Segregation in Peru  María Balarin and Aurora Escudero

179

10 School Segregation in the Spanish Quasi-­market Education System: Local Dynamics and Policy Absences  Xavier Bonal and Adrián Zancajo

201

11 The Production of Socio-­economic Segregation in Chilean Education: School Choice, Social Class and Market Dynamics  Cristián Bellei, Mariana Contreras, Manuel Canales and Víctor Orellana

221

Part IV  Conclusions 12 International Perspectives, Patterns and Challenges on School Segregation: An Interview with Professor Gary Orfield  Xavier Bonal and Cristián Bellei

243

Index

253

Notes on Contributors María Balarin is a Senior Researcher at the Group for the Analysis of Development (GRADE) where she has a portfolio of both applied and academic research focusing mostly on education and youth. Her academic work focuses on the field of education, where she is concerned with understanding how educational processes mediate relations between the state and society and the construction of educational policies. Her recent work focuses on the impact of educational markets on patterns of educational and social segregation, and on vulnerable youth transitions in the context of exclusionary citizenship regimes. Samir Barbana is a teaching-­assistant based at the faculty of Psychology and Educational Sciences, University of Louvain-­la-Neuve, UCL. He is also a researcher at the Groupe Interdisciplinaire de recherche sur la socialisation, l’éducation et la formation (GIRSEF), where he conducts studies on policy implementation. Tiago Lisboa Bartholo holds a PhD in Education and he is currently associate professor at the Federal University of Rio de Janeiro. His research concerns the distribution of educational opportunities and the robust evaluation of educational programmes and policies – especially those that can break the cycle of poverty. He has also conducted the evaluation of educational programmes and worked with large-­scale secondary data resources from Brazil and England. He has been visiting scholar in several European universities. Cristián Bellei holds a PhD in Education from Harvard University. He is an associate researcher at the Center for Advanced Research in Education and Professor in the Sociology Department at the University of Chile. He previously worked at the Chilean Ministry of Education and UNICEF in Chile. He has also been a consultant for international organizations such as UNESCO, the World Bank and the Inter-American Development Bank. His main research areas are educational policy, educational equity and school improvement. He has published extensively about quality and equity in Chilean education. Xavier Bonal is Professor of Sociology at the Universitat Autònoma de Barcelona (UAB) and Special Professor of Education and International Development at the University of Amsterdam (UvA). He is the director of the research group Globalisation, Education and Social Policies (GEPS) at the UAB and Coordinator of the GLOBED Project, an Erasmus Mundus Master on Education Policies for Global Development. He has widely published on sociology of education, education policy and globalization and education and development. He has worked as a consultant for international

viii

Notes on Contributors

organizations such as UNESCO, UNICEF, the European Commission and the Council of Europe. Willem R. Boterman is Assistant Professor of Urban Geography in the department of Geography, Planning and International Development Studies at the University of Amsterdam. He holds a Masters in political science and a PhD in human geography. His research has mainly focused on the interaction of demographic and class change, families, residential choice, housing and gentrification. His most recent work concentrates on issues of class and gender and social reproduction via school and residential practices, and on middle-­class disaffiliation, segregation and social and spatial polarization. He is currently working on a four-­year research project on the role of space in the reproduction of educational inequalities and school segregation. Manuel Canales holds a PhD in Sociology from the Universidad Complutense de Madrid, 1989. He is current Director of the Social Sciences Institute of Universidad de O′Higgins, Rancagua de Chile. He is also a senior lecturer at the Sociology Department at the Universidad de Chile. His main research areas are qualitative methodology and critical social aspects of the Chilean society. Samuel Charmillot is a lecturer and researcher at the Faculty of Psychology and Educational Sciences, University of Geneva (Switzerland). His research focuses on the links between the organization of education systems, school segregation and educational inequalities. Mariana Contreras is a sociologist from the University of Chile. She works as a research assistant at the Centre for Advanced Research in Education. She has conducted research and published about the sociocultural dimension of school choice and its relationship with socioeconomic school segregation in Chile. She is currently coordinating a multiple case study on high school improvement processes. Marcio da Costa holds a PhD in Sociology from the Instituto Universitário de Pesquisas do Rio de Janeiro. Since 1996, he has been Professor at the Federal University of Rio de Janeiro, where he created the Research Lab on Educational Opportunities and was its first coordinator. Presently he is the head of the Escola de Formação Paulo Freire, the institution in charge of teachers’ advanced training in the Department of Education of Rio de Janeiro. His main research interests focus on sociology of education, educational policies and educational evaluation. Vincent Dupriez is Professor of Education at the Université Catholique de Louvain (UCL) and senior researcher at the Interdisciplinary Research Group in Socialisation, Education and Training (GIRSEF). He has developed research in the areas of education policy, educational inequalities, school administration and comparative education. Currently his main research concerns the teaching profession and educational policies’ design and implementation.

Notes on Contributors

ix

Aurora Escudero holds a BA in Sociology from the Pontifical Catholic University of Peru (PUCP). She is currently a Research Assistant at the Group for the Analysis of Development (GRADE), where she has developed applied and academic research mostly focused on the field of education. Her recent work uses both mixed and qualitative methods in the fields of public policy, education and gender studies. Georges Felouzis is a full professor at the Faculty of Psychology and Educational Sciences, University of Geneva (Switzerland). His research focuses on education policies in Switzerland and France, on educational inequalities and on international comparisons in education. Barbara Fouquet-Chauprade is a senior lecturer and researcher at the Faculty of Psychology and Educational Sciences, University of Geneva (Switzerland). Her work focuses on the design, the implementation and the effects of education policies, on educational inequalities, and on school segregation and priority education policies. Stephen Gorard is Professor of Education and Public Policy, and Director of the Evidence Centre for Education, Durham University (https://www.dur.ac.uk). He is a Fellow of the Academy of Social Sciences, member of the British Academy grants panel, and Lead Editor for Review of Education. His work concerns the robust evaluation of education as a lifelong process, focused on issues of equity, especially regarding school intakes. He is author of around thirty books and over 1,000 other publications. He is funded by the Economic and Social Research Council (ESRC) to investigate measures of educational disadvantage and how such measures can be used for school improvement (ES/N012046/1). Natalia Krüger is an Assistant Researcher at CONICET (National Scientific and Technical Research Council) and an Assistant Professor in the Department of Economics at Universidad Nacional del Sur (National University of the South) in Bahía Blanca, Argentina. She holds a PhD in Economics from Universidad Nacional del Sur, and has carried out doctoral and postdoctoral research visits at Universidad de Barcelona (Spain), Universidade Federal Fluminense (Brazil) and Katholieke Universiteit Leuven (Belgium). Her research interests lie in the area of economics of education, with a focus on educational quality and equality of opportunities in Latin America. Roslyn Arlin Mickelson is Chancellor’s Professor and professor of sociology, public policy, and women and gender studies at the University of North Carolina at Charlotte, USA. Her interests include how race, gender, and social class shape educational opportunities, processes and outcomes. Her current research investigates social and organizational contexts that predict female and underrepresented minority college students’ success in science, technology, engineering and mathematics (STEM), and the cumulative disadvantages of school and classroom racial, ethnic and class segregation on educational outcomes from elementary school through to college.

x

Notes on Contributors

Víctor Orellana holds a Master in Social Sciences from the University of Chile. He is an assistant researcher at the Center for Advanced Research in Education at the University of Chile, and Director of Education at Nodo XXI Foundation. His main research areas are sociology of education, education quality, higher education and social stratification. Nadia Siddiqui is Assistant Professor at School of Education, Durham University. She is author of several peer-­reviewed journal articles on research methods, evaluations of school-­based interventions, and equity in education. Her research interests are in studying stubborn patterns of poverty and inequalities through population data sets and large-­scale surveys such as National Pupil Database (NPD, England), Higher Education Statistics (HESA) and Annual Survey of Education Report data (ASER, Pakistan). Using these secondary data resources, she has investigated the indicators of disadvantage that have deep associations with children’s academic attainment, well-­ being and happiness, and access to pathways for a successful life. Marie Verhoeven is professor of sociology and researcher at the Groupe Interdisciplinaire de recherche sur la socialisation, l’éducation et la formation (GIRSEF), Catholic University of Louvain, Belgium. Her research interests focus on: (i) the socializing and controlling functions of schooling; (ii) cultural diversity and integration policies in education; (iii) the intersection between socio-­economical inequalities and ethnic inequalities in schooling; (iv) segregation and systemic discrimination; and (v) social justice and social justice theory. She has been part of several national and international research projects (on the regulation changes in the educational field, educational policies, segregation or ethnic minority students’ school careers and identities). Adrián Zancajo is a postdoctoral research fellow at the Catholic University of Louvain and member of the Interdisciplinary Research Group in Socialisation, Education and Training (GIRSEF). He holds a PhD in Sociology from the University Autònoma of Barcelona. He is also a member of the Globalization, Education and Social Policies research group at the UAB. His main research topics focus on the political economy of educational privatization, education markets, educational inequalities and school segregation. He has collaborated with Education International, Open Society Foundation and Fundació Jaume Bofill. He is currently working on a research project on market decommodification reforms in Chile and Belgium.

1

Introduction: The Renaissance of School Segregation in a Context of Globalization Xavier Bonal and Cristián Bellei

1.  School segregation as a field of study in the 21st century The uneven distribution of pupils in schools, according to their social origin, ethnic group, sex or any other ascriptive characteristic, is an important topic in educational research. It has received special attention in the USA. The 1954 Supreme Court sentence in the case of Brown vs. Board of Education, in a context of historical apartheid, converted school segregation and desegregation policies into a central object of study for academics and researchers. Since the 1960s, especially after the publication of the Coleman Report (Coleman et  al., 1966), the amount of literature on this topic has grown extensively and is notably larger than that found in other parts of the world. A number of well-­known authors (Orfield, 2001; Orfield and Lee, 2005; Hanushek et al., 2003; Ogbu, 2003; Saporito, 2003) has focused on different aspects of the study of school segregation (scale, spatial dynamics, consequences for performance, social cohesion, effects of desegregation policies). These and other authors have constructed a body of research, which has maintained interest in school segregation for more than fifty years. While racial segregation obviously dominates most of the research on this topic in the USA, the study of socio-­economic segregation has also received growing attention in recent decades (Rumberger and Palardy, 2005; Palardy, 2013; see also the excellent review by Roslyn Mickelson in this volume, p. 123). Several court sentences abandoning the historical doctrine of ‘separate but equal’ in US schools have opened the door to desegregation policies, especially with regard to bussing plans in many school districts. Although these policies were active in the 1970s, they started to decline from the 1980s onwards. However, desegregation policies have remained controversial in US education (Noblit, 2015) and their efficacy can be questioned after decades of resegregation in US schools (Orfield, 2001; Frankenberg and Orfield, 2012). Nothing similar has occurred in other parts of the world, either because of a lack of a history of explicit apartheid, followed by massive desegregation processes, or due to lack of means and data with which to carry out the same type of research. However, in recent years, this has begun to change. In Europe, as the contributors of this volume show, better access to data, especially after the Programme for International Student Assessment (PISA), the crucial importance of migration movements and trends in the

2

Understanding School Segregation

direction of more market-­oriented reforms in education have all generated a growing interest in the study of school segregation in different countries and enhanced international comparisons (Benito et al., 2014; Alegre and Ferrer, 2010; Dronkers and Robert, 2008; Gorard and Smith, 2004). The study of ethnic segregation has been less central than it has been in the USA. Despite its importance, the discrimination and school segregation faced by Roma children in European societies have received much less attention than research on racial inequalities in the USA.1 The migrant conditions of children, socio-­economic variables or special needs proxies have actually concentrated the focus of most empirical research on school segregation in European countries. In Latin America, where social and economic inequalities are remarkably high, the study of school segregation has only been a priority in recent years. Research has mainly been focused on factors accounting for the high levels of exclusion from the school system, low academic performance and noticeable educational inequity. Nevertheless, internal and external migration movements, processes of urban polarization and new educational policies (such as the promotion of private schools and different forms of decreasing state responsibility) have increased educational differentiation among social groups. These trends are combined with more traditional forms of educational segmentation, which are reinforced in the form of educational access extended to low-­income populations: while, in urban areas, high socio-­economic status (SES) families tend to attend private or selective public schools, and in rural zones, indigenous people live and study in highly segregated environments. All these factors and the improvement in the quality of available data (mainly linked to international studies, such as the PISA) have facilitated an increasing number of studies on school segregation (many of them referred to in this book), motivated not only by a broad concern for equity, but also by its potential contribution to improve students’ academic performance. Additionally, globalization has expanded the processes of fragmentation and segmentation in large Latin American cities, which have also increased the level of interest among scholars in this emerging field. More than sixty years after the Brown vs. Board of Education sentence, the topic of school segregation not only remains an important area of educational research, but has gained momentum in recent decades. Globalization has undoubtedly impacted this renaissance in school segregation studies. Social inequalities have increased in many of the urban spaces in the globalized world. Economic growth and social development have been unequally distributed and generated growing processes of urban fragmentation and segmentation across neighbourhoods. As Musterd et  al. (2017) argue, globalization and polarization have taken place simultaneously, even in cities of countries with strong welfare regimes. Moreover, the increase in social inequalities since the mid–1970s and the incapacity of governments to use public and social policies to reduce them (Atkinson, 2015) have contributed to worsening socio-­ economic segregation in major cities. Global inequalities, economic globalization and political conflicts are also affecting migration movements all over the world. Despite the many restrictions on the movement of people, compared to the flow of goods and capital, the number of migrants in the world reached nearly 250 million in 2013 (3.4% See the works by Arabadjieva (2015), Cashman (2016) and Shattuck (2012).

1

Introduction

3

of the world’s population), compared to 120 million in 1990.2 Socio-­economic spatial segregation has increased as a result of all these trends, although its intensity and particular characteristics have varied in different countries and cities, depending on factors such as pre-­existing urban segregation, the process of economic restructuring, the development and transformation of welfare states or the characteristics of housing policies (Musterd and Ostendorf, 1998; Tammaru et al., 2015). Education has not been immune to these tendencies. Residential segregation, migration movements, economic inequalities and sometimes education policies themselves have mostly produced an increasing process of school segregation between the most disadvantaged social groups and the upper classes of society. Obviously, these processes have adopted different shapes and intensities in different parts of the world, from the reinforcement of traditionally segmented curriculum structures to the appearance of new divisions related to market dynamics and privatization processes. Inequalities are produced and reproduced for various reasons, and experienced in a distinct manner by different social groups. Social cleavages are imported into the education systems for economic, cultural, religious, linguistic, sexual and many other factors that divide societies. Indeed, what these divisions have in common is that they significantly impact on the life opportunities of underprivileged groups, whether they are girls, poor, black, religious or linguistic minorities. Hence, the relevance of school segregation as a field of study rests primarily on the fact that it is one of the most important factors explaining the reproduction of education inequalities. As noted in some country cases included in this book, research on factors predicting academic achievement has played a relevant role in boosting studies on school segregation, since a lack of educational inclusion has crucial consequences regarding the educational opportunities of the most vulnerable populations. The study of school segregation has made enormous progress in refining measures to estimate its magnitude, as well as in implementing more complex designs to identify its effects on different dimensions of students’ experiences. Contrary to certain positions that might find civic virtues in voluntary and spontaneous school segregation processes (Merry, 2012) and homogeneous school communities (Chubb and Moe, 1990), there is a great deal of evidence to show that school segregation has negative effects on the performance of the most disadvantaged students (Dupriez et al., 2008; Dumay and Dupriez, 2008; Thrupp et al., 2002; Rangvid, 2007). Despite the existence of relevant school effects (organizational and pedagogical practices), which can have a positive impact on learning processes and outcomes of students, the net effect of school composition variables on student performance tends to be greater than the net effect of variables related to pedagogic practices and organizational processes (Benito et al., 2014; see also the chapter by Natalia Krüger in this volume, p. 103). We have also learned that the concentration of higher- or lower-­ ability students has a greater effect on performance than ascriptive characteristics (Rumerger and Palardy, 2005; Van der Slik et al., 2006), as well as that compositional effects are asymmetric because they are strongest among the most underprivileged See https://siteresources.worldbank.org/INTPROSPECTS/Resources/334934-1199807908806/ 4549025-1450455807487/Factbookpart1.pdf. Retrieved on 3 October 2017.

2

4

Understanding School Segregation

students. Put another way, the educational performance of ethnic minorities or low SES students is more sensitive to composition effects than it is in the case of higher SES students (Hanushek et al., 2002; Andersen and Thomsen, 2011). Beyond inequalities in educational performance, research has also shown the negative effects of school segregation on other aspects of social life. Schools’ social and ethnic heterogeneity can be decisive in terms of students’ capacity to develop social capital and in promoting intercultural friendship networks (Van Houtte and Stevens, 2009; Tropp and Prenovost, 2008). Religious, linguistic or ethnic mixing can also facilitate a reduction in prejudice and higher levels of contact between opposed communities in contexts of political conflict (Hughes et  al., 2013). Schools’ racial, ethnic or social composition is also related to aspects such as violence, racial prejudice, labour market outcomes, wages or a sense of democracy (see the review by Mickleson in this volume, p. 123). Despite all this evidence, policies to tackle school segregation have been timorous at best. As we shall see later in this introduction, national, regional or local governments have been reluctant to introduce significant changes in education policies to achieve a more balanced distribution of students in schools. This resistance could be the result of a complex set of political, economic and ideological interests, which pressure education policy. Altering school choice models, making elite schools accept lower SES or ethnic minority students and changing school zones are decisions that face resistance from different sectors of the educational community, such as parents, private schools, school principals or even teachers. As discussed below, research on desegregation policies, which is correspondingly scarce, documents the technical, cultural and political difficulties accounting for their observed low levels of effectiveness.

2.  Unpacking the social mechanisms producing school segregation Understanding school segregation processes in a context of globalization is a complex task. It requires exploring the interaction between the different dimensions that generate spatial inequalities in schooling. Some of the factors inducing processes of school segregation are external to education systems and would require political action beyond education policy. Residential segregation, pockets of poverty in specific neighbourhoods, migration waves, demographic trends, forms of cultural closure and cultural emulation are potential factors that generate school segregation, which can only be addressed by implementing urban development policies, social policies or cultural actions to facilitate social integration. However, other causes of school segregation may be identified in the characteristics of education systems or in specific education policies that may favour polarization and an unbalanced distribution of underprivileged or highly privileged students. Early tracking and institutional differentiation, the presence of a large number of private schools and the capacity of schools to select their students are aspects that correlate with levels of school segregation (Alegre and Ferrer, 2010). In the same vein, models of school choice, the definition of catchment areas, levels of shared responsibilities to enrol at-­risk students and inspection

Introduction

5

systems to avoid student selection and fraud are factors that can be decisive in understanding how school segregation is produced and reproduced. As the case studies in this book show, the interaction between external and internal factors produces unique scenarios of school segregation, which engender inequalities of a different nature and intensity. Hence, policies to tackle school segregation must focus on different aspects, depending on the specific characteristics that this segregation adopts in different contexts. Reducing school segregation in some cities may require systems of student mobility, such as bussing in the USA. However, in some European cities characterized by lower levels of residential segregation, policies could focus on such aspects as reforming institutional differentiation or redefining school catchment areas. In Latin America, inequalities in the quality of schooling between public and private schools and high levels of urban segregation among the upper classes configure specific scenarios of extreme educational segregation between the rich and the poor. Reducing school segregation in Amsterdam requires different means than the ones needed to do the same in Santiago de Chile, Lima or Barcelona. The strongly contextual character of school segregation processes justifies the absence of universal and unique solutions to desegregate schooling. In fact, as we have learned from research, it may be the case that similar reforms produce contrasting outcomes when they are implemented in different contexts. The way in which policies are designed, the structural characteristics of social inequalities and the specificities of education systems are key aspects in understanding why even similar generic policies produce different effects in different contexts. For instance, while the 1988 Education Reform Act, which extended school choice in the UK, did not have an impact on school segregation (Gorard and Fitz, 1998; Taylor and Gorard, 2001), the universal voucher system in Chile and the expansion of the educational market significantly increased school segregation (Valenzuela et al., 2014). Likewise, while school mapping has shown some level of effectiveness in certain Spanish municipalities in mitigating school segregation (Bonal, 2012; Alegre et  al., 2010), in France, the carte scolaire has been ineffective in ensuring diversity in schools (Oberti, 2007a; Felouzis et al., this volume, p. 29). In any case, what this diversity of effects highlights is the importance of a specific set of conditions under which education policies take place. If similar policies have different effects, research should focus on the particular historical, social and institutional aspects of societies in order to understand the social mechanisms that trigger school segregation. Understanding why and how school segregation is produced and reproduced in different education systems requires disentangling the complexity of these social and educational factors, which the chapters included in this book explore. Research that is focused on the causal forces shaping school segregation has identified some regular patterns of interaction between social, cultural and educational factors, which trigger processes of educational exclusion, segregation and polarization. Both international comparison and individual case studies have focused on different factors in order to understand how social mechanisms cause school segregation. Despite the risk of oversimplifying the evidence, this section focuses on four main groups of factors: residential segregation and neighbourhood effects, institutional characteristics of education systems, the role of market reforms in education, and the

6

Understanding School Segregation

direct role of education policies regarding admission systems and compensatory policies. Residential segregation has been identified as one of the main causal forces of school segregation. However, the relationship between modes of residential segregation and school segregation is far from simple and linear. Firstly, we are confronted by a dual causal relationship. Residential segregation affects school segregation as much as differences in school quality impact on families’ residential patterns and choices, especially among the middle class (Frankenberg and Kotok, 2013; Boterman, 2013; Raveaud and van Zanten, 2007). Unequal schooling, middle-­class educational strategies and residential segregation interact locally in complex ways and significantly impact on educational attainment (Maloutas, 2007). In their search for social advantage, middle-­class families produce circuits of schooling in the educational marketplace, which reproduce spatial inequalities in school composition and academic performance (Ball et al., 1995). Secondly, residential segregation does not have one single and unique effect on the processes of school segregation and the educational performance of ethnic minorities. While a vast amount of research has provided evidence of the negative consequences of ethnic residential segregation and school segregation on the educational opportunities of students from minority groups (Rothstein, 2015; Gramberg, 1998; Andersson et al., 2010), some research has pointed out that ethnic density does not always has a negative effect on educational outcomes. The academic performance of youngsters from ethnic minorities especially depends on the level of integration of the respective ethnic minority in the host country, as well as the stability and quality of the neighbourhood. These variables explain why, for some ethnic minorities, their level of concentration may be more damaging than it is for others (Fleischmann et al., 2011). These disparities have been of particular interest to geographers, who have tried to assess whether there are specific neighbourhood effects on different social dimensions, with educational attainment being one of them. The existence of neighbourhood effects implies that there are aspects related to spatial characteristics, which impinge on social outcomes (attitudes, performance, opportunities) beyond the individual characteristics of residents (Friedrichs et  al., 2003). In education, the presence of neighbourhood effects would indicate that residential segregation affects education performance beyond school composition or school effects. While most research along these lines has shown that neighbourhood effects are less significant than institutional or peer effects (Del Bello et al., 2015; Sykes, 2011), some authors have pointed out the non-­linear and threshold aspects of neighbourhood effects, meaning that their impact is larger from specific levels of socio-­economic or ethnic concentration (Galster, 2014). Interestingly, the non-­linear aspect of neighbourhood effects underlines the fact that residential effects are stronger in the most affluent neighbourhoods (Duncan et  al., 1997; Helbig, 2010), suggesting that residential segregation could exacerbate the educational privileges of the most well-­off more than it would damage the opportunities of the most disadvantaged. In any case, the moderate role of neighbourhood effects is consistent with the fact, as confirmed by many studies, that school segregation is higher than residential segregation (Rangvid, 2007; Harris, 2017; Karsten et al., 2006). This demonstrates the

Introduction

7

importance of other factors beyond urban segregation in understanding how processes of spatial inequalities in schooling are produced and reproduced. Institutional differentiation is a decisive social mechanism of school segregation. International comparisons, based on the PISA, have highlighted a close relationship between early tracking (including different forms of horizontal stratification of schooling) and between-­school segregation (Murat, 2012; Alegre and Ferrer, 2010; Jenkins et al., 2008). Countries such as Austria, Belgium or Germany have higher levels of secondary school segregation than countries with comprehensive education systems (particularly Nordic countries). Social differences between tracks are greater than differences within tracks (Jenkins et  al., 2008), while socio-­economic background variables have a stronger effect on education inequalities in less comprehensive education systems (Duru-Bellat et al., 2004). The differences in educational and social value among tracks largely explain their different social composition and the higher levels of socio-­economic or ethnic segregation in less valued tracks. While the thesis formulated by Baudelot and Establet (1971) refers to the dual network of the French education system in the 1970s, it remains valid today for most societies, as proven by international comparisons. The evidence is so clear that the Organisation for Economic Co-­operation and Development has pointed out the negative effects of early tracking and low permeability between tracks concerning the educational opportunities of the most disadvantaged, while suggesting that upper secondary education student selection processes should be deferred (OECD, 2012). Academic and social segregation related to tracking increases educational inequity. Early tracking or high levels of within-­school ability grouping provides students with different learning environments, while having a negative impact on those students located in the lower tracks, along with any clear evidence of the benefits for high achievers and no significant effects on overall performance (Hanushek and Woessmann, 2006). Teacher and student expectations are also deeply affected by tracking, with low expectations for and from low performers and manifest consequences regarding their stigmatization. Lower tracks tend to receive fewer human and material resources, and are usually avoided by the best teachers (Oakes, 2005). Student trajectories are clearly marked by their prior allocation in lower tracks, with little opportunity to move to higher tracks or groups. Earlier allocation, therefore, inhibits the development of learning potentialities among many students (Blossfeld et al., 2016). A third crucial social mechanism concerning the reproduction of school segregation is related to the market-­oriented reforms that have spread globally in recent decades. These reforms have not necessarily enhanced direct privatization processes in education (although Chile and Peru, included in this volume, are examples of increasing direct private provision). Rather, they have promoted public–private partnerships and other measures emulating organizational and accountability systems of the private sector. Policy tools inspired by market reforms, such as voucher systems, charter schools, low-­fee private schools, free schools, higher levels of school choice, school-­ based management reforms and contracting out school services, have expanded as ‘good policies’ promoted by international organizations (Patrinos et  al., 2009) and national governments, with the support of private corporations (Robertson et  al., 2012).

8

Understanding School Segregation

As in the case of other social mechanisms, social and institutional contexts modify the effects of market reforms on school segregation. Policies enhancing school choice, for instance, may have different effects, depending on the level of residential segregation, but there will also be variations in the quality of schools and the extent to which the quasi-­market is regulated. The socio-­spatial characteristics of local education markets and parental choice strategies interact to produce heterogeneous effects across neighbourhoods and municipalities. A large number of studies confirms that the unequal choice opportunities of families with a high or low SES impact negatively on school segregation (see, among others, Gewirtz et  al., 1995; Denessen et  al., 2005; Easton, 2015; Bonal et al., 2017; Bellei et al., this volume, p. 221). Greater choice capacity facilitates processes of white flight and triggers processes of distinction with families looking for privileged schools. However, it is also the case that some international comparisons have found that segregation by parental occupation or country of birth is lower in countries with relatively little governmental control of schools and higher levels of choice (Gorard and Smith, 2004), and that the 1988 reform in the UK, which increased school choice, slightly reduced socio-­economic segregation (Gorard and Fitz, 2006), a finding that was also reported for Rio de Janeiro in Brazil (Bartholo and Da Costa, this volume, p. 65). These controversial results call for the need to unpack the relationship between choice and segregation in each specific context. As discussed, both catchment areas and free choice can facilitate school segregation, depending on the weight of other mediating factors within the complex relationship between school choice and segregation. There is less controversy, however, about the effects of diversity in educational supply on school segregation. Beyond the already-­mentioned relationship between the presence of private schools in education systems and school segregation (Alegre and Ferrer, 2010), there is other evidence confirming that the higher the diversity of schools, the stronger the mechanisms of social and academic student selection. This is true for free schools in the UK (Green et al., 2015) or Sweden (Bunar and Ambrose, 2016), for private independent and private subsidized schools in Chile (Valenzuela et al., 2014; Elacqua, 2012) or Spain (Bonal, 2012), for areas with a higher diversity of school supply in France (Oberti, 2007b), for low-­fee private schools in Peru (Balarin and Escudero, this volume, p. 179), for religious or pedagogical diversity in the Netherlands (Karsten et al., 2003), and for charter schools (Garcia, 2008) or magnet schools (Saporito, 2003) in the USA. Diversity in schooling stimulates processes of selection and self-­selection, as well as white flight, and can favour processes of cultural, religious or ethnic closure in specific schools. Of course, the mechanisms triggering school segregation work differently according to the modes of educational differentiation. However, despite the arguable benefits that a greater variety of curricular and pedagogic models can produce in terms of educational quality, there is no doubt that this diversity of school projects, accompanied by systems of enhancing choice, such as vouchers, tends to increase school segregation by SES or ethnic origin. Last but not least, educational policies that regulate admission systems and student allocation, as well as compensatory policies, also play an important role in triggering school segregation. Admission policies refer not only to the level of freedom in terms of school choice or the definition of catchment areas, but also to many other decisions

Introduction

9

regarding student distribution, such as classroom ratios, the number of places reserved for students with special needs or to students with a low SES, or decisions regarding the opening or closing of new classrooms in certain schools. Regulations on these aspects and everyday political decision-­making may either favour or reduce processes of school segregation. Research in this area has evidenced that reforms that provide schools with high levels of discretion in terms of student admissions, or that do not impede admission tests or certain forms of discriminatory school entrance requirements, increase the polarization of social factors, ethnicity or ability within school enrolment (Söderström and Uusitalo, 2010; Harris, 2012; Contreras et al., 2012; see also the chapters by Balarin and Escudero, and Krüger, in this volume). Active policies to ensure a balanced distribution of disadvantaged students can be decisive in terms of desegregation. Beyond the well-­known practice of bussing in the USA, other policies of affirmative action, such as reserving a number of places for students with certain characteristics, can result in a more equal distribution of at-­risk students. Despite these policies having been developed to a greater extent at the university level, in order to ensure places for minority students, some educational systems include similar measures for basic and post-­compulsory education. Reserving places for Roma children has been a common measure in some Eastern European countries (Miskovic, 2013; CHR, 2017). Meanwhile, in Spain, there is a legal obligation imposed on every school to reserve a minimum number of places for students from a low SES background or for late-­arrival migrants (Bonal, 2012). In Belgium’s French community, the reform of the admission system has enabled public authorities to intervene in order to ensure a more balanced distribution between privileged and underprivileged students (Dupriez et al., this volume). The implementation of such measures can be a decisive factor in reducing school segregation, as well as counterbalancing the effects of residential segregation or school competition. Finally, compensatory policies represent another mechanism, which can, somewhat paradoxically, contribute to school segregation. While compensatory policies can be powerful instruments with which to improve the learning conditions of children with social and academic disadvantage, they can also have unintended consequences for the unequal distribution of students. On the one hand, compensatory policies have a known stigmatization effect for certain schools, which may prevent middle-­class families from sending their children to them. As the analysis in this volume by Felouzis, Fouquet-Chauprade and Charmillot demonstrates, concerning the case of the Zones d’Action Prioritaire in France, these policies are designed to compensate for the effects of school segregation, rather than to prevent it. Nevertheless, despite gaining access to supplementary human or material resources, or despite focusing their pedagogic action on singular projects, such in the case of magnet schools, these types of measures may polarize educational demand between disadvantaged and privileged students. On many occasions, there is a trade-­off between targeted education policies, which are focused on the needs of specific groups or territories, and other policies, such as defining catchment areas or reserving places for students with special needs, thereby attempting to balance the distribution of vulnerable students among all schools. On the other hand, the specific design of compensatory policies may introduce biases that exacerbate school polarization among social groups. As Gorard and Siddiqui, in this

10

Understanding School Segregation

volume, show in the case of the UK, territorial focalization (and ignorance of the social composition of schools) may lead to failed systems of compensation and higher school segregation. Likewise, policies designed to facilitate access for poor children to supposedly high-­quality schools may, in fact, increase academic and social selection and aggravate school segregation. This occurred in the UK, on account of the Assisted Places Scheme policy (Easton, 2015) and the expansion of places in selective grammar schools (Gorard and Seddiqui, this volume), and in Colombia, after the launch of a programme of charter schools (Termes et al., 2016).

3.  Desegregation policies: political, cultural and institutional limits to desegregation In this section, based on the country cases included in the book, we analyse the policies and regulations intended to inhibit, control or reduce levels of school segregation. In order to understand the link between policies and the causal factors in the production of school segregation, we associate those policies with the educational context in which they have been implemented. Therefore, we have organized this section by distinguishing between traditional and market-­based school segregation dynamics, while discussing different educational policies aimed at tackling segregation, which are more typical within both frameworks. Meanwhile, the final part of the section discusses the more general issue of observably low priority assigned to desegregation policies on the educational policy agenda of the studied countries.

3.1  Policies tackling segregation in traditional settings Traditional forms of segregation within the public system were based on institutional segmentation: different kinds of curricula and educational aims in respect of different student populations. They represented the established way of channelling increasing social and academic diversity caused by expansion processes; in some way, they were the price to pay for the maintenance of the middle classes when incorporating the lower classes into the system. Once segregation trends were identified, a repertoire of more or less conventional policies was developed to tackle school segregation. However, the ultimate effectiveness of these policies has been questioned, given that in the medium term, privileged groups have always found ways to avoid undesirable schools.

Curriculum differentiation Many educational systems have channelled the process of schooling through differentiated forms of curricula, producing a segmentation that almost inevitably reflects a hierarchy. The most widespread example of this is the organization of a vocational channel oriented towards work as an alternative to general secondary education, which is clearly oriented towards higher education. Since the publics of both streams tend to be socially differentiated, this socio-­educational segmentation is

Introduction

11

reflected in the social composition of the institutions. Several countries, such as France and Belgium, have recognized this segmentation as inequitable and tried to modify it, either by extending the common curriculum or by delaying the moment of separation between the two tracks. A non-­segmented curriculum would make the difference in educational opportunities between institutions less evident by favouring the homogenization of their students. However, this is rarely achieved. Another form of curricular differentiation is the one based on the level of academic requirements or the performance of the schools, thereby creating academically selective public schools. In this case, given that it is recognized that the orientation of the curriculum is the same as for other public schools, the hierarchy of prestige or quality between the two is even more evident. Given the multiple ways in which the performance of students is related to differences in class and ethnicity, the consequence is that these schools also contribute to educational segregation. This produces inequality in relation to non-­selective schools, which governments have attempted to overcome by developing compensatory policies (such as Zones d’Action Prioritaire in France or targeted budgets in the UK). However, the social context makes it very difficult to reverse.

Within-­school tracking While segmented channels offer differentiated curricula between schools, another form of differentiation has been the organization according to performance level within the same institution, which is associated with different forms of internal differentiation, as well as with curricular options in the form of elective courses. Tracking and other forms of internal differentiation lack any univocal interpretation. For some, this constitutes another manifestation of school segregation (in this case, between classrooms in the same school), since the different tracks are associated with class, race and ethnicity, thus reproducing inequalities, as the US case shows. Therefore, the elimination of these practices is promoted as a key part of desegregation policies. However, allowing and even promoting some degree of internal differentiation have also been interpreted as the result of maintaining some degree of social heterogeneity within public schools. Typically, the middle classes will press for this type of differentiation in order to generate more opportunities for their children within the public system. The alternative is a system in which the middle class migrates to private schools. Thus, there is a trade-­off between segregation among and within schools. In the latter case, internal differentiation would ensure heterogeneity within schools at the cost of facilitating differentiated educational experiences. Certainly, these types of practices are rarely formulated as ‘social integration policies’, but rather as local socio-­ educational dynamics, spontaneous from the point of view of policies, but effective as educational micropolitics.

School zoning and unregulated admission processes School zoning has been implemented in many countries, although with important operational differences, in terms of coercive tactics for enrolment and the control of

12

Understanding School Segregation

the school offer. From a general perspective, by linking the social composition of schools with that of their urban environment, zoning reproduces residential segregation in schools. But the details in the zoning design matter when modulating the translation of residential segregation into school segregation, while policymakers have tried different alternatives, as the Spanish case clearly shows, by modifying the size of catchment areas. Likewise, it is important to decide whether admission decisions are administered centrally or at the school level, which relates to the degree to which families can choose the school within a zone; in the latter case, a quasi-­market situation can be set up, as we shall analyse below. In this regard, it is interesting to note that, since school segregation in many countries is greater than residential segregation, zoning has been proposed as a way to increase social heterogeneity within schools. However, in some cases studied in this book, zoning becomes a segregating factor in itself. An additional reason for school segregation is that not all parents respect the allocation made by public authorities; typically, the most educated and those with greater resources can influence these decisions, as shown in the cases of Peru and Argentina. In France, given the inadequate functioning of zoning in some contexts, school choice beyond the reference area was introduced as a solution, which apparently did not have the expected positive effect of higher social mixing. This is consistent with the evidence associated with the negative effects of school choice on segregation. More generally, in several public education systems, the admission practices of schools are not particularly transparent and easily lend themselves to arbitrariness. This is due to the fact that, on the one hand, schools try to select less problematic students or avoid minority students and those who are disadvantaged and hinder the teaching process or the school’s performance, while, on the other hand, families carry out different practices to influence the allocation process of school places. Several countries have introduced measures to make these systems more transparent, non-­ discriminatory and even progressive, but experience suggests unequal effects in this respect.

3.2  Desegregation policies in a market-­oriented context Beyond ‘traditional’ forms of segregation, based on fairly widespread characteristics of public education systems, there are forms of segregation associated with market or quasi-­ market dynamics, which are identified as producers of school segregation in several studies. Certainly, these dynamics can be more or less transparent and, although they are mainly associated with the private sector, they also occur within public education systems. In such cases, policies have tried to lessen the segregating effects of the school market, while maintaining its organizing principles, especially the choice of schools by families, thereby reducing the potential for these policies to reduce inequalities.

Privatization and school autonomy In most of the cases presented, private schools are considered to be a relevant factor in the production of school segregation. This happens because, unlike public schools, which

Introduction

13

are initially open to the whole community, private schools tend to specialize in subpopulations, which sometimes constitute particular communities. In systems that are historically based on private provision in order to guarantee pluralism, such as in the Netherlands or Belgium, these involve religious, ethnic or philosophical communities; in other countries, private education (which is generally more restricted and expensive) tends to be associated with upper and sometimes middle social classes, except in the Chilean case, where privatization under an open market regime has produced an enormously specialized offer according to the level of income of families throughout the entire range of social stratification. Certainly, the characteristics of the operation of private schools that produce segregation are also replicated in some cases within the public system, as in the case of Peru or Argentina; that is, when schools have the autonomy to differentiate their offer to diverse audiences, they impose admission requirements on students and families, thereby controlling the overall admission processes. Despite this evidence, in general, countries have not implemented policies to reverse privatization tendencies; on the contrary, in some cases, these policies have been openly promoted. Policies aimed at reducing the segregating and unequal effects of privatization have been rather indirect. For example, in Belgium, compulsory educational standards have been established for all schools in an effort to reduce the heterogeneity of the offer. Yet the most widespread policy in this regard has been the regulation of school admission processes to avoid arbitrary discrimination, make the processes more transparent and guarantee freedom of choice for families in a more productive way. However, as the Chilean and Belgian experiences show, even these fairly moderate policies to tackle segregation have faced enormous resistance from private schools, which feel threatened, given that their educational projects are based on homogeneous communities.

School choice Freedom of school choice for parents is a key mechanism found in school systems organized with a market logic. As we have seen, in traditional systems based on public education, some parents also find informal or indirect means to influence where their children will study; however, in the logic of the market, all parents are expected to be active choosers looking for the school that best suits their preferences. Given that ‘preferences’ are associated with the cultural, social and economic characteristics of families, school choice is usually recognized as a factor that produces school segregation. Some countries that allow school choice have tried to reduce their effects on segregation: on the one hand, by trying to equalize the conditions under which parents of different social classes or ethnic groups choose a school (as in Belgium, Chile or Brazil), by means of information campaigns and advertising indicators of school performance, or by reducing the costs associated with schooling, such as transportation and school fees; and on the other hand, by partially diminishing the freedom of parents to choose by organizing some form of ‘controlled school choice’ (as in some cities of the Netherlands, the USA or Spain). This is done by only allowing choice within a specific territory or by giving power to local authorities to maintain partial control over the admission process. For instance, some countries have introduced regulations to redistribute

14

Understanding School Segregation

students with higher learning difficulties among local schools in cases where the dynamics of choice produce a very high concentration of them in a small number of schools. It is interesting to note that, in the opposite direction, in some systems (such as in France or Rio de Janeiro in Brazil), authorities have begun to implement school choice policies as a way to combat the segregation produced by school zones, provide greater mobility to disadvantaged families and facilitate their integration into more diverse schools. The evidence provided by these examples, although limited so far, suggests that they have not significantly reduced school segregation; thus, broadly speaking, it coincides with other evidence available in other case studies in this book (the Netherlands, Chile and some districts in the USA), which reveals that parents’ school choice is not part of the solution to school segregation; rather it can actually be one of its causes.

Competition Finally, market dynamics are reflected in the competition between schools to attract specific families and thus obtain advantages over others, typically in the form of resources, but also prestige and better performance. Given the relationship between the characteristics of families and school choice mentioned above, competition between schools also tends to reinforce segregation by increasing the number of ‘desirable’ students from more advantaged families and, especially, highly ‘preventable’ students from lower social classes, ethnic minorities or migrant backgrounds. Of course, traditional systems also produce prestigious hierarchies among schools, but the dynamics of segregation are less intensive because there is reduced arbitrariness in the admission policies, families exercise less freedom of choice and schools do not benefit economically (at least formally) from these differences in prestige. In fact, the market situation exacerbates all these processes. In addition to some of the previously mentioned policies that control school autonomy and parental choice, some countries have implemented policies that are focused on reducing the effects of competition between schools. These policies include increasing the resources associated with ‘at-­ risk’ students (e.g. immigrants or low-­income families) in per-­capita financing schemes, giving preference to disadvantaged students in the admission process when there is excess of demand, avoiding the publication of performance rankings that may stigmatize certain schools, or modifying the geographical area within which schools compete for students.

3.3  The lack of priority concerning desegregation policies The previous analysis focused on policies that, with varying but generally low intensity, address some of the causes that create school segregation and therefore try to prevent it. Other policies, on the other hand, simply try to mitigate some of their effects and cannot therefore be considered as policies that tackle school segregation. For example, several countries (such as the USA or Chile) have implemented school improvement programmes targeting poorly performing schools with a high concentration of at-­risk students, while others allocate compensatory funds to schools where the most

Introduction

15

disadvantaged students attend (as in the UK) or to urban areas where such students mostly live (such as in France or Belgium), so that local or school authorities can provide better conditions and better inputs for their teachers and students. Although the educational effects of these policies are varied – some of them have been quite successful – their effectiveness with regard to school segregation has been questioned. In some cases (such as the Zones d’Action Prioritaire in France), undesired effects have been documented, such as the stigmatization of schools or areas ‘benefiting’ from these policies. Paradoxically, these policies may end up reinforcing the level of segregation and the consequences that they were intended to alleviate. The review of international experiences also allows for the identification of another pattern. There is, in general, an absence of public policies directly aimed at reducing school segregation once this has occurred. The well-­documented US experience involving direct policies of racial desegregation continues to be the most resolute instance of policies of this kind. In most of the studied countries, school segregation has not been defined as a matter of educational policies and, at most, political action has been limited in order to partially alleviate its consequences. Furthermore, in some cases in which school segregation has been identified as a problem to be solved and become part of the policy agenda, the measures that have been implemented either try only to address its causes indirectly (such as increasing the voucher value for low-­ income students in Chile) or may be qualified as ‘voluntary’ desegregation policies, based, for example, on school choice mechanisms (as in the case of France and the Brazilian city of Rio de Janeiro). This relative lack of priority to tackle school segregation in the context of educational policies is in sharp contrast to the increasing amount of available scientific evidence, including that from the analysed cases, in terms of their relevance, magnitude and consequences. The question then is how to understand this apparent distance between the academic relevance given to school segregation and the low priority given to this problem by decision makers or even activists. Firstly, as several authors point out in their studies, the evidence regarding segregation in education has been presented relatively recently in almost all countries (with the exception of the USA, although in this case, the racial segregation of schools was not an undiscovered phenomenon, but an institutional feature of the US educational system). In Europe and Latin America, school segregation was, at best, subsumed under more general themes about inequity in education, rather than the object of direct political discussion. In addition, available evidence for its negative consequences remains scarce outside of the USA, which clearly reduces the pressure on governments to take action. Likewise, the priority agenda for governments is always limited. Policies that dominate current international debates in education, such as privatization, school autonomy and test-­based accountability, are actually potential factors contributing to segregation. Thus, far from giving priority to desegregation policies, governments are attracted to global education policies that are potentially inconsistent with a desegregation strategy. As such, there is a clear incompatibility between potentially opposing agendas. So far, it is quite clear which agenda is winning this dispute. Secondly, even in contexts in which policymakers have become aware of the relevance of segregation and are interested in advancing policies to combat it, it is not

16

Understanding School Segregation

clear what tools they have at their disposal. Although the evidence is quite convergent concerning some of its causes, the same cannot be said with regard to policies to combat it. Indeed, it is possible to state that there is no consensus about some of the basic measures that have been proposed and analysed in this book. For example, some governments promote vouchers and school choice as ways to increase families’ mobility and allow them to self-­desegregate, while others promote zoning to decrease the dynamics of the school market or pursue compensatory policies that target specific groups of students. Somewhat paradoxically, the same desegregation measures used in one specific context have been identified as factors that produce school segregation in other contexts. Likewise, the US experience highlights the complexity in implementing policies of forced desegregation, since, in the medium term, some of its effects seem to be reversed. Forced desegregation may carry the price of an eventual abandonment of the public education system by the middle classes, as has happened in some countries. Certainly, context matters when it comes to assessing the appropriateness of one policy or another. Given the sensitivity around this issue, social actors have reacted to policies in ways that are not always predictable, which can ultimately inhibit the effects of specific policies, as well as generate new negative effects. In short, the path of desegregation policies is highly uncertain. Finally, the political economy of desegregation policies is tremendously complex and potentially very conflictive. Desegregation policies are concerned with controlling institutional dynamics, which are deeply rooted in both the public and the private sectors. They also attempt to modify sociocultural practices of difficult and slow change, as well as impinge upon the social interests of the upper and middle classes to the benefit of the lower classes, minority ethnic groups and migrants – that is, those sectors with less voice and power in society. The few experiences of direct desegregation policies, and even some indirect ones, have shown the extent to which they can be explosive in terms of social conflict. Thus, it is not surprising that decision makers have mainly chosen to pursue only policies that deal with the effects of segregation or, at most, do not help to increase it. This is certainly a first step prior to openly promoting the policies of integration and diversity in schools. To advance decisively in this direction, it is clear that most of the countries analysed have not yet arrived at a social consensus on the value of education in diverse contexts; in short, this is a pending social debate. Finally, the complexity of the causes of segregation, which certainly include the weight of urban segregation and new forms of exclusion based on immigration, in which social class mixes with issues of language and race more than before, demands an approach that goes beyond educational policies. To tackle those multidimensional causes requires a cross-­sectorial approach, which is generally absent from these public policies.

4.  International perspectives on school segregation: structure and summary of the book Overall, the different country cases included in this book provide a comprehensive review of the current debates on school segregation, demonstrating the dynamism of

Introduction

17

this academic field and highlighting the complexities involved in properly identifying both the causes and the consequences of school segregation. More importantly, they provide strong grounds to consider school segregation, according to social class, race and ethnicity, as a highly relevant problem, which should be a key priority on educational policy agendas. Finally, the analysed experiences also provide information with which to prevent simplistic approaches to this issue, since its complete resolution depends not only on institutional changes, but also on cultural and social support, which should be ultimately based on an equity-­oriented vision of educational policy. Part I,‘Patterns of school segregation and social inequalities’, contains three chapters, which clearly show how school segregation should be understood in its widest context, while linking it to the educational institution and also the society and the history that has moulded it. The French case is paradigmatic in this sense. Although formally committed to meritocratic ideals, such as equality of opportunity and common schooling, Felouzis, Fouquet-Chauprade and Charmillot demonstrate how the policies adopted in France, with the very aim of realizing these ideals in the context of educational democratization, have not produced higher educational equality. Even more, they have contributed to the consolidation of, and increase in, social and ethnic segregation within the French school system. The explanation of this paradox stems from the educational policies that have a sociocultural dimension and produce reactions from different social actors, whose responses and strategic actions can cause undesirable effects that are hard to anticipate. Thus, despite the fact that the collège unique has tried to homogenize the school experience, diverse ways of school competition and selection have maintained certain levels of differentiation. Likewise, although school mapping allows for diversity in schools, this is limited by urban segregation and the avoidance strategies of families with higher cultural and economic capital. Finally, although priority education zones are intended to reduce the effects of segregation, they often end up reinforcing it via the stigmatization produced by targeted policies. Thus, social inequalities have found certain interstices to be transformed into school segregation. The French-­speaking Belgian system provides a good counterpoint. According to the analysis of Dupriez, Barbana and Verhoeven, the institutional differentiation of educational provision, which was linked to the existence of socially segregated populations, occurred at an early stage. This produced an educational system, which, through its mission, failed to prioritize integration into a common project and reinforced a culture of anti-­heterogeneity in the school. Instead of a common educational system, Belgian society has recognized the value of preserving diverse school networks based on different sociological communities (historically, Catholics and liberals), thereby promoting the freedom of teaching, school choice and private provision. This sociocultural segregation, explicitly imported into the educational system, has generated the conditions necessary for the emergence of a quasi-­market, which has slowly detached itself from its origins in a context of religious tolerance and is now based on academic and social inequalities. Although, recently, there has been awareness of the serious educational segregation that the system produces, the authors claim that the historical and social roots are so deep, the weak egalitarian policies that have been implemented are not enough to overcome the prevalent resistance.

18

Understanding School Segregation

Finally, the situation of Brazil illustrates very well the difficulties faced by educational systems, which have recently expanded, in facilitating social integration, particularly in the context of vastly unequal societies. Bartholo and Da Costa have shown how the existence of a private schools sector channels the segregation of the upper and middle classes according to a market logic, while also allowing for another segment of the private sector focused on the lower social classes. Additionally, they emphasize that school segregation processes also occur within the public system, albeit in less explicit ways (for instance, entry selection, which is associated with different types of school prestige). The authors stress that these dynamics interact with the socio-­demographic features of cities, adding to the social class, and differences in skin colour, thus producing a triple segregation: social, racial and academic. Interestingly, these complex forces seem to annul the potential positive effects on social integration of a recent policy involving the random assignment of students to Rio de Janeiro schools, possibly because certain social actors (in this case, middle-­class families and the local administration) rapidly learn how to take advantage or ‘get around’ the new regulations. Part II of the book, ‘School segregation and student performance’, delves into the additional educational inequity that school segregation means for disadvantaged students, providing empirical arguments about several negative effects. To start, Gorard and Siddiqui demonstrate the degree of complexity involved in the definition of disadvantaged students and how compensatory educational policies should identify them in order to be effective: that is to say, even in contexts where social segregation in schools has been recognized as a relevant problem and policies have been designed to address these issues, the distance before success is achieved remains too great. Basically, the authors argue that relevant educational disadvantage accumulates over time, and should thus be measured longitudinally in order to attend to the poverty trajectory experienced by students and their level of segregation. Based on their own estimations, the authors demonstrate that, when measuring poverty longitudinally and estimating school segregation accordingly, it is possible to explain a significant part of the lower educational achievement of disadvantaged students. These factors are more relevant than other variables, such as geographical area or the selective nature of grammar schools, which are commonly mentioned in political debates in England. Natalia Krüger, for her part, argues that, in Argentina, the recent expansion of secondary education (traditionally, an educational level that was socially and academically selective) has refined the internal segregation mechanisms of the educational system in both the private sector (which operates with market rules and produces explicit segregation) and the public sector (where wealthier families take advantage of the weakness of the state institutions to obtain advantages by informal and opaque methods). Thus, despite Argentina being recognized as a society with fewer inequalities than other Latin American countries, the estimations show that the level of social segregation in education is significantly higher than the average of OECD countries. The author concludes with the argument about how school socio-­ economic segregation is a relevant factor when explaining the mathematics performance of Argentinian students (she estimates a standardized coefficient of a quarter of a standard deviation, which is associated with the effect of the socio-­ economic composition of schools on the PISA mathematics score at the student level),

Introduction

19

which is higher than other studied factors, such as educational resources, curriculum structure or school competition. This section ends with a contribution from Roslyn Mickelson, who summarizes the impressive evidence accumulated in the social sciences concerning the effects of ethnic, racial and socio-­economic compositions of schools in the USA, where educational segregation has been studied more than in any other country. Mickelson organizes her findings in a model, which shows how education in schools with a diverse intake can contribute to the construction of a society that is fairer, multi-­ethnic and democratic. Firstly, the evidence presented sets out the generalized academic benefits (in terms of mathematics, language and, to a more moderate extent, science performance, as well as access and graduation from college) for students who attend schools/classrooms that are diverse in their composition, especially those belonging to disadvantaged groups. Secondly, evidence is presented for the non-­academic benefits of education in diverse schools, e.g. less prejudice and fear, fewer racial stereotypes, an increase in confidence, respect and intergroup acceptance, and more inter-­racial friendships and relationships. In addition, the author presents evidence of the medium- and long-­term effects on the civic commitment of young people (e.g. volunteer work, enrolment with community groups, and different forms of political participation), and also of adults, including work setting, social relations and community life. The irony of the US case is that, after the success of racial desegregation policies in the 1960s and 1970s, a resegregation process took place, driven by demographic and urban changes, as well as by educational policies and practices: that is, factors that have proven to be highly elusive with regard to the increasingly weak policies of racial, ethnic, social and academic integration in schools. The four cases included in Part III, ‘Market dynamics and school segregation’, directly address certain issues, which, despite having different centralities, have also been present in some of the other studied countries. This prompts the question about the extent to which school segregation is produced or increased by the (more or less explicit) market dynamics of educational systems. The analysis of the Peruvian case by Balarín and Escudero is noteworthy, as it confirms the functioning of a de facto market, which has neither been created nor regulated by educational policies; indeed, it is characterized by the authors as an ‘un-­governed school market’. In other words, a dual situation exists in Peru. The private sector (which has increased in size, fostered by policies that promote privatization in recent decades) operates as a free market, regulated by price (including an important number of low-­fee schools) and without state control. The authors consider privatization and price differentiation as explanatory factors in the noticeable increase in socio-­economic segregation of education in Peru during the last decade. However, in the same period, segregation also increased within the public education sector, which is explained by families’ school choice, discrimination processes in the admission to schools and economic contributions (formally voluntary) charged to families; over time, these dynamics are reinforced by the spatial segregation of cities. According to the authors, these practices reflect the presence of a market within the public sector, which is highly opaque and unequal. The Spanish case, as analysed by Bonal and Zancajo, is different. There is a consolidated public/private dual system, in which the state subsidizes private schools

20

Understanding School Segregation

(even religious schools), respects freedom of school choice and conducts a light-­touch inspection of school admission processes, in both the private and the public sectors (despite regulations that formally forbid the discrimination of students). According to the authors, this educational quasi-­market produces segregation based on the social class of families, residential segregation, parents’ preferences, admission policies and explicit or covered fees (in principle, schools should be free). This scenario has become more complex with the increase in the number of immigrants, which reinforces residential, as well as educational, segregation between public and private schools. Some middle-­class families leave the public system to avoid immigrants and students who are at-­risk in social and educational terms. Finally, the authors emphasize how these general dynamics are very hard to govern because there is a clear absence of national policies to prevent school segregation and because key decisions regarding school admission processes and the control of student selection depend on municipal governments. This produces a highly differentiated system locally, in which some municipalities develop desegregation measures, while others tend to preserve the interests of the most affluent and more pressuring groups. The Dutch experience, together with the Belgian one, is well known as a highly established system, which privileges freedom of teaching and school choice among parents. As Boterman explains in his chapter, by attempting to guarantee religious and pedagogical pluralism since the beginning of the twentieth century, public and private (the majority) schools have received the same treatment, while the right of parents to choose a school is guaranteed by the constitution. Thus, in a system that is openly segregated by ethnicity and religion, segregation by social class and place of residence appears to be less important. However, according to Boterman, this historical conceptualization has started to change. On the one hand, increasing the residential segregation of immigrants and creating schools for minority religious groups have exacerbated the ethnic-­religious segregation of schools; on the other hand, the evidence suggests that an important proportion of families, especially from the middle class, use social criteria when choosing a school by looking for a homogeneous ethnic-­class composition. According to the author, this would explain his findings, which show that schools are equally or more segregated than cities, despite the expectation that free school choice should separate school segregation from urban segregation by offering greater mobility to families. Finally, the Chilean educational system, as analysed by Bellei, Contreras, Canales and Orellana, represents an extreme case of a consolidated educational open market, which operates with high levels of deregulation. In Chile, public and private schools (with or without profit motives, including religious schools) openly compete to obtain state funding, which is allocated via a common voucher system. While families can choose from among all schools, the schools’ admission processes are discriminatory and subject to no particular control by the state. Private schools are allowed to charge families without losing access to the state voucher. There are also private schools that are very expensive and without any school subsidy. The authors argue that these market dynamics (which have in turn accelerated privatization) explain the high levels of socio-­economic segregation in the Chilean education system. In a society with extreme levels of social inequalities and a weak public education system, the authors show how

Introduction

21

families base their school choice on a social rationality: whereas the upper and middle classes enclose themselves in socioculturally homogeneous environments, the lower-­ middle class fearfully rejects public, free and non-­selective schools because of their association with the lower class, which is considered to have risky behaviours. This in turn fosters an increase in school segregation ‘from the demand side’. The book concludes with an interview with Professor Gary Orfield. Drawing on his copious academic and activist experiences, he provides a complete state-­of-the-­art description of the field and highlights current trends in school segregation in the USA and globally. Current changes in the racial composition of societies, residential segregation and education policies themselves have produced new tendencies in school segregation, which will make this problem even more difficult to tackle. Professor Orfield identifies elements that can help us to understand regional differences in school segregation dynamics and the challenges that education policies will face in the future. The role of research is crucial in providing evidence of the damaging effects of school segregation, a strategy that Professor Orfield has developed over many years in the context of the Civil Rights Project.

References Alegre, M. A. and Ferrer, G. (2010), ‘School Regimes and Education Equity. Some Insights Based on PISA 2006’, British Educational Research Journal, 36 (3): 433–62. Alegre, M. A., Benito, R. and González, I. (2010), ‘Measures and Determinants of Student Body Socioeconomic Diversity: Evidence from Spain’, Journal of School Choice, 4 (1): 23–46. Andersen, S. C. and Thomsen, M. T. (2011), ‘Policy Implications of Limiting Immigrant Concentration in Danish Public Schools’, Scandinavian Political Studies, 34 (1): 27–52. Andersson, E., Östh, J. and Malmberg, B. (2010), ‘Ethnic Segregation and Performance Inequality in the Swedish School System: A Regional Perspective’, Environment and Planning A., 42 (11), 2674–86. Arabadjieva, K. (2015), ‘Challenging the School Segregation of Roma Children in Central and Eastern Europe’, International Journal of Human Rights, 20 (1): 33–54. Atkinson, A. (2015), Inequality. What Can Be Done? Cambridge, MA: Harvard University Press. Ball, S., Bowe, R. and Gewirtz, S. (1995), ‘Circuits of Schooling: A Sociological Exploration of Parental Choice in Social Class Contexts’, The Sociological Review, 43 (1): 52–78. Baudelot, C. and Establet, R. (1971), L’école capitaliste en France. Paris: Editions Maspero. Benito, R., Alegre, M. A. and Gonzàlez, I. (2014), ‘School Segregation and its Effects on Educational Equality and Efficiency in 16 OECD Comprehensive School Systems’, Comparative Education Review, 58 (1): 104–34. Blossfeld, H. P., Buchholz, S., Skopek, J. and Triventi, M. (eds.) (2016), Models of Secondary Education and Social Inequality: An International Comparison. Cheltenham: Edward Elgar. Bonal, X. (2012), ‘Education Policy and School Segregation of Migrant Students in Catalonia: The Politics of Non-­decision-making’, Journal of Education Policy, 27 (3): 401–21. Bonal, X., Zancajo, A. and Verger, A. (2017), ‘Making Poor Choices? Demand Rationalities and School Choice in a Chilean Local Education Market’, Journal of School Choice, 11 (2), 258–81.

22

Understanding School Segregation

Boterman, W. R. (2013), ‘Dealing with Diversity. Middle-­class Family Households and the Issue of “Black” and “White” Schools in Amsterdam’, Urban Studies, 50 (6): 1130–47. Bunar, N. and Ambrose, A. (2016), ‘Schools, Choice and Reputation: Local School Markets and the Distribution of Symbolic Capital in Segregated Cities’, Research in Comparative and International Education, 11 (1): 34–51. Cashman, L. (2016), ‘New Label No Progress: Institutional Racism and the Persistent Segregation of Romani Students in the Czech Republic’, Race Ethnicity and Education, 20 (5): 595–608. Chubb, J. E. and Moe, T. M. (1990), Politics, Markets, and America’s Schools. Washington, DC: Brookings Institution Press. Commissioner of Human Rights (2017), Fighting School Segregation in Europe through Inclusive Education: A Position Paper. Council of Europe. Contreras, D., Sepúlveda, P. and Bustos, S. (2012), ‘When Schools Are the Ones that Choose: The Effects of Screening in Chile’, Social Science Quarterly, 91 (5): 1349–68. Del Bello, C. L., Zenou, Y. and Patacchini, E. (2015), Neighborhood Effects in Education Neighborhood Effects in Education. IZA Discussion Paper No. 8956. Institute for the Study of Labour, Bonn. Denessen, E., Driessen, G. and Sleegers, P. (2005), ‘Segregation by Choice? A Study of Group-­specific Reasons for School Choice’, Journal of Education Policy, 20 (3): 347–68. Dronkers, J. and Robert, P. (2008), ‘School Choice in the Light of the Effectiveness Differences of Various Types of Public and Private Schools in 19 OECD Countries’, Journal of School Choice, 2 (3): 260–301. Dumay, X. and Dupriez, V. (2008), ‘Does the School Composition Effect Matter? Evidence from Belgian Data’, British Journal of Educational Studies, 56 (4): 440–7. Dupriez, V., Dumay, X. and Vause, A. (2008), ‘How Do School Systems Manage Pupils’ Heterogeneity?’, Comparative Education Review, 52 (2): 245–73. Duru-Bellat, M., Mons, N. and Suchaut, B. (2004), Caractéristiques des systèmes éducatifs et compétences des jeunes de 15 ans. L’éclairage des comparaisons entre pays. Les Cahiers de l’IREDU (Bourgogne, IREDU, CNRS-Université de Bourgogne). Easton, S. (2015), ‘Featured Graphic. Parental “Choice” and Sociospatial Segregation in a UK City: Access to the Best State-­funded Secondary Schools for Black Somali Pupils’, Environment and Planning A, 47 (10): 2021–2. Elacqua, G. (2012), ‘The Impact of School Choice and Public Policy on Segregation: Evidence from Chile’, International Journal of Educational Development, 32 (3): 444–53. Fleischmann, F., Phalet, K., Neels, K. and Deboosere, P. (2011), ‘Contextualizing Ethnic Educational Inequality: The Role of Stability and Quality of Neighborhoods and Ethnic Density in Second-­generation Attainment’, International Migration Review, 45 (2): 386–425. Frankenberg, E. and Kotok, S. (2013), ‘Demography and Educational Politics in the Suburban Marketplace’, Peabody Journal of Education, 88 (1): 112–26. Frankenberg, E. and Orfield, G. (2012), The Resegregation of Suburban Schools A Hidden Crisis in American Education. Cambridge, MA: Harvard Education Press. Friedrichs, J., Galster, G. and Musterd, S. (2003), ‘Neighbourhood Effects on Social Opportunities: The European and American Research and Policy Context’, Housing Studies, 18 (6): 797–806. Galster, G. (2014), ‘Nonlinear and Threshold Aspects of Neighborhood Effects’, Köln Z Soziol (Suppl) 66: 117–33. Garcia, D. (2008), ‘The Impact of School Choice on Racial Segregation in Charter Schools’, Educational Policy 22 (6): 805–29.

Introduction

23

Gorard, S. and Fitz, J. (1998), ‘The More Things Change. . . . The Missing Impact of Marketization’, British Journal of Sociology of Education, 19 (3): 365–76. Gorard, S. and Fitz, J. (2006), ‘What Counts as Evidence in the School Choice Debate?’, British Educational Research Journal, 32 (6), 797–816. Gorard, S. and Smith, E. (2004), ‘An International Comparison of Equity in Education Systems’, Comparative Education, 40 (1): 15–28. Gramberg, P. (1998), ‘School Segregation: The Case of Amsterdam’, Urban Studies, 35 (3): 547–64. Green, F., Allen, R. and Jenkins, A. (2015), ‘Are English Free Schools Socially Selective? A Quantitative Analysis’, British Educational Research Journal, 41 (6): 907–24. Hanushek, E. and Woessmann, L. (2006), ‘Does Educational Tracking Affect Performance and Inequality? Differences-­in-Differences Evidence across Countries’, Economic Journal, 116: 63–76. Hanushek, E. A., Kain, J. F., Markman, J. M. and Rivkin, S. G. (2003), ‘Does Peer Ability Affect Student Achievement?’, Journal of Applied Econometrics, 18 (5): 27–44. Hanushek, E. A., Kain, J. F. and Rivkin, S. G. (2002), ‘New Evidence about Brown v. Board of Education: The Complex Effects of School Racial Composition on Achievement’, NBER Working Paper, 8741, National Bureau of Economic Research. Harris, R. (2012), ‘Local Indices of Segregation with Application to Social Segregation between London’s Secondary Schools, 2003–08/09’. Environment and Planning A., 44 (3): 669–87. Harris, R. (2017), Measuring the Scales of Segregation: Looking at the Residential Separation of White British and Other Schoolchildren in England Using a Multilevel Index of Dissimilarity. Transactions of the Institute of British Geographers. School of Geographical Sciences University of Bristol. Helbig, M. (2010), ‘Neighbourhood Does Matter! Socio-­structural Neighbourhood Characteristics and Educational Success’, Kolner Zeitschrift Fur Soziologie Und Sozialpsychologie, 62 (4), 655–79. Hughes, J., Campbell, A., Lolliot, S., Hewstone, M. and Gallagher, T. (2013), ‘Inter-Group Contact at School and Social Attitudes: Evidence from Northern Ireland’ Oxford Review of Education, 39 (6): 761–79. Jenkins, S. P., Micklewright, J. and Schnepf, S. V. (2008), ‘Social Segregation in Secondary Schools: How Does England Compare with Other Countries?’, Oxford Review of Education, 34 (1): 21–37. Karsten, S., Felix, C., Ledoux, G., Meijnen, W., Roeleveld, J. and Van Schooten, E. (2006), ‘Choosing Segregation or Integration? The Extent and Effects of Ethnic Segregation in Dutch Cities’, Education and Urban Society, 38 (2), 228–47. Karsten, S., Ledoux, G., Roeleveld, J., Felix, C. and Elshof, D. (2003), ‘School Choice and Ethnic Segregation’, Educational Policy, 17 (4): 452–77. Merry, M. S. (2012), ‘Segregation and Civic Virtue’, Educational Theory, 62: 465–86. Miskovic, M. (ed.) (2013), Roma Education in Europe: Practices, Policies, and Politics. New York, NY: Routledge. Murat, M. (2012), ‘Do Immigrant Students Succeed? Evidence from Italy and France’, Global Economy Journal, 12 (3): 1–20. Musterd, S. and Ostendorf, W. (1998), Urban Segregation and the Welfare State. Inequality and Exclusion in Western Cities. New York: Routledge. Musterd, S., Marcińczak, S., van Ham, M. and Tammaru, T. (2017), ‘Socioeconomic Segregation in European Capital Cities. Increasing Separation between Poor and Rich’, Urban Geography, 38 (7): 1062–83.

24

Understanding School Segregation

Noblit, G. W. (2015), ‘Introduction’, in Noblit, G. W. (ed.), School Desegregation. Breakthroughs in the Sociology of Education. Sense Publishers, Rotterdam. Oakes, J. (2005), Keeping Track: How Schools Structure Inequality (2nd ed.). New Haven, CT: Yale University Press. Oberti, M. (2007a), L’école dans la ville: ségrégation–mixité–carte scolaire. Paris: Presses de Sciences-Po. Oberti, M. (2007b), ‘Social and School Differentiation in Urban Space: Inequalities and Local Configurations’, Environment and Planning A., 39 (1): 208–27. OECD (2012), Equity and Quality in Education: Supporting Disadvantaged Students and Schools, OECD Publishing. Ogbu, J. (2003), Black American Students in an Affluent Suburb. A Study of Academic Disengagement. Lawrence Erlbaum Associates: Mahwah, NJ. Orfield, G. (2001), Schools More Separate: Consequences of a Decade of Resegregation. Civil Rights Project, Harvard University. Cambridge, MA. Orfield, G. and Lee, Ch. (2005), Why Segregation Matters: Poverty and Educational Inequality. Civil Rights Project, Harvard University. Cambridge, MA. Patrinos, H. A., Barrera-Osorio, F. and Guáqueta, J. (2009), The Role and Impact of Public Private Partnerships in Education. Washington DC: World Bank. Rangvid, B. S. (2007), ‘School Composition Effects in Denmark: Quantile Regression Evidence from PISA 2000’, Empirical Economics, 32 (2): 359–88. Raveaud, M. and van Zanten, A. (2007), ‘Choosing the Local School: Middle Class Parents’ Values and Social and Ethnic Mix in London and Paris’, Journal of Education Policy, 22 (1), 107–24. Robertson, K., Mundy, K., Verger, A. and Menashy, F. (eds.) (2012), Public Private Partnerships in Education: New Actors and Modes of Governance in a Globalizing World. Cheltenham, UK: Edward Elgar. Rothstein, R. (2015), ‘School Policy is Housing Policy: Deconcentrating Disadvantage to Address the Achievement Gap’, in Race, Equity, and Education: Sixty Years from Brown (pp. 27–43). Economic Policy Institute. Rumberger, R. W. and Palardy, G. J. (2005), ‘Does Segregation Still Matter? The Impact of Social Composition on Academic Achievement in High School’, Teachers College Record, 107: 1999–2045. Saporito, S. (2003), ‘Private Choices, Public Consequences: Magnet School Choice and Segregation by Race and Poverty’, Social Problems, 50 (2): 181–203. Shattuck, J. (2012), Ten Years After: A History of Roma School Desegregation in Central and Eastern Europe (Rostas I., Ed.). Central European University Press. Söderström, M. and Uusitalo, R. (2010), ‘School Choice and Segregation: Evidence from an Admission Reform’, Scandinavian Journal of Economics, 112 (1): 55–76. Sykes, B. (2011), Spatial Order and Social Position. Amsterdam: University of Amsterdam. Tammaru, T., Marcinczak, S., Van Ham, M. and Musterd, S. (2015), Socio-Economic Segregation in European Capital Cities: East meets West. London: Routledge. Taylor, C. and Gorard, S. (2001), ‘The Role of Residence in School Segregation: Placing the Impact of Parental Choice in Perspective’, Environment and Planning A., 33 (10): 1829–52. Termes, A., Bonal, X. and Verger, A. (2016), ‘Concession Schools: An Emblem of Charter Schools in Latin America’, Compare, 46 (6), 984–9. Thrupp, M., Lauder, H. and Robinson, T. (2002), ‘School Composition and Peer Effects’, International Journal of Educational Research, 37 (5): 483–504. Tropp, L. R. and Prenovost, M. (2008), ‘The Role of Intergroup Contact in Predicting Children’s Interethnic Attitudes: Evidence from Meta-Analytic and Field Studies’, in

Introduction

25

Sheri R. Levy and M. Killen (eds.), Intergroup Attitudes and Relations in Childhood Through Adulthood, 236–60. New York: Oxford University Press. Valenzuela, J. P., Bellei, C. and de los Ríos, D. (2014), ‘Socioeconomic School Segregation in a Market Oriented Educational System. The Case of Chile’, Journal of Education Policy, 29 (2), 217–41. Van Der Slik, F. W. P., Driessen, G. W. and De Bot, K. L. (2006), ‘Ethnic and Socioeconomic Class Composition and Language Proficiency: A Longitudinal Multi-­level Examination in Dutch Elementary Schools’, European Sociological Review 22 (3): 293–308.

Part I

Patterns of School Segregation and Social Inequalities

2

School Segregation in France: The Role of Public Policies and Stakeholder Strategies Georges Felouzis, Barbara Fouquet-Chauprade and Samuel Charmillot

1.  School segregation in France: an overview From a scientific and policy standpoint, the issue of school segregation is inextricably linked to the analysis of educational policies. With regard to France, it is useful to understand how segregative phenomena are closely related to specific societal developments and are therefore part of the very functioning of the French educational system. To this end, we shall present three educational policies that are emblematic of the French educational system and that have contributed to school segregation as we see it today. These are the collège unique (i.e. comprehensive school), the carte scolaire (i.e. school mapping) and the ‘priority education’ policies. All three are part of recent French policy on education and give a fairly accurate picture of its functioning. After a brief introductory presentation of the present state of knowledge on school segregation in France, we shall show how the implementation of these three policies produces effects that indirectly contribute to segregation in the French educational system. School segregation can be broadly defined as the unequal distribution of the school population according to particular characteristics, such as social origin, ethnic origin, sex or any other ascriptive characteristic. A more restrictive definition considers that spatial separation of pupils in schools, classes, tracks or different ability groups does not suffice to speak of school segregation. Two further conditions are necessary. Firstly, the separation of pupils must lead to specific forms of inequality or exclusion (Barthon, 1998). In this sense, if 95 per cent of the pupils of a school come from privileged socioeconomic backgrounds, this situation cannot be described as segregative. School segregation is therefore characterized by its negative consequences for those individuals who experience and are directly affected by it. One simple demarcation criteria is that in all cases of segregation, individuals have little or no choice: they cannot leave their neighbourhoods or their ghettoized schools because they do not have the economic, social, cultural or regulatory means, as in, for example, the carte scolaire (school mapping) system. Secondly, for segregation to be described as ‘educational’, its consequences must come from ‘the specific actions of the institutions and the educational actors’ (Van Zanten, 2012), and not only result from factors outside the school. This segregation

30

Understanding School Segregation

can be the indirect result of a school policy aimed at grouping pupils into educational units according to their academic level. Numerous studies have shown that tracked educational systems – which separate pupils according to their academic performance – indirectly lead to separation according to adjoining and correlative criteria such as social and ethnic origin. This phenomenon of interweaving segregation has been observed in the United States (Gamoran and Mare, 1989; Lucas and Berends, 2007; Oakes, 1985), Belgium (Franquet et al., 2010), Germany (Baumert et al., 2006; Schnepf, 2002), Switzerland (Felouzis and Charmillot, 2013) and France prior to the Haby reform1 (Baudelot and Establet, 1971). School segregation, however, is not necessarily the consequence of an explicit policy of separating pupils. It can also result from more diffuse processes, deeply rooted in society and linked to the combined practices of different social agents. This, rather, is the situation that we encounter in France. The collège unique (comprehensive school) was in fact developed to counter the tracked system that prevailed until the mid-1970s. It supported and enabled the democratization of secondary education. However, it did not guarantee social and ethnic diversity in educational institutions for various reasons: the desire of families that their children be educated with members of their own social or ethnic group (Felouzis and Perroton, 2009), urban segregation ‘imported’ into the school through school mapping (Van Zanten, 2012; Oberti, 2007), the development of private education (Merle, 2011a), the diversification of the public educational offer (Barthon and Monfroy, 2006; Trancart, 2000), competition between schools (Broccolichi et  al., 2010), uncontrolled easing of school mapping after 2007 (Merle, 2011b), etc. These end-­to-end phenomena are not necessarily the result of political will, but rather of a ‘tyranny of small decisions’ (Schelling, 1978) or of social actors’ strategies to optimize their positions (Maresca, 2003). Research on school segregation in France has come rather late and mostly focused on local geographical spaces (see, for example, Ben Ayed et al., 2013; Fack et al., 2014; Felouzis, 2003; Ly et al., 2014; Merle, 2010; Trancart, 1998). A more complete picture of the extent and structure of school segregation on a national scale has been obtained only very recently, thanks in particular to the work of Ly and Riegert (2015). Based on data from the Directorate of Evaluation, Foresight and Performance (DEPP) for students enrolled between 2007 and 2013, Ly and Riegert calculated a standardized exposure index. This index estimates the extent to which a group of students are exposed to members of their own group in a given spatial unit. It is therefore a measure of the degree of potential contact between groups within a school. Another property of this index is that it allows us to break down segregation into two sources: between-­school and within-­school segregation. It is therefore possible to estimate which part of the total segregation is due to differences between the schools and which part is due to the composition of the classes. Using this index, Ly and Riegert were able to measure social and academic segregation in secondary 1 (collège) and secondary 2 (lycée) education.2 The Haby reform took place in 1975. Secondary 1 education (which is called collège in France) went from a tracked system to a comprehensive system (collège unique). 2 Collège lasts for four years, from the sixth (Year 7 in the UK) to the third grade (Year 10). It takes students from the age of 11 to 14. The Lycée (sixth form college) lasts for three years: the second (Year 11), the first (Year 12) and the Terminale (Year 13); students are between 15 and 17 years of age. 1

School Segregation in France

31

Figure 2.1  Between-­school and within-­school segregation at the national level, measured by the exposure index for the 2007 cohort. Source: Ly and Riegert, 2015, p. 33. Notes: 1. This figure should be read as follows: for the cohort of students entering sixth grade in 2007, academic segregation in first grade was 20.6 per cent between schools and 38 per cent between classes. ‘Good students’ represent 22.4 per cent of the total students. 2. The term ‘SPC+’ refers to students whose parents belong to the highest socio-­professional categories. 3. Sixth grade (Year 6 in the UK) to third grade (Year 10) refers to secondary 1 education (collège). Second grade (Year 11) to Te (Year 13) refers to secondary 2 education (lycée).

In secondary 1 education, between-­school social segregation3 is 16–17 per cent depending on the grade level (Figure 2.1). When within-­school social segregation is taken into account, the exposure index increases by 4–6 percentage points depending on the grade level concerned. This decomposition of the total segregation between the Social segregation is measured by taking as reference the most socially advantaged students (SPC+). These are students whose parents belong to the following socio-­professional categories: business leaders, liberal professions, civil service, intellectual and artistic professions, business executives, heads of schools, teachers, etc.

3

32

Understanding School Segregation

between-­school and within-­school level indicates that within-­school segregation has a limited effect, since classroom composition accounts for only about 20 per cent of total segregation. The authors also measure academic segregation, taking as a reference the proportion of ‘good students’.4 In secondary 1 education, between-­school academic segregation remains relatively low, varying between 7 per cent and 9 per cent depending on the grade level considered. When class composition is taken into account, academic segregation increases to 13–18 per cent depending on the grade level. The structure of academic segregation is different from that observed for social segregation, since within-­school segregation accounts for 50 per cent of the total segregation. In secondary 2 education, academic segregation increases strongly. This sudden increase in the exposure index is mainly due to the separation of pupils into tracks at the start of secondary 2 (general, technological and vocational tracks). These measurements of social and academic segregation calculated on a national scale conceal strong geographical disparities. A more detailed analysis of social segregation indeed reveals significant variation across regions, with an exposure index ranging from 2 per cent to 27 per cent. The authors note that while within-­school segregation varies little from one region to another, between-­school segregation appears to be much lower in rural areas. The low population density in rural areas means that schools recruit from larger catchment areas, which encourages social diversity. Conversely, more urbanized regions have higher segregation indices. This is due, on the one hand, to strong residential segregation in urban areas. On the other hand, the geographical proximity between schools has the effect of creating competition, with some schools being more or less avoided, while other schools are highly sought after, which in turn accentuates the effects of residential segregation. So far, little research has been done in France regarding ethnic segregation, and there are several reasons for this. Firstly, it took France a long time to see itself as ‘a country of immigration’ (Noiriel, 2016). Until the 1970s, foreign workers, mainly from the former French colonies and from southern Europe, were perceived as populations only temporarily present in France and were supposed to return to their country of origin, although this was not what happened in reality. Secondly, during that same period, French sociology was strongly influenced by Marxist theory, which did not allow for thinking in terms of migratory inequalities (van Zanten, 2012). Thirdly, the notion of the French nation, conceived as ‘one and indivisible’ and which considers the citizen in a neutral way (without a singular affiliation), is a brake on reasoning in terms of inequality as linked to the ethnic origins of individuals. Finally, the French statistical system reflects this third point, since the official available databases do not contain information that identifies individuals of immigrant origin (Simon, 2008). Debates around what is known in France as ‘ethnic statistics’ or the ‘statistics of origin’ are therefore often heated and virulent, preventing pragmatic reasoning concerning the inequalities linked to the ethnic origin of individuals (Felouzis, 2008). ‘Good students’ are those who have not had to repeat years between the sixth and the fourth grade, who have passed the brevet school certificate – a national assessment at the end of the third grade – in the general track and achieved an average within the top 25 per cent of the awarded marks.

4

School Segregation in France

33

However, since the early 2000s, research on the links between ethnicity and school segregation has been emerging. Felouzis et al. (2007), for example, studied the social and ethnic segregation in a local context in the southwest of France (Bordeaux Academy) and their analysis highlights the extent of ethnic segregation. Using a concentration index, they calculate the percentage of students who should change schools to have a balanced distribution of students of a given group across all schools. Considering first social origins, they show that 29 per cent of pupils from disadvantaged backgrounds should change schools in order to achieve a balanced distribution. The concentration index increases to 62 per cent for non-­native students and, if only students from the Maghreb, Sub-Saharan Africa and Turkey are considered, no less than 89 per cent of pupils should change schools. This concentration of non-­native students in certain schools has consequences for their academic results and educational paths. The authors estimate that, all other things being equal, students in schools where the percentage of non-­native students exceeds 20 per cent score on average 0.4 points lower (out of 20) in the brevet school certificate (Felouzis, 2003). Recent research in France suggests that academic, social and ethnic segregation is an important factor in explaining inequalities in academic achievement and in educational paths. We shall now discuss three educational policies that were conceived to reduce inequalities, but in fact indirectly contributed to reinforcing school segregation.

2.  Educational policies and the question of segregation 2.1  Comprehensive school in France According to Mons (2007), the collège unique (comprehensive school) is defined as ‘a school organisation whose primary characteristic is to bring together almost an entire generation within a unique educational structure until the end of the first cycle of secondary education’ (p. 89). In order to be considered a comprehensive school, the school system must (1) present a single syllabus without tracking, (2) have comparable schools and (3) offer strictly the same opportunities. However, ever since its inception, the collège unique has proposed parallel paths, thus derogating from a comprehensive school model (Prost, 2013). In fact, it was only twenty years after the 1975 Haby reform that schools began to be defined as comprehensive schools. Until then, some pupils continued to be streamed at the end of the fifth and fourth grade towards ‘technological’ or ‘social integration’ tracks. Unsurprisingly, orientation towards these non-­general tracks was strongly correlated with the social origin of the pupils: whereas 98 per cent of the children of executives entered the general track in the third grade, only 78 per cent of working-­class children did (Mons, 2007). This disparity is not the result of pupils’ academic achievement alone, but also of the orientation procedures themselves, which Duru-Bellat and Mingat (1993) have shown to be socially unequal. The uniformity of education offered in the comprehensive school is also called into question by the presence of certain ‘elitist’ classes (Vasconcellos and Bongrand, 2013).

34

Understanding School Segregation

This is the case, for example, in the ‘European’ or ‘international’ classes5 in which 5.1 per cent of secondary school pupils were enrolled in 2012. Moreover, the creation of ‘preserved classes’, a strategy used by some schools, particularly segregated ones, to retain families who have the symbolic and economic means to choose other schools, also contributes to the erosion of the idea of a comprehensive school. This is the case, for example, in the ‘scheduled classes’ (CHAM), which provide artistic education or sports-­study classes. These distinctions through the choice of options contribute to reinforce segregation by creating situations of competition between public schools within the same geographical area in order to attract (or retain) the ‘best’ students. It also has important repercussions for a school career (Duru-Bellat and van Zanten, 2012). For example, 26 per cent of school teachers’ children take German as their first language, compared to only 8 per cent of children of working-­class and unemployed parents. Thus, although it is theoretically no longer possible to distinguish between different tracks in secondary 1 education, in reality, the choice of options reintroduces a strong distinction between classes within schools and between pupils’ educational paths. For Ly et al. (2014), this is one of the main sources of within-­school segregation, since ‘of the 16% of social segregation . . . which cannot be explained by chance, 62% . . . can be explained by the distribution of options in classes. If schools did not create their classes according to pupils’ choices of options, two thirds of the segregation cases might not have taken place’ (p. 5). Thus, in the early 1980s, in his preface to Paty’s work, 12 colleges in France (1981), Crozier noted that differentiation according to tracking, which prevailed prior to the Haby reform, had already been replaced by the differentiation of the schools. The idea was that schools became the functional substitutes for tracks. This analysis has not been contradicted, since schools are today highly differentiated according to the academic, social and ethnic nature of their student body. In a recent book, Ben Ayed (2015) argues that behind a unifying discourse lies a strong resistance to the comprehensive school, legitimizing the separation of pupils and the differentiation of educational paths. He pointed out that ‘during the contemporary period, educational policies always articulated a seemingly unifying discourse, while emphasising the need for differentiation in teaching to adapt to the “diversity” of the pupils. Recently, one of the central themes of the school debate was the criticism and the questioning of the comprehensive school’ (Ben Ayed, 2015, p. 16). Comprehensive school thus continues to be controversial, both for policy makers and for parents who doubt its effectiveness (Mons, 2007). Finally, it should be noted that although no structural reform has changed the organization of the comprehensive school over the last twenty years, this does not mean that it has not evolved. During this period, the practices and the implementations of certain mechanisms have modified its organization. In particular, there has been a homogenization in the educational offer, related to the ending of certain tracks in European and international classes focus on learning a foreign language. The European sections teach the chosen language for an extra two hours per week. The international sections offer specific courses in history-­geography and literature of the language concerned.

5

School Segregation in France

35

secondary 1, and a significant drop in repetition rates (Caille, 2014). This has had a positive impact on the educational paths of students in secondary 1 education, which have become more linear, and on their orientations in secondary 2, with an increase in orientations towards general and technological tracks. However, the effects of this homogenization on equity must be put into perspective, as options continue to be available. Moreover, along with this homogenization in the educational offer, there is a strong diversification at the territorial level, producing an amplification of school segregation (see, for example, Barthon and Monfroy, 2006; Trancart, 2000).

2.2  The carte scolaire: a tool against segregation? One cannot really describe the carte scolaire (school mapping) as a ‘school policy’. Rather, it could be seen as a series of provisions aimed at matching educational facilities, on the one hand, and families’ schooling needs, on the other, within a given local area. Introduced in 1963, the general underlying principle of school mapping was to condition the assignment of a pupil to a particular public school based on where she or he lives. The primary function of school mapping was to regulate the flow of pupils among schools, but gradually it was assigned another objective: to ensure social diversity. However, research on the effects of school mapping yields an identical finding: it contributes little to fostering the mixing of diverse groups of pupils. In the best of cases, school mapping merely reproduces the segregation of the residential space in the educational space. In some cases, it even helps accentuate segregation further still. This is what is meant, for example, when Oberti (2007) argues that ‘school mapping applies unequally to different social groups. It reinforces the protection of the most favoured, weakens social diversity in the ordinary schools, and finally accentuates the relegation and disqualification of the most disadvantaged’ (p. 262). One can therefore question the reasons for the inability of school mapping to ensure social diversity. A first explanation is that its student management objectives are more important than any social objectives. Van Zanten and Obin (2008) consider that this system has been more effective in its role of regulating student flows than in preserving social diversity in schools. Moreover, the history and the evolution of school mapping as reported by these two authors show that school mapping has never been as rigid as intended. Unlike in South Korea, Japan or Greece, school mapping in France has always been deployed along with a possibility of derogation (Mons, 2007). A second explanation is related to the size of school catchment areas. As the school catchment areas operate on a reduced territorial basis, school mapping can therefore contribute to diversity only where residential segregation also involves small spatial units: small neighbourhoods, islets, etc. If residential segregation involves larger units, for example at the municipality level, school mapping will have no effect, except to replicate the structure of the residential space in the school space. In other words, the areas of recruitment where the most disadvantaged pupils reside are too large and too homogeneous for school mapping to have any real effect in terms of diversity (Oberti, 2005). The segregative effects of school mapping also come from the strategies of parents who seek to enrol their children in schools other than those imposed on them by the catchment area they are living in. These avoidance strategies contribute to the

36

Understanding School Segregation

reinforcement of school segregation for two reasons. On the one hand, because the schools which are most avoided have a higher proportion of pupils from immigrant and disadvantaged backgrounds. On the other hand, it is mainly wealthier parents who possess the economic and cultural means to implement these strategies. These avoidance practices can take many forms. A first way of avoiding a school catchment area is to move to an area that guarantees assignment to a preferred school. In a study of the Paris region, Korsu (2004) shows that these practices remain marginal. He estimates that the rate of students who leave the local public sector following a move at the end of primary school is only 5–6 per cent of the total number of pupils and that the number of moves that are really for educational reasons is probably even lower. While avoidance by moving is limited, it still raises the question of the interaction between schools and residential choice. Oberti and Barthon (2000) indicate that the choice of housing, especially in large cities, is increasingly dependent on schools and not just on the proximity of work and the family. Thus, ‘the classic elements of urban segregation (property market logic, gentrification of city centres, housing policies, income inequalities, location of enterprises and investments, etc.) would be supplemented by a dimension related to the choice of schools’ (p.  308). There is therefore not only an impact of urban segregation on school segregation, but also a ‘backward effect’ of school segregation on urban segregation (Fack and Grenet, 2009). Control of space (access to local resources, place of residence, mobility) would thus become a resource in its own right (Oberti, 2005; François and Poupeau, 2009). This means that the advantages associated with spatial position are not only symbolic (Pinçon and Pinçon-Charlot, 2009), but also offer ‘added value’ to access the public and private resources of the school. Control of this spatial dimension thus represents a ‘spatial capital’, which, like cultural capital, contributes to the reproduction of school inequalities (Barthon and Monfroy, 2011). Derogations are another means used to circumvent school catchment areas. Parents may request that their child be enrolled in a school other than that imposed on them by school mapping. These requests for derogation are often justified for organizational reasons (proximity to the workplace, presence of another family member, etc.), but they can also be used informally to control the school attended by the child. Various studies carried out on the use of derogations show that the proportion of pupils enrolled in public institutions outside the sector remains a minority. Chausseron (2001) indicates that this affects about 10 per cent of French pupils enrolled in the fifth grade. However, these requests for derogations are likely to vary considerably depending on the geographical area, and are particularly noticeable in urban areas. For example, Maresca (2003) estimates that in some municipalities in the Paris region, more than half of families seek to avoid the public school imposed upon them at the end of primary school. Finally, some families resort to private education in order to bypass their catchment area. Chausseron (2001) indicates that 20 per cent of families use private education, again with important differences depending on the geographical location. In Paris, for example, the share of the private sector reached 34 per cent at the beginning of secondary 1 (Gillotte and Girard, 2005).

School Segregation in France

37

Even if avoidance strategies concern only a minority of students, they nevertheless produce effects that contribute to the accentuation of school segregation. This occurs firstly because only certain schools are the object of these avoidance strategies. The most requested schools are generally characterized by a high proportion of pupils who did not repeat years and by a limited presence of pupils from immigrant or disadvantaged backgrounds (van Zanten, 2012). Indeed, in the absence of clear information on the quality of schools, family choice depends on the reputation of the schools, which is collectively constructed within ‘judgement networks’ (Felouzis and Perroton, 2007). The criteria that define the educational level of a school is therefore filtered by the social and ethnic characteristics of its public. Because of its visibility, ethnic origin in particular is an indicator of the social identity of a school, and although it is rarely mentioned explicitly, it is often a reason for circumvention. Thus, ‘the ethnic criteria may ultimately summarise risky school contexts for the avoiding families’ (Lorcerie, 2003, p. 94). Secondly, avoidance strategies are conditioned by the resources available to families. It is mainly wealthier parents who possess the economic and cultural means to implement these strategies. Conversely, parents from disadvantaged backgrounds are ‘captives’ of the system, having neither the financial means to afford private education for their children, nor the ‘skills’, and particularly access to information, to enable these strategies. In the end, strategies to avoid designated school catchment areas contribute to reinforcing school segregation because ‘bad schools’ are perceived to have a high proportion of pupils of immigrant and disadvantaged backgrounds and because the more favoured families manage to escape these schools. However, pupil transfers, resulting from avoidance strategies, do not universally have the same effect. As Felouzis’ analysis (2005) on segregation in the Bordeaux academy shows, the consequences for school segregation are greatest in the most disadvantaged urban areas: In the wealthiest catchment areas, there is a game of musical chairs: the secondary school of the catchment area is not filled with local students, who are now in private education, but takes those from other catchment areas. This is a neutral operation for the school in terms of number of students as well as social and ethnic composition of its public. In the most stigmatised schools, what is lost to one side is not compensated by the other. The loss of pupils is either to the private sector or to other schools, without the contribution of other pupils compensating for these losses. There is then a process of self-­reproducing segregation. p. 45

Faced with the impossibility of school mapping to deliver diversity, certain politicians, such as Nicolas Sarkozy in 2007, proposed relaxing school mapping in order to promote greater freedom of choice. Merle (2011a, b) evaluated the effect of this relaxation of school mapping in three French cities (Paris, Bordeaux and Lille) and observed an increase in the avoidance of the most disadvantaged schools to the benefit of schools which recruit from the most favoured social categories, mechanically producing a reinforcement of the segregative phenomenon.

38

Understanding School Segregation

The easing of school mapping in 2007 was accompanied by measures designed to guarantee social diversity. In particular, that requests for derogations from scholarship students should be given priority over requests from wealthier families wishing to change schools on the grounds of preference or geographical proximity.6 However, an analysis conducted by Merle (2011a) shows not only that scholarship students rarely request derogation, but also that this criterion was not automatically given priority over other reasons. The author concludes that there is a phenomenon of ‘relative, or even absolute, ghettoisation for the most disadvantaged schools, combining a loss of enrolment and loss of pupils from advantaged backgrounds’ (p.  47). In order to continue to attract students from advantaged backgrounds and to reduce their flight to more favoured schools, the separation of pupils in different classes according to their academic level has been further strengthened in these schools. Between-­school segregation is therefore complemented by within-­school segregation whose negative effects have been demonstrated both on pupils’ competences (Duru-Bellat and Mingat, 1997; Ly et al., 2014), and on non-­cognitive dimensions such as discipline (Debarbieux, 1996) or educational well-­being (Fouquet-Chauprade, 2014). In his evaluation of the effect of the easing of the catchment areas in the Rennes academy, Merle (2011b) also shows that 2000–06 was characterized by a period of gentrification of the private schools, while the period following the easing of school mapping (2006–09) was characterized by an increase in social segregation.

2.3  ‘Priority education’ policies: tackling the consequences of school segregation Priority education policies are directly associated with policies to combat the effects of segregation. In France, they do not aim to prevent segregation, nor do they seek desegregation (as is the case, for example, of the USA or South Africa). Instead, they are intended only to mitigate or even eliminate its most negative consequences. It should be noted that the choices made in France in terms of ‘positive discrimination’ for students from disadvantaged backgrounds focus on a territorialized policy. Therefore, they are not about acting in favour of socially disadvantaged groups, but they are rather targeting disadvantaged areas, because of higher proportions of students from disadvantaged backgrounds and with academic difficulties. We do not intend to trace back the history of this policy7 which first saw the light of day in France in the early 1980s. Rather, we shall focus on its evolution over the last twenty years. Merle (2012) demonstrates the rapid proliferation of priority education. In 1997, the Priority Education Zones (ZEP) educated 14.3 per cent of the pupils in secondary 1 education. At the beginning of the 1997–98 school year, and following the Moisan and Simon report (1997), the ZEP became the Priority Education Networks (ZEP-REP). Seven years later, 20.4 per cent of the students were educated through If parents request a derogation, it is accepted only if there are places available in the requested school. If the number of applications for a school exceeds its capacity, priority is given to students receiving a scholarship. 7 On this point, see, for example, the work of Robert (2009). 6

School Segregation in France

39

these networks. Aware of the dilution of resources and the low relevance of such a facility, the Ministry of Education decided in 2006 on the creation of an Ambition for Success Network (RAR), which had to allow a refocusing of resources on a smaller number of schools: less than 5 per cent of pupils in secondary 1 education were enrolled in RAR. However, the ZEP-REP did not disappear and became the School Success Networks (SSN), which concerned 15 per cent of students. The new RAR policy aimed at concentrating more resources towards the schools that were most in need. However, overlaying the RAR and ZEP-REP tended to blur the message. In 2010 the Collège and Lycée for ambition, innovation and success (CLAIR) appeared, which was more explicitly aimed at improving school climate. This system extended to primary schools the following year (ECLAIR). In 2011, the majority of the RAR disappeared. This demonstrates that priority education has become increasingly complex by superimposing recent measures onto older ones. Moreover, since the early 1990s, priority education policies have been linked to urban policies, which further complicates their implementation, but probably gives more coherence to the fight against school inequalities (Lorcerie, 2010). At the same time, specific mechanisms are added to these policies (refresher courses, educational support, etc.) organized by the local authorities or the schools themselves. The so-­called ‘refounding’ of the school policy established in 2014, which aimed at reducing inequalities and promoting the success of all, tried to reduce this overlapping of measures and to refocus efforts on a smaller number of schools by thoroughly revising the catchment areas. The other objective was that of a ‘rubbing out’ of threshold effects. This policy established REP+ which focused mainly on pedagogical innovation and the modification of teachers’ working time by granting them non-­teaching hours to promote collective work, consultation, etc. Ultimately, these priority education policies face a major challenge: the confusion between the objectives of the policies and the means by which they can be achieved (Robert, 2009). They aim to restore equality between students by combating the deleterious effects of segregation, even though the nature and sources of the phenomena on which they want to act are very disparate (Fouquet-Chauprade and Dutrévis, 2015). This leads to a profusion of objectives and measures that hinder their effectiveness: do we want to take action on class sizes, on the construction of innovative pedagogical projects, on collective works, on school-­family relations or on the school climate? Many of the objectives of priority education are therefore vague and variously measurable. An evaluation report by the DEPP (CIMAP, 2013) notes that ‘priority education has gradually been assigned a wide range of operational objectives, both for pupils and their families (social, health, cultural, educational, orientation, vocational integration objectives, etc.), without these being truly formalised, nor always implemented, nor truly evaluated’ (p. 16). Assessments of these priority education policies are rather nuanced or even negative in terms of their effects on inequalities in competences and orientation, as well as on the school climate (see, for example, Armand and Gille, 2006; Meuret, 1994; Moisan and Simon, 1997). Moreover, ‘labelling effects’ are far from being anecdotal, since the system leads to a process of stigmatization of the schools and pupils attending those schools, ultimately causing an increase in school segregation (Fouquet-Chauprade and Dutrévis, 2015; Merle, 2012). Finally, far from being a policy of combating segregation,

40

Understanding School Segregation

priority education ends up being a ‘source of social and academic segregation’ (Merle, 2012, p. 63). Schools integrating priority education are avoided by families who can afford to do so (van Zanten, 2012), students who attend these schools are stigmatized (Fouquet-Chauprade, 2014; Merle, 2012), academic expectations are revised downwards, and the school climate is often tense, even violent (Debarbieux, 1996).

3.  Conclusion The three distinct school policies that we have examined in relation to school segregation are similar in a number of ways. Firstly, all three policies, each in their own way and to varying degrees, try to establish a system of compulsory education that is as unified and egalitarian as possible. This is indeed the leading principle of the collège unique, one of the goals of the carte scolaire and the main objective of priority education. Secondly, they all fail in doing so, either because their objectives are not always explicit, or because the means implemented are not sufficiently directly linked to these objectives. The case of the comprehensive school is exemplary. The Haby reform of 1975 aimed at accompanying school massification (assuming that this would lead to the democratization of education) and limiting social inequalities in access to the baccalauréat. However, from the 1980s onwards, differentiation between schools was such that it was necessary to establish a priority education policy as early as 1981. The latter was intended to limit inequalities in learning outcomes and educational paths between pupils linked to social segregation between schools. Successive evaluation reports and ad hoc studies have shown that this objective has not been achieved, indicating that giving more resources to segregated schools is not enough to reduce educational inequalities. Finally, the way in which these policies are implemented produces perverse effects which contribute to reinforcing segregation. This is particularly the case of school mapping. Its application, avoidance and relaxation all too often lead to confine disadvantaged and immigrant students to less attractive schools, while other students succeed in avoiding them. This is also the case with the priority education policy which does not aim to combat segregation, the risk being to legitimize it under the pretext of fighting against its consequences. The question then arises of how to explain this very singular situation of the French educational system with regard to the persistence and the extent of educational segregation and the consequent inequalities. The sources of the problem are, of course, multiple and one cannot reduce them to a single cause. They are the consequences of the actions of all actors involved in the educational field: parents, teachers, administrators, and policy makers. Yet it is clear from our analysis of public education policies that many of the choices made in the past have not improved equity in school. For a long time, it had been thought possible to combat school inequalities by promoting a double process: massification and compensation. Massification, which consisted of an unprecedented opening of secondary and higher education, was supposed to bring egalitarian qualities. While this may be true

School Segregation in France

41

regarding access to the baccalauréat for all tracks, this is not true for the ‘elite’ tracks in secondary 2 and higher education. This is also not the case for school achievement at the end of primary and secondary 1 education, which today depends more on the ascriptive characteristics of pupils (their social origin, their migratory path, their sex, etc.) than it did before (Felouzis et al., 2016). Compensation is linked to the fact that massification – in particular, in secondary 1 education – was very quickly carried out in a segregative manner. From the beginning of the 1980s, this segregative massification produced schools in which the conditions for education were unfavourable through the sheer concentration of social, economic and educational problems. The idea was therefore to compensate for this cumulative disadvantage by the allocation of a surplus of human and operational resources to these schools. However, these compensation policies did not produce the desired effects. They failed to limit inequality of achievement and of educational paths between students, because they merely distributed resources instead of focusing on student learning. School segregation is a powerful factor in the production of inequality. It produces conditions that are unfavourable to learning through the effects of the school climate, the composition of the school public and the quality of teaching. A school policy aimed at limiting the extent of inequality in student achievement would explicitly target segregation and would have as its primary objective diversity in schools – that is to say, coexistence of pupils of all academic levels within the same educational units, until the end of compulsory education. Finally, all these findings call for real policies of desegregation which are still missing in France.

References Armand, A. and Gille, B. (2006), La contribution de l’éducation prioritaire à l’égalité des chances des élèves. MENESR/IGEN/AGEAN/. Barthon, C. (1998), La ségrégation comme processus dans l’école et dans la ville. Revue européenne de migrations internationales, 14 (1): 93–103. Barthon, C. and Monfroy, B. (2006), Une analyse systémique de la ségrégation entre collèges: l’exemple de la ville de Lille. Revue française de pédagogie, 156, 29–38. Barthon, C. and Monfroy, B. (2011), Choix du collège et capital spatial: étude empirique appliquée aux collégiens lillois. Espace, population, sociétés, 2, 321–35. Baudelot, C. and Establet, R. (1971), L’école capitaliste en France. Paris: Maspero. Baumert, J., Stanat, P. and Watermann, R. (2006), Schulstruktur und die Entstehung differenzieller Lern- und Entwicklungsmilieus. In J. Baumert, P. Stanat and R. Watermann (Eds.), Herkunftsbedingte Disparitäten im Bildungswesen: Differenzielle Bildungsprozesse und Probleme der Verteilungsgerechtigkeit (pp. 95–188). VS Verlag für Sozialwissenschaften. Ben Ayed, C. (2015), La mixité sociale à l’école. Tensions, enjeux, perspectives. Paris: Armand Colin. Ben Ayed, C., Broccolichi, S. and Monfroy, B. (2013), Quels impacts de l’assouplissement de la carte scolaire sur la ségrégation sociale au collège? Tendances nationales et déclinaisons locales contrastées. Éducation & formations, 83, 39–57. Broccolichi, S., Ben Ayed, C. and Trancart, D. (2010), École: les pièges de la concurrence. Comprendre le déclin de l’école française? Paris: La Découverte.

42

Understanding School Segregation

Caille, J.-P. (2014), Les transformations des trajectoires au collège: des parcours plus homogènes mais encore liés au passé scolaire et à l’origine sociale. Éducation & formations, 85, 5–30. Chausseron, C. (2001), Le choix de l’établissement au début des études secondaires. Note d’information n°42. Vanves: MEN. CIMAP (2013), Evaluation de la politique de l’éducation prioritaire. Rapport de diagnostic. MEN/SGMAP. Debarbieux, E. (1996), La violence en milieu scolaire. 1. État des lieux. Paris: ESF. Duru-Bellat, M. and Mingat, A. (1993), Pour une approche analytique du fonctionnement du système éducatif. Paris: PUF. Duru-Bellat, M. and Mingat, A. (1997), La constitution de classes de niveau dans les collèges; les effets pervers d’une pratique à visée égalisatrice. Revue française de sociologie, 38 (4): 759–89. Duru-Bellat, M. and Van Zanten, A. (2012), Sociologie de l’école. Paris: Armand Colin. Fack, G. and Grenet, J. (2009), Sectorisation des collèges et prix des logements à Paris, Actes de la Recherche en Sciences Sociales, 180 (5): 44–62. Fack, G., Grenet, J. and Benhenda, A. (2014), L’impact des procédures de sectorisation et d’affectation sur la mixité sociale et scolaire dans les lycées d’Ile-­de-France. Rapport IPP n°3. Felouzis, G. (2003), La ségrégation ethnique au collège et ses conséquences. Revue française de sociologie, 44 (3): 413–47. Felouzis, G. (2005), De l’école à la ville: comment se forment les collèges ghettos? À partir d’une enquête statistique dans l’Académie de Bordeaux. Informations sociales, 125 (5): 38–47. Felouzis, G. (2008), Les catégories ethniques en sociologie: éléments pour un débat. Revue française de sociologie, 49 (1): 127–32. Felouzis, G. and Charmillot, S. (2012), Les enquêtes PISA. Paris: PUF. Felouzis, G. and Charmillot, S. (2013), ‘School Tracking and Educational Inequality: A Comparison of 12 Education Systems in Switzerland’, Comparative Education, 49 (2): 181–205. Felouzis, G. and Perroton, J. (2007), Les « marchés scolaires »: une analyse en termes d’économie de la qualité. Revue française de sociologie, 48 (4): 693–722. Felouzis, G. and Perroton, J. (2009), Grandir entre pairs à l’école. Ségrégation ethnique et reproduction sociale dans le système éducatif français. Actes de la recherche en sciences sociales, 180 (5): 92–100. Fouquet-Chauprade, B. (2014), Bien-être et ressenti des discriminations à l’école. Une étude empirique en contexte ségrégué. L’année sociologique, 64 (2): 421–44. Fouquet-Chauprade, B. and Dutrévis, M. (2015), Les impensés des politiques d’éducation prioritaire: les dimensions non-­académiques. Education Comparée, 13, 187–200. François, J.-C. and Poupeau, F. (2009), Le sens du placement scolaire: la dimension spatiale des inégalités sociales. Revue française de pédagogie, 169, 77–97. Franquet, A., Friant, N. and Demeuse, M. (2010), (S’) orienter dans l’enseignement secondaire technique et professionnel en Communauté française de Belgique: la part du choix. L’orientation scolaire et professionnelle, 39 (4): 1–15. Gamoran, A. and Mare, R. D. (1989), ‘Secondary School Tracking and Educational Inequality: Compensation, Reinforcement, or Neutrality?’, American Journal of Sociology, 94 (5): 1146–83. Gillotte, O. and Girard, P. (2005), La sectorisation, l’affectation et l’évitement scolaire dans les classes de sixième à Paris en 2003. Éducation & formations, 75, 137–49.

School Segregation in France

43

Givord, P., Guillerm, M., Monso, O. and Murat, F. (2016), Comment mesurer la ségrégation dans le système éducatif? Une étude de la composition sociale des collèges français. Éducation & formations, 91, 21–51. Korsu, E. (2004), L’évitement scolaire par déménagement. Ville-école-­intégration Diversité, 139, 107–18. Lorcerie, F. (2003), L’effet “outsider”. En quoi l’ethnicité est-­elle un défi pour l’école? VEI Enjeux, 135, 86–102. Lorcerie, F. (2010), Le faux départ des ZEP. In S. Broccolichi, C. Ben Ayed and D. Trancart (Eds.), Ecole: les pièges de la concurrence. Comprendre le déclin de l’école française (pp. 37–56). Paris: La Découverte. Lucas, S. R. and Berends, M. (2007), ‘Race and Track Location in U.S. Public Schools’, Research in Social Stratification and Mobility, 25 (3): 169–87. Ly, S. T., Maurin, E. and Riegert, A. (2014), La mixité sociale et scolaire en Île-­de-France: le rôle des établissements. Rapport IPP n°4. Ly, S. T. and Riegert, A. (2015), Mixité sociale et scolaire et ségrégation inter- et intraétablissement dans les collèges et lycées français. Paris, CNESCO. Maresca, B. (2003), Le consumérisme scolaire et la ségrégation sociale dans les espaces résidentiels. Cahier de recherche du CREDOC n°184. Merle, P. (2010), Structure et dynamique de la ségrégation sociale dans les collèges parisiens. Revue française de pédagogie, 170, 73–85. Merle, P. (2011a), Concurrence et spécialisation des établissements publics et privés. Revue française de sociologie, 52 (1): 133–69. Merle, P. (2011b), La carte scolaire et son assouplissement. Politique de mixité sociale ou de ghettoïsation des établissements? Sociologie, 2 (1): 37–50. Merle, P. (2012), Éducation prioritaire. Cinq principes pour une refondation. La vie des idées. Meuret, D. (1994), L’efficacité de la politique des zones d’éducation prioritaire dans les collèges. Revue française de pédagogie, 109, 41–64. Moisan, C. and Simon, J. (1997), Les déterminants de la réussite scolaire en zone d’éducation prioritaire. Paris: La Documentation Française. Mons, N. (2007), Les nouvelles politiques éducatives. La France fait-­elle les bons choix? Paris: PUF. Noiriel, G. (2016), Le Creuset français. Histoire de l’immigration (XIXe–XXe siècle). Paris: Points. Oakes, J. (1985), Keeping Track: How Schools Structure Inequality. New Haven and London: Yale University Press. Oberti, M. (2005), Différenciation sociale et scolaire du territoire: inégalités et configurations locales. Sociétés contemporaines, 59–60 (3–4): 13–42. Oberti, M. (2007), L’école dans la ville: ségrégation–mixité–carte scolaire. Paris: Presses de Sciences-Po. Oberti, M. and Barthon, C. (2000), Ségrégation spatiale, évitement et choix des établissements. In A. Van Zanten (Ed.), L’École, l’état des savoirs (pp. 302–10). Paris: La Découverte. Paty, D. (1981), Douze collèges en France: le fonctionnement réel des collèges publics. Paris: La Documentation Française. Pinçon, M. and Pinçon-Charlot, M. (2009), Sociologie de la bourgeoisie. Paris: La Découverte. Prost, A. (2013), Du changement dans l’école. Les réformes de l’éducation de 1936 à nos jours. Paris: Seuil. Robert, B. (2009), Les politiques d’éducation prioritaire. Les défis de la réforme. Paris: PUF.

44

Understanding School Segregation

Schnepf, S. V. (2002), ‘A Sorting Hat that Fails? The Transition from Primary to Secondary School in Germany’, Innocenti Working Paper No. 92. Florence: UNICEF Innocenti Research Centre. Simon, P. (2008), Les statistiques, les sciences sociales françaises et les rapports sociaux ethniques et de ‘race’. Revue française de sociologie, 49(1), 153–62. Trancart, D. (2000), L’enseignement public: les disparités de l’offre d’enseignement. In A. Van Zanten (Ed.), L’école: l’état des savoirs (pp. 54–62), Paris: La Découverte. van Zanten, A. (2012), L’école de la périphérie. Scolarité et ségrégation en banlieue. Paris: PUF. van Zanten, A. and Obin, J.-P. (2008), La carte scolaire. Paris: PUF. Vasconcellos, M. and Bongrand, P. (2013), Le système éducatif. Paris: La Découverte.

3

Structural and Systemic Dimensions of School Segregation in French-­speaking Belgium Vincent Dupriez, Samir Barbana and Marie Verhoeven

Notwithstanding more than a century of policies aimed at school democratization, the education system in French-­speaking Belgium is characterized by the persistence of sharp social inequalities in outcomes. These inequalities reflect considerable phenomena of segregation observed between streams and between schools (Lafontaine, 2005; Lafontaine et al., 2003; Jacobs et al., 2009). The structural causes of this situation are well documented in the literature (see in particular Lafontaine and Monseur, 2011; Dumay et  al., 2010; Danhier et  al., 2014) and researchers generally agree that segregation between school environments can be attributed in large part to the combined effects of a quasi-­market regulation of the system and a considerable institutional differentiation of educational provision. Following a brief presentation of the definition of school segregation that will be used in this chapter, we review these two main mechanisms and their political and socio-­historical roots. We then describe several indicators of social, academic (i.e. according to students’ educational attainment) and ethnic segregation in the education system in French-­speaking Belgium and bring out how, paralleling the spread of such indicators, the concept of segregation has gradually entered public debate on the school. In the final section we address the policies that have been adopted to combat school segregation and the accompanying social inequalities.1

1.  What is school segregation? Concepts and definitions Speaking of school segregation is hardly an innocent choice, given the complex, multi-­ dimensional nature of the concept, as well as its normative and political ‘charge’. For purposes of situating the definition used in this chapter, we shall limit ourselves here to a few quick remarks on the dimensions and contours of the phenomenon. It should be remembered first of all that the literal meaning of segregation is simply ‘an action of separating or driving out’ (Grafmeyer, 1994). But more fundamentally, segregation can also designate ‘a body of deliberate norms and procedures aimed at This overview has been made possible by financial support from the F.R.S.-FNRS Fund for Scientific Research (Wallonia-Brussels Federation) within the framework of FRFC contract no. 2.4579.11.

1

46

Understanding School Segregation

preventing certain kinds of contact, especially mixtures or confusions’ (Brun, 1983, p. 76). In this case, it goes along with the idea of unequal power relations between social groups, where some groups have the power to exclude others from a certain number of spaces or advantages in order to maintain a privileged access for themselves. The literature thus tends to reflect a fluctuation between two definitions of segregation. The first, which is more descriptive, designates a ‘simple inequality in the dividing up of the groups or their behaviours, an inequality resulting in a more or less clear, more or less readable appreciation within the urban landscape’, while the second identifies segregation with an ‘exclusion’ expressing a more general, profound discrimination that is rigorous and sometimes institutionalized’ (Roncayolo, 1994, pp. 13–14). In the educational field, Delvaux proposes a fairly descriptive definition of segregation as ‘the translation of social differences in space. It manifests itself when individuals, classified in distinct social categories by the society and endowed with a differentiated social value, are separated in space and thus led to little interaction’ (Delvaux, 2005, p. 276). Measuring segregation would then entail comparing the actual situation of students having diverse characteristics with a fictive situation that would guarantee a random distribution of these characteristics within a reference population. This rationale underpins most dissimilarity indexes constructed on the basis of international databases. However, some researchers (e.g. van Zanten, 2001) go further by taking the normative dimensions of segregation into account by foregrounding the unequal social relations on which segregation is based (the underlying processes of social or institutional segregation), but also its negative effects for the populations thus segregated, such as cultural or economic marginalization or unequal access to legitimate resources, recognition and so on. In addition, documenting segregation scientifically also presumes choosing one (or several) scale(s) of observation (e.g. school, local area) with regards to the phenomenon and fixing the criteria used for measuring segregation. In this chapter, we shall attempt to document segregation in the descriptive sense of the term, paying particular attention to the academic (difference in attainment levels between students), social and ethnic dimensions of the phenomenon. But we shall not ignore the structural and normative aspects of the question, which will be addressed by examining the historical, structural and cultural mechanisms which, within French-­ speaking Belgium’s education system, can be viewed as factors of exclusion leading to an unequal access to educational resources.

2.  The sources of segregation: the construction of Belgium’s education system 2.1  Deeply rooted social and institutional separations In Belgium, the school was historically constituted less as an institution intended to integrate young people as a whole into a common educational project (as was the case in France, for example) than as the juxtaposition of educational programmes differentiated according to socially segmented populations.

Structural and Systemic Dimensions of School Segregation in Belgium

47

The Belgian education system2 rapidly came to be structured around schooling networks rooted in sociologically distinct communities (Catholic and liberal/secular). In order to understand this system, it is necessary to recall the way in which relations between the state and civil society in Belgium were historically constructed over time. Following the Belgian Revolution of 1830, an objective alliance was formed between liberals and Catholics. While the preceding rulers, successively Austrian, French and Dutch, had limited the Church’s role in education, the independent Belgian state opted to promote freedom in this domain, notably by allowing congregations, parishes and dioceses to open schools. As of 1830, the new political power thus guaranteed freedom of speech, freedom of the press and freedom of education. The first Belgian Constitution (1831, art. 17) formalized freedom of education and introduced a twofold principle still in effect today: freedom of educational provision and freedom of school choice (for the male household heads). In the area of education as in other fields, a ‘weak’ state was thus established. On the whole, educational initiatives were left to local powers and religious authorities, who were free to open schools and assume the role of ‘organizing power’ (i.e. the education authority). Whether public or private, these education authorities were responsible for staff recruitment, the definition of curriculum content and student evaluation. It was only later that the state gradually came to be involved in educational questions, most often in exchange for public resources allocated for school operations. In more general terms, Belgian society is profoundly marked by the historically conflict-­ridden relations between Catholic and secular organizations, which has led to the creation of a system of so-­called ‘consociational democracy’3 (Lijphart, 1977; Dumont, 1999). This model is based on a state with limited powers that oversees public life in close cooperation with representatives of the secular and Catholic milieus, which are responsible for internally structuring the main institutions and services in many sectors such as health, social services, culture and education. In the case of education, by the end of the nineteenth century, this double historical legacy (a relatively weak public authority and hostile inter-­community negotiations) was to feed a school war between these communities, each of which was endowed with its own collective social identity and conceptions of education. In the late 1950s, this ‘school war’ gave rise to a historic compromise (the School Pact), which in turn consolidated the principle of family school choice and a government commitment to fund all the schools, whether public or private, so that families opting for a private institution (most often Catholic) would not be financially penalized. Since then, the educational landscape has been clearly structured around three main groups (known as ‘networks’): public schools coordinated by the central government (which, for the French-­speaking part of the country, is now the French-­speaking Community From 1830 to 1988, the Belgian state was responsible for the entire education system. Since 1 January 1989, each of the linguistic and cultural communities controls its own education system. This article focuses on education in the French-­speaking part of the country, which came under the responsibility of the French-Speaking Community of Belgium in 1989. 3 Lijphart considers that consociationalism is an alternative form of democracy observed in segmented societies in which the power is split between elites and institutions linked to different autonomous society’s sectors or ‘pillars’. 2

48

Understanding School Segregation

of Belgium); public grant-­aided schools (organized by the municipalities and provinces); and the private grant-­aided schools, which are essentially Catholic. The schools of the grant-­aided networks are organized within federations, the most powerful of which is the General Secretariat for Catholic Education, responsible for representing and defending the interests of the Catholic schools. In other words, in Belgium, the force of the freedom of education principle (freedom of provision and freedom of choice) is deeply rooted in a specific sociological and historical context. As in England or Chile, Belgian families can choose their school. But this freedom of school choice has never been explicitly defended in the name of an educational market or competition between schools, but rather in that of the household head’s freedom to select an educational environment consistent with his or her philosophical convictions.

2.2  Educational quasi-­market and school segregation Nonetheless, in recent decades, the legal recognition of this double freedom has given rise to a more liberal version of parental school choice. Today, families are more likely to justify this choice on criteria that are educational or personalized (i.e. the search for a school environment adapted to each child’s abilities and preferences) than on a strictly ‘denominational’ basis. Following Vandenberghe (1998), a large number of researchers consider that in Belgium as well, within each of the three school systems – French-, Dutch- and German-­speaking – it is legitimate to speak of the presence of an educational quasi-market. Here, it is worth reminding that the term ‘educational quasi-­market’ is generally taken to refer to a form of education system governance including the following three features: (1) for the families, parental school choice; (2) for the schools, the possibility of differentiating and defining their educational provision; and (3) funding of the schools based on the number of students registered. Many studies have drawn attention to the way that this type of regulation constitutes a space fostering the differentiation of educational provision and the (re)production of forms of educational self-­ segregation. Delvaux and Joseph (2003) thus demonstrate how, within local competition spaces, the quasi-­market results in social and symbolic hierarchies between schools, accompanied by flows of students between ‘donor’ and ‘recipient’ schools. This situation creates what Gewirtz et  al. (1995) call ‘circuits of schooling’, namely statistically recurrent flows of students between schools that are clearly identified by stakeholders in the education field. Moreover, it has been shown that this context combining the historical autonomy of schools and parental choice underlies significant differences between schools, in particular with regards to the actual curriculum and student performance.‘Competitive interdependencies’ between schools lead each of them to develop their own ‘educational niche’, defined as ‘the social and educational outcome of a process of interaction between a school and its public, in an overall context of competition between schools’ (Dupriez and Cornet, 2005, p. 181). In other words, by adapting to its environment, the school is led to develop its own culture or specific organizational identity (Draelants and Dumay, 2011). As a result, segregation between schools is closely associated with

Structural and Systemic Dimensions of School Segregation in Belgium

49

the educational quasi-­market and the inequalities between schools structuring it (Maroy and Delvaux, 2008). The creation of educational niches in fact facilitates peaceful coexistence between differentiated schools and thus validates a form of division of educational labour tied to the social distribution of the school population (Dupriez and Wattiez, 2016).

2.3  The legacy of a ‘differentialist’ conception of education Along with the historic divide opposing Catholic and secular communities (until the early twentieth century at least), a social divide emerged on the basis of the implicit representation that each social class should have a specific form of education. This vision was concretely reflected in education systems compartmentalized according to social groups: primary school was thus traditionally conceived as the ‘people’s school’, while the middle and secondary schools were in principle intended for the children of the bourgeoisie and the elites, with their own preparatory sections (Grootaers, 1998). There is probably a link between this legacy and the differentialist conception of education which still prevails in Belgium despite numerous political schemes intended to move towards the construction of a ‘common’ school free of such divisions. Indeed, the political authorities decided in 1924 that study programmes in primary education (ages 6 to 12 years) should coincide, regardless of where pupils attended school. After World War II, different initiatives were undertaken in order to reduce differences between secondary education programmes, which, at that time, imposed strict separations between students depending on their education and training choices. With the adoption of a major law on the ‘reform’ of secondary education in 1971, students beginning secondary school no longer chose between general, technical or vocational programmes; regardless of the school, they entered a two-­year observation cycle. A largely common curriculum was aimed at permitting all students to try their hands at a variety of disciplines (within general education, but also artistic and technological subjects), conceived as test activities preparing each student for a positive programme choice as of the third year. This law on secondary education reform clearly corresponded to a more ‘inclusive’ school, on the whole offering the same curriculum to all students up to the age of 14 years. It should be noted, however, that this new programme was not adopted by all schools and that where it was adopted, it nonetheless envisaged a specific pathway, outside of this common curriculum, for all students who had not successfully completed their primary studies. In the French-­speaking part of the country, this was the case for 5–10 per cent of the students, who were thus not concerned by the common curriculum. In general, after one year of secondary education, these students were directed towards a vocational training programme in a vocational school. Since the early 2000s, there have been several reforms of this lower secondary cycle but in the end, they all reflect the same ambiguity: the political and educational authorities’ desire to provide a school that is almost entirely common until the age of 14 comes up against the inability of the education system and its stakeholders to handle students with uneven levels of motivation and performance within such a common curriculum. In the end, each time a greater degree of integration for all students is

50

Understanding School Segregation

considered, the system opens a back door (differentiated lower secondary cycle, selection for specialized education, increased grade repetition, etc.) which permits different treatment of the students who are the most distant from school culture. More broadly, some authors point to a persistent ‘anti-­heterogeneity culture’ in French-­speaking Belgium, ‘steeped in the principle that it is impossible to educate all children together’ and inseparable from a ‘culture of grade repetition’ based on the conviction that it is more ‘profitable’ in educational terms to create the most homogeneous classes possible around students’ ‘attainments’ and ‘talents’ (Lafontaine, 2005). In this respect, Belgium is distinguished by its particularly high grade-­repetition rates.4 Crahay and Delhaxhe (2004) propose the term ‘culture of differentiation’ to describe education systems, including those in Belgium, where multiple processes, ranging from teachers’ beliefs to concrete schemes, are at work to separate students prematurely in the course of their education. In the case of French-­speaking Belgium, this involves a high use of grade repetition in primary and secondary schools alike, significant channelling of students towards special education programmes and the separation of students into sharply distinct streams by the third year of secondary school. In short, most specialists of education in Belgium suggest that the educational quasi-­market system and emphasis on differentiation are, along with the residential segregation, the main factors responsible for school segregation.

3.  The emergence of indicators and the concept of segregation in public debate In this section, we attempt to trace the gradual emergence of the segregation issue in public debate on education, in particular through scholarly research and quantitative data that have permitted an assessment of differences between schools in relation to their populations. We also consider the criteria that have been used to bring out certain kinds of segregation within the education system. This section is divided into three parts. In the first, we underline the absence of references to segregation until quite recently. We then recall the way the processing of data taken from the first PISA (Programme for International Student Assessment) surveys contributed to the characterization of the education system as a space of social and academic segregation. In the final part, we single out complementary analyses contributing to a better understanding of the dimensions of segregation.

3.1  The absence of references to segregation To begin with, it must be emphasized that the concept of segregation is recent in debates on education in Belgium. This is not to say that segregation did not exist before. According to the Wallonia-Brussels Federation’s Education Indicators for 2013, one out of ten pupils was already one year behind at the end of the first year of primary school and more than 20 per cent repeated at least once during their primary schooling. By the third year of secondary school, more than 50 per cent of the students were at least one year behind.

4

Structural and Systemic Dimensions of School Segregation in Belgium

51

On the contrary, we have already pointed to the double cleavage – confessional and social – that structured Belgian education for many years. However, this double divide was largely taken for granted, and the term ‘segregation’ was rarely used to characterize the blatant separation of school populations. Historically, the semantics of equal opportunity (and the right to education for all) were more often invoked to foster progressive educational measures and arguments for the development of the common school. In a fundamental text like that of the 1959 School Pact, there is no mention of the implications of segregation, much less the promotion of social diversity. Rather, the division between the three school networks is to some extent justified in terms of a ‘non-­discrimination’ principle, according to which each pupil should have access to an education corresponding to his or her convictions. Similarly, the arguments in favour of the 1971 secondary education reform law focused above all on the idea of equal opportunity and the development of all talents. The segregation of school publics through programme choices met with no condemnation. When education became the responsibility of the three linguistic communities (Flemish-, French- and German-­speaking), the political authorities of French-­speaking Belgium defined (in an important Decree or Educational Act published in 1997, the so-­called ‘Missions Decree’) four priority missions for the education system. One of these was to ‘provide all pupils with equal opportunities for social emancipation’, a choice of wording that echoed preliminary discussions emphasizing the extent of social inequalities at school and the dependence of educational pathways on the parents’ qualifications and professions. Another decree was then adopted in 1998 in order to ‘provide all pupils with equal opportunities for social emancipation, in particular through the implementation of positive discrimination’. It is important to note the compensatory rationale that underlies this decree: in order to guarantee equal opportunities for emancipation, the education system should provide more resources for schools with student bodies from disadvantaged sociocultural backgrounds. However, the decree did not question the fact that many students from disadvantaged backgrounds were grouped together within the same schools.

3.2  Segregation measured on the basis of PISA data The arrival of the PISA surveys in the early 2000s was to allow the systematic documentation of student population differences between schools and facilitate comparisons across countries and education systems. It was also during this same period that the term ‘segregation’ gradually entered debates on the school. Academic researchers to begin with (e.g. Lafontaine et al., 2003), followed by the public authorities (Contrat pour l’École, 2005; Indicateurs de l’enseignement, 2006), drew attention to different forms of inequality and segregation within the education system in French-­ speaking Belgium. One of the first significant documents addressing this issue was the Liège University report (Lafontaine et al., 2003) on the PISA 2000 outcomes of French-­speaking Belgian students. In addition to the students’ performance level, which was close to the average for the OECD as a whole, researchers firmly stressed the extent of the differences in scores between students:

52

Understanding School Segregation What is striking in the French-­speaking Community of Belgium . . . more than the average, is the extent of the spread of the outcomes. With a standard deviation of 111, the French-­speaking Community of Belgium is, along with Germany, the education system where the heterogeneity of the performances is most pronounced. Lafontaine et al., 2003, p. 55

Adopting an argument fairly close to the one presented above, the University of Liège researchers attributed the considerable disparity in scores to the functioning of the education system: Without too much risk of being mistaken, we might argue that the disparity observed in performances is related to the way our education system is structured . . . a high rate of grade repetition, hierarchical organisation of streams, significant disparities between schools [and] de facto segregation by social and ethnic backgrounds all contribute to standardising student groups. Lafontaine et al., 2003, pp. 55–6

The report indicates as well that the majority of the low achievers come from socially and culturally disadvantaged families. More specifically, the authors argue that: Of all the education systems across the countries participating in the PISA surveys, the French-­speaking Community of Belgium shows the highest impact of parental socio-­professional status on students’ reading scores. A student in French-­speaking Belgium whose parents exercise a profession at the bottom of the income scale is more likely to figure among the lowest 25% in reading than is the case elsewhere. Lafontaine et al., 2003, p. 85

Finally, the authors emphasize the extent to which differences between pupils’ scores largely depend on differences between schools: In the French Community, the proportion of variance between schools is very high (56.2%) and the same is true in Flanders (54.6%). This means that Belgium’s education system shows one of the highest degrees of variance between schools. This proportion is also high in Austria (60%) and Germany (59.8%), as well as Hungary (67%), Poland (63%) and The Czech Republic (53.4%). The lowest degree of variance is found in the Nordic countries (7–20%). Lafontaine et al., 2003, p. 162

Later studies based on secondary analyses of the PISA 2000 and 2003 databases also highlight these differences between schools, often measured by dissimilarity indicators. Dupriez and Vandenberghe (2004) thus bring out that, on the basis of an attainment level indicator (a score below the first quartile of the distribution of the education system concerned), it would be necessary in French-­speaking Belgium to transfer 61 per cent of 15-year-­old students, in order to obtain an equitable distribution of low-­ achievers across all schools. If we consider the most socio-­economically disadvantaged pupils, 45 per cent of them would have to be transferred. In both cases, French-­speaking

Structural and Systemic Dimensions of School Segregation in Belgium

53

Belgium’s education system is, along with that of Germany, the most segregated among the OECD countries participating in the PISA 2000 survey. Two years later, an official publication of the General Administration for Education in French-­speaking Belgium took note of this appraisal and thus helped to circulate it within the education system. The Education Indicators (2006) thus affirmed that ‘Among the European countries participating in PISA, the French-­speaking Community is one of the education systems where segregation between schools by socio-­economic characteristics and student attainment level is the most pronounced’ (Ministère de la Communauté française de Belgique, 2006, p. 64). Among other uses of PISA databases to establish segregation in the education system (and often to denounce it), the significant recourse to such material by non-­ governmental organizations is also worth mentioning. The most active in this respect is probably the Appel Pour une Ecole Démocratique (Call for a Democratic School, APED), which began to produce its own secondary analyses as soon as the results of the PISA 2000 survey were released (2003). The goal of this initiative is not only to denounce existing segregation but to demonstrate its relationship to the dynamics of school markets and recourse to early selection of secondary education streams. Along with other organizations, academics and the main teachers unions, the APED is also active within the Plate-­forme de Lutte contre l’Échec scolaire (Platform against School Failure); since its founding in 2003, this coalition has actively criticized school segregation, unequal opportunities in school access and the resulting educational and social exclusion. These repeated appeals, coming from the associative sector and the trade unions in particular, have helped to establish various policies for reducing inequalities that we shall consider in the final section of this chapter.

3.3  Three additional segregation-­related issues: ethnicity, staff and tracking The publications we have cited here, mainly based on data from international surveys, allow an assessment of the high level of social inequalities in French-­speaking Belgium’s education system and the equally high level of school segregation. What emerges most clearly is the degree of academic (i.e. student attainment level) and social segregation. Several publications (Lafontaine et al., 2003; Demeuse et al., 2005; Dupriez and Dumay, 2006) link these two phenomena, explaining, at least in part, the extent of social inequalities in the area of schooling by the high level of segregation as a source of major differences in learning opportunities offered to students (see also Dupriez et al., 2008). Beyond this clear trend, we now address three additional segregation-­related issues that have also been measured and analysed: the ethnic dimension of segregation, differences in teaching staff according to the school attended and differences between educational tracks (or streams) at secondary level.

Ethnic segregation and inequalities A vivid debate is raging among scholars about the nature and amplitude of ethnic segregation within the French-­speaking school system, particularly regarding its

54

Understanding School Segregation

interaction with socio-­economic individual factors and its embedment in general structural patterns of segregation (see e.g. Lafontaine and Monseur, 2011; Hirtt, 2014; Danhier et al., 2014). Although it is difficult to draw unambiguous conclusions from this complex debate, several key elements might be pinpointed. First, considering the available data on the school system as a whole (unfortunately based on nationality, not on more complex ways of identifying ethnicity), ethnic-­based segregation appears to be less pronounced than academic or social segregation (Ministère de la Communauté française de Belgique, 2006). However, the mapping of these existing data reveals obvious ethnic segregation patterns, as ethnic minority students are often concentrated in the most deprived areas, especially in urban contexts. This observation is particularly true for extraEuropean Union and ‘postcolonial’ minorities such as Moroccan and Turkish groups or sub-Saharan African groups (Delvaux et al., 2007; Delvaux and Serhadlioglu, 2014). This spatial ethnic segregation is largely reflected in schooling segregation, which has been constantly demonstrated since the first studies on migrant students schooling (see Ouali and Rea, 1994, Jacobs and Rea, 2007, Delvaux et al., 2007). However, school segregation cannot be interpreted as a simple reflect of urban segregation (Delvaux et al. 2007, Delvaux and Serhadlioglu, 2014), as in many cases, the indicators of (social or ethnic) segregation between schools tend to be sharper than those of urban segregation. This might be partly explained by contrasting family strategies in a quasi-­market context, tending to reinforce social and ethnic grouping; however, this has much to do with the phenomenon of creation of educational niches (cf. supra) leading to forms of coupling between differentiated school contexts and differentiated school populations (Verhoeven, 2011). Finally, the obvious interaction between socioeconomic and ethnic factors in our context makes the explanatory debate very complex. Whereas considerable differences between the performances of ‘native’ students and those from immigrant families (Jacobs et  al., 2007, 2009; Danhier et  al., 2014) are systematically observed,5 most studies show that these differences tend to become statistically insignificant once controlled by family professional status or educational level. In other words, the relative weakness of the outcomes of students from immigrant backgrounds has more to do with their socio-­economic status than their migrant origin, although some researchers recognize a specific effect tied to the language spoken at home (Crutzen and Lucchini, 2007; Lafontaine et al., 2003) or to ethnic stigmatization in orientation (Ouali and Rea, 1994; Delvaux et al., 2007).

Segregation and teaching staff As has been the case in other countries, in particular in the United States (Clotfelter et  al., 2005), Belgian researchers have studied the relationship between the type of For instance, PISA 2012 data analysis still reveal important attainment gaps in mathematics or French between ‘native’ students (born in Belgium to Belgian parents), ‘first-­generation’ students (born abroad) and so-­called ‘second-­generation’ students (born in Belgium to two parents born abroad) (Danhier et al., 2014). The same research report indicates that in 2012 the proportion of students attaining minimal skills in math (Level 2 as defined by PISA) was lower among students from foreign backgrounds than among native-­born students (ibid., p. 36).

5

Structural and Systemic Dimensions of School Segregation in Belgium

55

students enrolled in a school and the characteristics of the teaching staff working there. Delvaux et al. (2013) thus show that the most disadvantaged schools (with regards to the students’ socio-­economic level) are also those with the greatest number of beginning teachers (less than five years of professional experience). This relationship is especially strong in the Brussels region, where the proportion of beginning teachers working in the 25 per cent most disadvantaged schools is nearly 50 per cent higher than those working in the 25 per cent most privileged schools. A recent McKinsey report (2015), carried out at the request of the Wallonia-Brussels authorities, indicates that one in five teachers (19.2 per cent) working in the most disadvantaged schools does not have an educational qualification preparing them for the teaching profession. This rate is only 10 per cent in the 25 per cent most privileged schools (McKinsey and Company, 2015, p.  110). Dumay (2014), meanwhile, demonstrates that the socio-­ economic level characterizing the students of a given school shows a significant covariation with the teacher turnover rates. In other words, the schools with socio-­ economically disadvantaged student bodies are more exposed to teaching staff rotation than the others. This statistically significant relationship is stronger in basic education than at secondary level. These different data on teaching staff thus draw attention to the objective characteristics penalizing schools with the most disadvantaged student bodies. Such a situation does not result from any explicit, organized desire on the part of school stakeholders. Rather, in the absence of a centralized regulation of teaching career management, the sum of individual choices made by teachers and school administrations leads to the results presented here.

Segregation between streams The PISA surveys also bring out the extent to which the students’ programme choices (at age 15) are associated with the parents’ qualification level, and that of the mother in particular. Here, the results are unambiguous: the higher the mother’s qualification, the greater the probability of pursuing a general stream (Dupriez and Verhoeven, 2007). In addition, Monseur and Lafontaine (2012) stress that French-­speaking Belgium has one of the highest correlation rates between the student’s socio-­economic background and the probability of being in a programme leading to a qualification, even when the student’s academic performance is taken into account. Research drawing on the socio-­economic indicator proper to the education system in French-­speaking Belgium (Indicateurs de l’enseignement, 2015) also brings out considerable segregation between streams. For example, the disparity between the branches of secondary education emerges from the beginning of lower secondary, with a large gap between the mean socio-­economic indicators of the common first cycle (attended by the large majority of students aged 12–14 years) and that of the so-­called ‘differentiated’ first grade (for students who do not have their primary school certificate). This gap widens in the second and third cycles, when general education becomes the stream for the most privileged students, followed by the technical streams and, finally, the vocational programmes.

56

Understanding School Segregation

4.  School segregation and public policy In this final section, we review the policies intended to act on school segregation and social inequalities. These appear at the end of the 1980s, when the French-­ speaking community was mandated to supervise education in French-­speaking Belgium. Even if the reforms we analyse here are all committed to addressing inequalities between students, they do not explicitly confront segregation. As indicated above, this concept is fairly recent in Belgium and was only introduced as a means of defining the reality of the education system in the early 2000s, in the wake of what can be described as a ‘PISA shock’ (Niemann, 2015). Moreover, the reforms emerge within a political context where successive Ministers of Education – nearly all situated at the left of the political spectrum – often invoked the argument of reducing inequalities in order to legitimize their actions. This argument, and later that of ‘desegregation’, were also present in the discourse developed by a multitude of school stakeholders with their own respective agendas: the OECD, grass-­roots activists, industry professionals and researchers (Mangez, 2011; Draelants, 2009). In such a context, two categories of policies can be identified. The first may be described as compensatory. Like the positive discrimination briefly mentioned above, these policies borrow their cognitive and normative content from abroad, above all the United States (Rochex, 2010). In French-­speaking Belgium, three types of compensatory policies have emerged, as we shall explain in greater detail below. A second group, which we would qualify as structural, is aimed not at distributing resources according to the school population concerned but rather at reforming the structure of the system in order to reduce segregation and related inequalities.

4.1  Three compensatory policies As elsewhere in Europe, these compensatory policies – inspired by the ‘affirmative action’ first introduced in the United States in the early 1960s – emerge in a context where studies revealing the mechanisms by which the school system transforms social inequalities into educational inequalities had been widely disseminated (Grootaers, 1998; Beckers, 1998). In French-­speaking Belgium, the ethnic dimension of segregation is often dissociated from the socio-­economic dimension. The compensatory policies thus targeted the poorest categories of the school population without ever referring to their ethnic status. In this respect, such policies must be distinguished from those applied in the USA, where affirmative action reforms were related to larger ethnic issues (Bickel, 1998) in addition to the struggle against poverty (Frandji et al., 2011; Demeuse, 2002). In French-­ speaking Belgium, the compensatory policies inspired by affirmative action were to take two distinct forms: first of all, the definition of Priority Education Areas (Zones d’éducation prioritaires or ZEPs) and then the establishment of positive discrimination. The latter was subsequently adapted to give rise to ‘differentiated supervision’ (encadrement differencié).

Structural and Systemic Dimensions of School Segregation in Belgium

57

Priority Education Areas (ZEPs) French-­speaking Belgium’s ZEPs were created in 1989. They corresponded to territories characterized by their socio-­economic and educational precariousness, where public authorities sought to encourage educational, cultural and social initiatives benefiting the students there and, more generally, the neighbourhood as a whole. In order to carry out these socio-­educational initiatives, a wide variety of local projects was implemented in ZEPs areas (e.g. homework assistance programmes, intercultural activities). Beyond the influence of affirmative action logics, the ZEP reform was inspired by a nearly identical policy implemented eight years earlier in France, where the ZEPs promoted a kind of territorialization of education in order to respond to major differences in academic success from one area to another. In France, these policies were also closely linked with urban planning initiatives aimed at giving new life to certain poor neighbourhoods (de Villers, 2010) and thus furthered comprehensive projects between schools and neighbourhoods. In French-­speaking Belgium, the ZEPs also attempted to support such collaborations between the schools and other neighbourhood stakeholders, but they remained totally independent of urban planning policies. In contrast to France, moreover, there is no school mapping in Belgium, and this situation was to raise questions about the beneficiaries of the resources allocated to the ZEPs. Certain schools with privileged student bodies were located in neighbourhoods targeted by the ZEPs, while schools attended by disadvantaged students were sometimes situated outside the ZEPs and thus did not have access to the latter’s resources. Indeed, this was one of the criticisms raised with regards to the ZEPs, which were gradually abandoned by the political authorities. Depending on the territories and dimensions evaluated, measurements and assessments of these Priority Education Areas have brought out uneven results. Available information does not permit a clear position on either the ability of these policies to reduce segregation (Demeuse, 2002) or their influence on students’ academic performances.

Positive discrimination As already mentioned in this chapter, positive discrimination was established in French-­speaking Belgium following the Missions Decree (1997), which sought to offer each student equal opportunities for social emancipation. The aim was different from that of the ZEPs: positive discrimination was intended above all to support educational activities within the schools. Accordingly, following the 30 June 1998 decree establishing positive discrimination, increased means were allocated to the schools in two forms: subsidies and additional staff. In order to identify the schools eligible for these means (12–13 per cent of the total), an aggregated socio-­economic index was attributed to each student on the basis of a series of indicators related to each student’s home neighbourhood. The quantitative indicators produced in support of this policy were thus intended to identify the schools with the most disadvantaged student bodies rather than the schools located in the most disadvantaged neighbourhoods.

58

Understanding School Segregation

Differentiated supervision On 30 April 2009, a new decree on differentiated supervision was issued with slightly broader purposes than the preceding one. Beyond the educational initiatives, civic aims were now targeted as well (e.g. reduction of violence). The number of schools potentially benefiting from the programme (25 per cent), its duration (five years rather than three previously) and its budget were also expanded. The decree was driven by the logic of empowerment: the schools are basically autonomous in the choice of the activities they organize with the additional resources, depending on their objective and local contexts. At the same time, the criteria for drawing up the project in order to obtain a subsidy have become more precise. It is now necessary to define the objectives to be achieved in terms of indicators, with fixed techniques for monitoring the teams’ ability to attain these objectives.

4.2  Structural policies The objective of this second category of policies is a (moderate) reform of the education system’s organizational structure and, more specifically, the educational quasi-­market system, whose role in the production of segregation has already been indicated above. Here we would note that the attempts to reform this system have not been accompanied by a reassessment of the larger principle – consociational democracy – underlying it in Belgium. These reforms were to address two dimensions of the quasi-­ market: freedom to propose a differentiated school provision and freedom of school choice. The third feature (basing the level of school funding on the number of students recruited) would never be the target of public policies, if not (very indirectly and with little force) through the positive discrimination and differentiated supervision measures.

Freedom to propose a differentiated offer: towards common school standards? The first policy analysed here attempted to act upon educational provision which traditionally relies on decentralized actors (the freedom of the local educational authorities) and reflects a highly differentialist view of education. In the 1990s, a scheme for gradually defining basic skills and standards was set up in consultation with the schooling networks and then applied, for the first time in French-­speaking Belgium, to all schools, regardless of their network. This process of defining standards and basic skills is similar to the standards-­based reforms movement found in other countries during the same period. In such systems, however, the reforms generally subscribed to a logic of accountability, which was not the case with the basic skills in French-­speaking Belgium at the time; rather, basic skills mainly reflected the aim of having common teaching standards for all the schools and all the networks (Mangez, 2011) with a view to equality among students. This initiative, undertaken by the centre-­left political majority, was supported by a large majority of school stakeholders.

Structural and Systemic Dimensions of School Segregation in Belgium

59

Several years later, the spread of non-­certifying external assessments was associated with a fairly similar political argument and supported by comparable coalitions. Ultimately, these assessments – whose results are confidential, under threat of heavy penalties – lent support to two objectives. On the one hand, they gradually provided flexible guidelines for teachers’ classroom practice; on the other, they generated information regarding the state of students’ knowledge and skills. As of 2009, the teachers and their schools were required to introduce certifying external assessments (once again subject to confidentiality). This development marked a major transformation in French-­speaking Belgium. Indeed, student evaluations had always been a prerogative of the teachers and the local education authorities. These certifying external exams are administered today at the end of the sixth and eighth years of compulsory education, and determine students’ promotion or grade repetition. Such a decision is thus no longer in the hands of local educational teams. In addition, the exams disseminate a common teaching model across the education system in a more binding way than was previously the case (Carette and Dupriez, 2009). They also limit schools’ recourse to grade repetition for certain students, which traditionally constituted a highly effective means of filtering and selection (Barbana and Dupriez, 2015).

Reforming school choice The second institutional characteristic coming under increased regulation in recent years is the handling of parental school choice. It should be recalled that in Belgium, public authorities traditionally do not regulate student enrolments. Each family contacts the school where it wishes to enrol its child and the rest of the procedure depends on a dialogue between the headteacher and the family concerned. An initial step illustrating the intervention of the public authorities on this point is found in the Missions Decree (1997), where two articles specify the legal and administrative elements that must accompany a student’s school enrolment and/or refusal: decision-­making criteria, time limits to be observed, procedure to be followed and, overall, the obligation for all schools to accept any student subscribing to the institution’s charter. The lack of an available place therefore became the only legitimate motive for refusing a student. However, the question of school choice emerged above all in the thick of the politico-­media debates of the mid–2000s. Some associations denounced an unequal treatment of families and the duplicitous strategies of school administrators who informed certain families that there were no longer any places available when such places remained open for other families. The Missions Decree was thus not respected everywhere. Following a broad consultation among education stakeholders (the Contract for the School), which was to advocate ‘the end of ghetto schools’, the Minister of Education proposed adjustments in the functioning of the venerable Belgian ‘School Choice’. A first decree (2007) thus introduced standards (in particular, a common date for the beginning of enrolments in all secondary schools) aimed at guaranteeing the ‘first come, first served’ principle (Demeuse and Friant, 2011). As a result, schools could no longer announce a year in advance (as had sometimes been the case) that there were no places available.

60

Understanding School Segregation

This decree, specifying the way in which school choice was to be implemented, was part of an explicit commitment to ‘desegregation’. However, it affected only one segment of the education system (enrolment in the first year of secondary education), and its intervention remained quite modest: by establishing the ‘first come, first served’ principle, it complicated the strategies of certain schools that were accustomed to selecting their student body. However, the decree is still based on the principle of free choice. Subsequent decrees (2008 and 2010) moderately redefined the terms of such a regulation and introduced criteria for giving priority to certain students, in particular on the basis of socio-­economic indicators. These limited attempts to place school choice within a more equitable framework gave rise to heated opposition from certain stakeholders drawing on particularistic, community-­centred arguments (Delvaux and Maroy, 2009) to defend the legitimacy of a differentiated educational provision at secondary level, notably according to differences in students’ talents and ambitions.

5.  Conclusion While the phenomenon of segregation is not new, the term itself has only gained ground in the literature on the education system in French-­speaking Belgium over the past fifteen years. This gradual arrival on the research agenda was supported by the dissemination of quantitative indicators, both national and international, confirming the extent of such segregation between schools, especially in secondary education. Research efforts generally converge, moreover, to indicate two major sources of segregation: on the one hand, the Belgian tradition of ‘free choice’ conducive to a kind of ‘self-­segregation’ within the education system and, on the other, a logic of differentiating students and learning pathways, which is present in primary school but even more evident at secondary level with its different education and training streams. Such an analysis associating inequalities in outcomes, differentiation between schools and structural mechanisms has to some extent been integrated into the programme of the centre-­left political parties who, for the past two decades, have almost invariably been responsible for education in French-­speaking Belgium. Over the last ten years, moreover, new reforms have been adopted with the objective of providing support for the schools with the largest proportions of students from disadvantaged backgrounds but also, more fundamentally, that of introducing moderate systemic transformations in order to regulate parental school choice and limit the mechanisms of pathway differentiation. Resistance to these reforms, coming from the general public, parents and a large number of educational stakeholders, nonetheless limits their effects (which are difficult to analyse and measure with the requisite accuracy). Such resistance illustrates how difficult it remains, despite empirical evidence leading to relatively precise assessments, to regulate highly institutionalized individual liberties in a historical, social and cultural context where individual rights are increasingly demanded and perceived as a key feature of modern societies.

Structural and Systemic Dimensions of School Segregation in Belgium

61

References Barbana, S. and Dupriez, V. (2015), Comment le pilotage par les résultats affecte-­t-il la coordination locale du travail? Le cas de l’enseignement secondaire belge francophone. In L. Progin, J.-F. Marcel, D. Perisset and M. Tardif (eds.), Transformation(s) de l’école: vision et division du travail (pp. 135–53). Paris: L’Harmattan. Beckers, J. (1998), Comprendre l’enseignement secondaire: évolution, organisation, analyse. Bruxelles: De Boeck. Bickel, R. D. (1998), ‘A Brief History of the Commitment to Inclusion as a Facet of Equal Educational Opportunity’, New Directions for Student Services, 1998 (83): 3–13. Brun, J. and Chauvire, Y. (1983), La ségrégation sociale: questions de terminologie et de méthode. Espace, populations, sociétés, 1 (1): 75–85. Carette, V. and Dupriez, V. (2009), La lente émergence d’une politique scolaire en matière d’évaluation des élèves. Quinze ans de transformations en Belgique francophone. Mesure et évaluation en éducation, 32 (3): 23–45. Clotfelter, C. T., Ladd, H. F. and Vigdor, J. (2005), ‘Who Teaches Whom? Race and the Distribution of Novice Teachers’, Economics of Education review, 24 (4): 377–92. Crahay, M. and Delhaxhe, A. (2004), L’enseignement secondaire inférieur: entre culture de l’intégration et culture de la différenciation. Paris: PIREF. Crutzen, D. and Lucchini, S. (2007), Etat des savoirs concernant l’éducation et la scolarité des enfants issus de l’immigration en Communauté française de Belgique. In M. Martiniello and A. Rea (eds.), L’Immigration et intégration en Belgique francophone. Etat des savoirs (pp. 287–332). Louvain-­la-Neuve: Academia Bruylant. Danhier, J., Jacobs, D., Devleeshouwer, P., Martin, E. and Alarcon, A. (2014), Vers des écoles de qualité pour tous? Analyse des résultats à l’enquête PISA 2012 en Flandre et en Fédération Wallonie-Bruxelles. Bruxelles: Fondation Roi Baudouin. Delvaux, B. (2005), Ségrégation scolaire dans un contexte de libre choix et de ségrégation résidentielle. In M. Demeuse, A. Baye, M.-H. Straeten, J. Nicaise and A. Matoul (eds.), Vers une école juste et efficace: vingt-­six contributions sur les systèmes d’enseignement et de formation: une approche internationale (pp. 275–95). Bruxelles: De Boeck. Delvaux, B., Desmarez, P., Dupriez, V., Lothaire, S. and Veinstein, M. (2013), Les enseignants débutants en Belgique francophone: trajectoires, conditions d’emploi et positions sur le marché du travail. Les Cahiers de Recherche en Education et Formation du Girsef, 92, 1–156. Delvaux, B. and Joseph, M. (2003), Les espaces locaux d’interdépendance entre écoles: étude de cas en Belgique francophone. Rapport de recherche: Changement des modes de régulation et production sociale des inégalités dans les systèmes d’éducation: une comparaison européenne. Louvain-­la-Neuve: Cerisis-UCL. Delvaux, B. and Maroy, C. (2009), Débat sur la régulation des inscriptions scolaires en Belgique francophone: où se situent les désaccords? Les Cahiers de Recherche en Education et Formation, 68, 3–25. Delvaux, B., Martiniello, M., Rea, A., Verhoeven, M., Devleeshouwer, P., Druetz, J., Herman, B. and Poncelet, A. (2007), Analyse des parcours scolaires des jeunes d’origine ou de nationalité étrangère en communauté française de Belgique. Bruxelles: Research report (ULB-ULG-UCL). Delvaux, B. and Serhadlioglu, E. (2014), La ségrégation scolaire, reflet déformé de la ségrégation urbaine. Différenciation des milieux de vie des enfants bruxellois. Les Cahiers de recherche du GIRSEF, 100, 3–112.

62

Understanding School Segregation

Demeuse, M. (2002), Analyse critique des fondements de l’attribution des moyens destinés à la politique de discrimination positive en matière d’enseignement en Communauté française de Belgique. (Doctoral dissertation, Université de Liège). Demeuse, M., Baye, A., Straeten, M.-H., Nicaise, J. and Matoul, A. (2005), Vers une école juste et efficace. 26 contributions sur les systèmes d’enseignement et de formation. Bruxelles: De Boeck. Demeuse, M. and Friant, N. (2011), ‘School segregation in the French Community of Belgium’, in J. Bakker, E. Denessen, D. Peters and G. Walraven (eds.), International Perspectives on Countering School Segregation (pp. 169–87). Apeldoorn: Garant Uitgevers NV. de Villers, J. (2010), Le traitement différencié des élèves de milieu populaire. Bruxelles: FAPEO. Draelants, H. (2009), Réforme pédagogique et légitimation. Le cas d’une politique de lutte contre le redoublement. Bruxelles: De Boeck. Draelants, H. and Dumay, X. (2001), L’identité des établissements scolaires. Paris: PUF. Dumay, X. (2014), Décrire et prédire le turnover des enseignants dans les établissements de la Fédération Wallonie-Bruxelles: Eléments d’analyse à partir de bases de données administratives. Les Cahiers de Recherche en Education et Formation du Girsef, 95: 1–47. Dumay, X., Dupriez, V. and Maroy, C. (2010), Ségrégation entre écoles, effets de la composition scolaire et inégalités de résultats. Revue française de sociologie, 51 (3): 461–80. Dumont, H. (1999), Le pluralisme « à la belge »: un modèle à revoir. Revue belge de droit constitutionnel, 1999 (1): 23–31. Dupriez, V. and Cornet, J. (2005), La rénovation de l’école primaire: Comprendre les enjeux du changement pédagogique. Bruxelles: De Boeck. Dupriez, V. and Dumay, X. (2006), ‘Inequalities in School Systems: Effect of School Structure or of Society Structure?’, Comparative Education, 42 (2): 243–60. Dupriez, V., Orianne, J.-F. and Verhoeven, M. (2008), De l’école au marché du travail, l’égalité des chances en question. Bern: Peter Lang. Dupriez, V. and Vandenberghe, V. (2004), L’école en Communauté française de Belgique: de quelle inégalité parlons-­nous?. Les Cahiers de Recherche en Education et Formation du Girsef, 27: 3–26. Dupriez, V. and Verhoeven, M. (2007), Du droit à l’éducation à l’égalité des résultats. In M. Frenay and X. Dumay (eds.), Un enseignement démocratique de masse: Une réalité qui reste à inventer (pp. 19–44). Louvain-­la-Neuve: Presses Universitaires de Louvain. Dupriez, V. and Wattiez, R. (2016), Niches éducatives, identités catégorielles et marchés scolaires. In H. Draelants and X. Dumay (eds.), Les écoles et leur réputation: L’identité des établissements en contexte de marché (pp. 85–102). Bruxelles: De Boeck. Frandji, D., Ayed, C. B., Emin, J.-C., Moignard, B. and Vandenbroucke, A. (2011), Pluralité des catégories de « cibles », tensions et implicites dans les élaborations. In M. Demeuse, D. Frandji, D. Greger and J.-Y. Rochex (eds.), Les politiques d’éducation prioritaire en Europe. Tome II, Quel devenir pour l’égalité scolaire? (pp. 53–104). Lyon: ENS Editions. Gewirtz, S., Ball, S. and Bowe, R. (1995), Markets, Choice, and Equity in Education. Maidenhead: Open University Press. Grafmeyer, Y. (1994), Sociologie urbaine. Paris: Nathan Université. Grootaers, D. (1998), Histoire de l’enseignement en Belgique. Bruxelles: Centre de Recherche et d’Information Socio-Politiques. Hirtt, N. (2014), Pourquoi les systèmes éducatifs de Belgique et de France sont-­ils les champions de l’inégalité sociale? Bruxelles: Aped.

Structural and Systemic Dimensions of School Segregation in Belgium

63

Jacobs, D. and Rea, A. (2007), Les jeunes Bruxellois, entre diversité et adversité: enquête parmi les rhétoriciens des écoles de la Ville de Bruxelles. Brussels Studies, 8: 1–17. Jacobs, D., Rea, A. and Hanquinet, L. (2007), Performances des élèves issus de l’immigration en Belgique selon l’étude PISA: Une comparaison entre la Communauté Française et la Communauté Flamande. Bruxelles: Fondation Roi Baudouin. Jacobs, D., Rea, A., Teney, C., Callier, L. and Lothaire, S. (2009), L’ascenseur social reste en panne: les performances des élèves issus de l’immigration en Communauté française et en Communauté flamande. Bruxelles: Fondation Roi Baudouin. Lafontaine, D. (2005), Hétérogénéité, mon cher souci. Revue Internationale d’Education, 40: 108–10. Lafontaine, D., Baye, A., Burton, R., Demonty, I., Matoul, A. and Monseur, C. (2003), Les compétences des jeunes de 15 ans en Communauté française en lecture, en mathématiques et en sciences: résultats de l’enquête PISA 2000. Cahiers du Service de Pédagogie Expérimentale, 13–14: 7–230. Lafontaine, D. and Monseur, C. (2011), Quasi marché, mécanismes de ségrégation sociale et académique en Communauté française de Belgique. Education Comparée, 6: 69–90. Lijphart, A. (1977), Democracy in Plural Societies: A Comparative Exploration. New Haven: Yale University Press. Mangez, C. (2011), Evaluer et piloter l’enseignement: analyse d’instruments de la politique scolaire en Belgique francophone. (Doctoral dissertation, Université Catholique de Louvain). Maroy, C. and Delvaux, B. (2008), Logiques d’établissements, interdépendances compétitives et inégalités sociales. In V. Dupriez, J.-F. Orianne and M. Verhoeven (Eds.), De l’école au marché du travail, l’égalité des chances en question (pp. 205–32). Bern: Peter Lang. McKinsey & Company (2015), Contribuer au diagnostic du système scolaire en FWB. Rapport à la Vice-Présidente, Ministre de l’Education, de la Culture et de l’Enfance. Bruxelles: McKinsey & Company. Monseur, C. and Lafontaine, D. (2012), Structure des systèmes éducatifs et équité: un éclairage international. In M. Crahay (ed.), Pour une école juste et efficace (pp. 185–219). Bruxelles: De Boeck. Niemann, D. (2015), Le «choc PISA» en Allemagne et les réformes de l’éducation. Administration & Éducation, 145 (1): 39–44. Ouali, N. and Rea, A. (1994), La scolarité des élèves d’origine étrangère (The schooling of pupils of foreign origin). Cahiers de Sociologie et d’Economie Régionales, 21–2: 7–56. Rochex, J.-Y. (2010), Les trois « âges » des politiques d’éducation prioritaire: une convergence européenne? In C. B. Ayed (Eds.), L’école démocratique. Vers un renoncement politique? (pp. 94–108). Paris: Armand Colin. Roncayolo, M. (1994), Préface. In J. Brun and C. Rhein (eds.), La ségrégation dans la ville: concepts et mesures (pp. 13–18). Paris: L’Harmattan. Van Zanten, A. (2001), L’école de la périphérie. Paris: PUF. Vandenberghe, V. (1998), L’enseignement en Communauté française de Belgique: un quasi-­ marché. Reflets et perspectives de la vie économique, 37 (1): 65–76. Verhoeven, M. (2011), ‘Multiple Embedded Inequalities and Cultural Diversity in Educational Systems. A Theoretical and Empirical Exploration’, European Educational Research Journal, 10 (2): 189–203.

64

Understanding School Segregation

Legal references Circulaire ministérielle de la Communauté française de Belgique ayant pour objet « Expériences pilotes de zones d’éducation prioritaires » (12 April 1989). Décret de la Communauté française de Belgique définissant les missions prioritaires de l’enseignement fondamental et de l’enseignement secondaire et organisant les structures propres à les atteindre (23 September 1997). Décret de la Communauté française de Belgique visant à assurer à tous les élèves des chances égales d’émancipation sociale, notamment par la mise en œuvre de discriminations positives (22 August 1998). Décret de la Communauté française de Belgique relatif à l’évaluation externe des acquis des élèves de l’enseignement obligatoire et au Certificat d’études de base au terme de l’enseignement primaire (2 June 2006). Décret de la Communauté française de Belgique portant diverses mesures visant à réguler les inscriptions et les changements d’école dans l’enseignement obligatoire (3 July 2007). Décret de la Communauté française de Belgique visant à réguler les inscriptions des élèves dans le premier degré de l’enseignement secondaire et à favoriser la mixité sociale au sein des établissements scolaires (18 July 2008). Décret de la Communauté française de Belgique organisant un encadrement différencié au sein des établissements scolaires de la Communauté française afin d’assurer à chaque élève des chances égales d’émancipation sociale dans un environnement pédagogique de qualité (9 July 2009). Décret de la Communauté française de Belgique définissant les missions prioritaires de l’enseignement fondamental et de l’enseignement secondaire et organisant les structures propres à les atteindre en ce qui concerne les inscriptions en première année du secondaire (18 March 2010).

4

Patterns of School Segregation in Brazil: Inequalities and Education Policy Tiago Lisboa Bartholo and Marcio da Costa

1.  Introduction This chapter discusses the distribution of educational opportunities in Brazil, and presents a review of the most recent publication related to school segregation and educational inequality in the country. The empirical analysis investigates patterns of between school segregation in four different state capitals: Rio de Janeiro, São Paulo, Belo Horizonte and Curitiba. It uses secondary data provided by INEP (Instituto Nacional de Estudos e Pesquisas Educacionais) and the Rio de Janeiro Educational Department for all children enrolled in primary education (first to fifth grade – 6 to 11 years old), including public and private schools. The term segregation here refers to an uneven distribution of students with similar characteristics across a school system. Three different indicators of potentially disadvantaged students were calculated using the Segregation Index (GS) (Gorard et al., 2003; Gorard, 2009): a) parents’ education; b) colour;1 and c) age-­grade distortion. This latter variable summarizes information on all students that have not followed a regular age-­grade flow in different educational transitions. The concept of segregation should not be considered a synonym of discrimination or unfairness. As measured here, using the Segregation Index (Gorard et al., 2003) or Dissimilarity Index (Duncan and Duncan, 1955), school segregation is, to some extent, almost unavoidable. Nonetheless, it is important to be aware of the phenomenon, describe patterns over time and create models to analyse the potential impact of educational policies on school segregation. Evidence from many different countries suggests that school segregation is a universal phenomenon and has to be considered as a consequence of residential segregation, educational policies and parental choice, which are presumed to correlate with social, economic and cultural isolation (Harris, 2012). In the case of Brazil, we prefer to use this denomination instead of race, given that it is a little less imprecise. The classification adopted by the national organ concerned with demography expresses a certain confusion in the national context by using colour and ethnic origin in the same table for self-­definition of the population.

1

66

Understanding School Segregation

The debate about the social impact of clustering students has different characteristics depending on the country of interest. In the United States, for example, the main concern has been with segregation by ethnic background, more specifically, black pupils (Goldhaber, 1999; Hill and Lake, 2010; Sikkink and Emerson, 2008; Saporito, 2003). During the 1950s, several indices of residential segregation were created, initially concerned with the racial division in the United States, later to be extended to other fields, including the public school admission patterns. European researchers from different countries, such as England or the Netherlands, have been more focused on segregation by first (or home) language and poverty (Gorard et al., 2003; Gramberg, 1998). In Brazil, the concern with school segregation is not recent, although the few previous studies used to analyse a very small number of schools, mainly focused on clustering by poverty (Consorte, 1959; Costa, 2008). Most of the empirical research regarding the subject of school segregation addresses two complementary issues. The first concerns the possible effects of this phenomenon. What are the impacts of clustering pupils with similar characteristics? Are there any potential benefits or deleterious effects of intentionally clustering pupils? Secondly, the debate about impact should be followed by a different question: what is the role of policy on the levels of school segregation? Understanding segregation patterns and how they occur are important for future public policies aimed at achieving more equitable educational systems. It could be argued that concentrating potentially disadvantaged students (for example, living in shantytowns or those from poorer, less educated family backgrounds) in one area (or in a certain school) can be efficient when seeking to implement focused policies as ameliorative packages. On the other hand, there is an increasing amount of evidence suggesting that clustering pupils with similar characteristics can have an impact on how they are treated at school, the quality of teaching, the overall levels of attainment, post-­compulsory education, and also an increasing association between academic achievement and socio-­economic status (Haarh et al., 2005; EGGRES, 2005; Rosenthal and Jacobson, 1968). Recent publications using data from Rio de Janeiro public schools corroborate the school composition effect or school-­mix-effect theory (Bartholo; Costa, 2016). The subject of school segregation had been historically scarce in Brazilian educational research. There are a few qualitative studies and even fewer studies using secondary data for a large proportion of the students. Nonetheless, the subject has gained recent relevance with the demonstration that educational inequality in Brazil has been increasing over the past ten years, despite the decline of social inequality during the same period. José Francisco Soares, former president of INEP,2 presented figures in the opening lecture in ABAVE 3 Conference 2017, Salvador, Brazil, demonstrating the increase

The National Institution of Education Research, INEP, is in charge of producing Prova Brasil, standardized test for all public schools and Annual School Census, http://portal.inep.gov.br/web/ guest/inicio. 3 ABAVE – National Association of Education Assessment, http://ixreuniao.abave.org.br. 2

Patterns of School Segregation in Brazil

67

in differences in performance between the first and fifth quintile of average SES (socio-­economic status) in Prova Brasil,4 from 2005 to 2015. During the same period, considering all states in the country,5 income inequality presented a reduction, along with an increase in annual expenditure per student. We believe that studies about school segregation can help understanding of this apparent contradiction. Data suggest that social stratification among schools in Brazil is high, not only considering the private and public systems, but also among public schools. Aware of the effects of clustering pupils, researchers should be able to understand the mechanisms that cause segregation or at least describe elements/variables associated with it. The real challenge in this type of research is to disentangle each one of these elements (explanatory variables) and show how they impact segregation. Maybe the most relevant question for the educational field is: what is the role of policy on the levels of segregation? Is it possible to interfere in the segregation process simply by changing the legislation regarding student enrolment and transfer? The article presents two main sets of results. The first describes patterns of school segregation by colour and age-­grade distortion from 2007 to 2016 in different Brazilian state capitals,6 considering all students enrolled in public and private schools. A second analysis, focused on Rio de Janeiro public municipal schools, presents evidence about the impact of educational policies on school segregation by parental education and colour.

2.  The debate about school segregation in Brazil The theme of school segregation has been addressed recently and has scarcely been disseminated in Brazil. Despite an old work by Consorte (1959), a study of limited scope that observed differences in composition among public schools in one particular region in Rio de Janeiro, the matter was not effectively explored until some incursions of urban sociology converged with initial studies by our research team at the Rio de Janeiro Federal University. Disregarding the discussions, in general, normative, about inclusive policies for students with significant physical and/or cognitive deficiencies, outside the scope of our interest, it may be affirmed that it was only in the twenty-­first century that a limited area of investigation of school segregation was constituted. In the general policy agenda, the discussions about inequality of educational opportunity generally did not address the phenomenon in the large public municipal and state educational systems. Based on traditional politico-­ideological approaches that run counter to public and private systems as the principal source of inequality in the distribution of school opportunities, the focus tended not to shed light on the evident inequality within the public systems.

A national assessment in Portuguese and Mathematics, applied to all students enrolled in the fifth and ninth grades in public schools since 2005 in Brazil. 5 Federal policies, such as Fundef (1996) and Fundeb (2006), were important to reduce inequality in student per capita investment among states and cities. 6 Rio de Janeiro, São Paulo, Belo Horizonte, Salvador and Curitiba. 4

68

Understanding School Segregation

Attention to the severe problems of general quality in our schooling systems, expressed by poor school performance indicators, even considering the upper strata in our school and social stratification, perhaps explains the political inattention to how unequally this quality is distributed. Since 2007, a federal educational policy has attempted to combine performance with school flows, by means of a single indicator, the IDEB (Basic Education Development Index), measured in almost all the country’s public schools (Soares, 2009). However, such measurement is tangential to the question – that is, not dealing with it directly, given that the mean performance in standardized tests is counterbalanced by an indicator of the proportion of students that are regularly promoted to each grade. The latter indicator may be taken as a measurement of inequality in student progress by school grade, but it does not in fact become a measurement of the unequal distribution of school opportunities, even though it could arouse some concern for ‘those left behind’. The creation of the IDEB, turned into a yardstick of the quality of Brazilian education, and measured school by school, can be interpreted as a clear sign of concern for the differences among schools. This does not necessarily lead to concern regarding school segregation, considered as the composition of the student population, but it opens up possibilities for such. Nationwide, it was only in 2007 that the National Student Census (Censo Escolar), conducted annually by the Ministry of Education, began to collect information individualized by student. Until then, data about students were collected in aggregated form and provided little information about their home background characteristics. Therefore, with the introduction in 2005 of Prova Brasil, the national examination held in all public schools in the fifth and ninth grades, containing a socio-­economic questionnaire for the students and the individualized School Census, it became viable to gather more detailed information on the phenomenon of school segregation. Since 2007, Prova Brasil and the School Census have generated the IDEB. Certainly, the countrywide diffusion of the literature and research about school effectiveness (Alves and Franco, 2008; Brooke and Soares, 2008), jointly with the enrichment of the information bank, has played an important role in the matter of segregation appearing on the agenda. Part of the work that addresses the theme of school segregation is aimed at a general treatment of the subject, which identifies differences in skin colour, gender or place of residence, associated to attainment, and, eventually, to its effects on income distribution. We can certainly affirm that the treatment of the theme with attention to how students are distributed among schools, because these processes are hierarchized, including internally, and, above all, how such administrative and political mechanisms tend to preserve or even multiply inequalities prior to school, is restricted to a small number of researchers. Thus, even if the effective inequality in schooling is broadly recognized and, more recently, with the development of the educational information systems, measured more precisely, the studies about how they are produced according to policies and organizational practices of educational systems are still quite scarce. Our observation points out that the terms ‘school segregation’ or ‘educational segregation’ tend to be studied along three general lines. The first deals with their effects

Patterns of School Segregation in Brazil

69

on indicators of educational performance and general life opportunities of the people (Alves and Soares, 2007, 2008, 2009; Bartholo and Costa, 2016, Costa and Guedes, 2009; Costa and Koslinski, 2006; Flores and Scorzavafe, 2009). The second line of investigation usually portrays the phenomenon according to its socio-­spatial dimensions. The ‘geography of opportunities’ usually associates the differences among schools with aspects of the territories in which they are located and/or with the characteristics of their students’ places of residence (Koslinski et al., 2013; Lima and Araujo, 2016; Morandi, 2016; Paiva and Burgos, 2009; Ribeiro and Kaztman, 2008; Torres and Gomes, 2002). The third line seeks to delve into the school dimension of the phenomenon itself, into how it is produced and reproduced, essentially based on the conjugation among schools and bureaucratic and family action. This line may be subdivided into studies of segregation among schools and school systems (Alves, 2010; Bartholo, 2013, 2014; Bruel, 2016; Carvalho, 2014; Carvalho et  al., 2016; Costa and Bartholo, 2014; Costa and Koslinski, 2006, 2011, 2012; Costa and Nogueira, 2009; Costa et  al., 2013; Koslinski et  al., 2014; Koslinski and Carvalho, 2015) and intra-­school segregation (Bartholo and Costa, 2014; Costa and Koslinski, 2008). Recent studies in Rio de Janeiro municipal public schools suggest selection bias by school principals or other members of school bureaucracy to select students based on specific characteristics (Bartholo and Costa, 2014; Koslinski and Carvalho, 2015). The attempt above at general classification of the works about school segregation has led us to characterize this chapter as an endeavour that does not exactly fit into any of the three categories created. In effect, here we seek principally to describe the phenomenon in national terms and its recent evolution. There is no intention to establish relations that propitiate some type of causal inference, but to stimulate debate, portraying its dimension in seeking recognition of its relevance in the national agenda. Our most recent path heads towards identifying processes by which the state bureaucracy, irrespective of the deliberate intentions of its actors, and the educational policies can mitigate or accentuate negative externalities of the system recognized as being of low quality, but also with high levels of inequality among them.

3.  Design and methods The chapter presents two sets of independent data: a) from the National Education Census (INEP) from 2007 to 2016; and b) administrative data from the Rio de Janeiro Education Department from 2006 to 2013. The Segregation Index (GS) was assessed considering all the available indicators of potential disadvantage, widely known in Brazil and also in other countries, to correlate with student achievement. Massey et al. (1996) identify five dimensions of segregation: evenness, exposure, concentration, centralization and clustering. In this research, the concepts of interest are evenness and exposure. Segregation here is a measure of an uneven distribution of pupils with similar characteristics across different schools. The GS indicates the exact proportion of disadvantaged pupils who would have to change school for there to be no segregation in terms of the specific characteristics expressed in the indicator. The formula below describes the GS:

70

Understanding School Segregation GS = 0.5 * {Σ|Fi / F – Ti / T|}

where: 1) ‘Fi’ is the number of potentially disadvantaged pupils in school ‘i’, where ‘i’ varies from 1 to the number of schools; ‘F’ is the total number of potentially disadvantaged pupils in Rio de Janeiro public municipal schools; ‘Ti’ is the total number of pupils in school i; and ‘T’ is the total number of pupils in Rio de Janeiro public municipal schools (Gorard et al., 2003). Gorard (2009) highlights four desirable properties that such indices must present, regardless of research field: 1) organisationally invariant, such that, if a school is broken into two, or if two schools merge, with the same proportion of FSM [Free School Meals], pupils in all the values of the index remain the same; 2) size or scale invariant, such that, if the number of both FSM and non-FSM pupils is multiplied by a constant in all schools, the value of the index remains the same; 3) compositionally invariant, such that, if the number of FSM pupils is multiplied by a constant in all schools, the value of the index remains the same (equivalent to the margin-­free criterion in sex segregation analysis); and 4) affected by transfers, such that, if an FSM pupil moves from a school with more FSM pupils to a school with less, the value of the index goes down. Gorard, 2009, p. 644

One of the main issues debated by researchers is whether the index changes just by a simple shift in the numbers of potentially disadvantaged pupils in a specific region. This is a crucial element, especially in situations where the researcher is interested in segregation patterns over time. The results will be presented in two separate sections. First, the chapter describes segregation patterns over time in four different Brazilian state capitals from 2007 to 2016. Bartholo and Costa (2014) presented a similar analysis, albeit with a shorter trend. The aim is to describe trends over time and highlight the impact of stratification by the public and private sectors in primary education in Brazil. Two characteristics of potentially disadvantaged students were calculated using the National School Census: colour and age-­grade distortion. Unfortunately, information for parental education is not available in the National School Census for private and public schools in Brazil. A second analysis uses administrative data from the Rio de Janeiro Education Department,7 and presents initial evidence of the impact of a new allocation policy, called the ‘lottery’ system, on patterns of school segregation. The aim here is to present evidence of the potential impact of educational policy on school segregation. This

Our research group at the Federal University of Rio de Janeiro (LaPOpE – Laboratório de Pesquisa em Oportunidades Educacionais) has an agreement with the Rio de Janeiro Education Department since 2009, to have access to administrative data for research purposes. This particular data set has more variables (such as parental education) with less missing data compared to the National School Census.

7

Patterns of School Segregation in Brazil

71

particular policy aimed to prevent school principals from selecting students based on adscript characteristics. In this context, it had a desegregation purpose, but initial evidence suggests an opposite effect.

4.  School segregation patterns in four Brazilian cities The data presented in this section provide the results of the Segregation Index for ten consecutive years in four large Brazilian cities that practise different school enrolment policies. Two characteristics regarding segregation can be directly observed in data from the Brazilian Educational Census. The first concerns the skin colour dimension. The second characteristic refers to schooling itself, the mismatch between student age and grade matriculation. This is not an important aspect in many countries, but it is extremely relevant in Brazil and in some other countries, where the phenomenon of examination failure is relatively common, along with other associated phenomena, such as dropout and provisional abandonment, albeit less observed for the age ranges considered. Both aspects are quite well correlated with the social-­economic dimension, which cannot be considered here, given the non-­existence of information about the latter in the National School Census, the basis for the comparisons presented herein. The organization of grades and levels at the schools is as follows: a) pre-­school for pupils aged 4–5 years is compulsory, starting in 2016; b) fundamental education is compulsory, attended by pupils aged 6–14, usually divided into five initial grades (first segment) and the last four grades (second segment); and c) High school caters for pupils aged 15–17 and is not compulsory. This chapter focuses on the initial five grades of fundamental education (students aged 6–11 years). It is also important to highlight that around 18 per cent of the pupils in the whole country are enrolled in private schools. Nonetheless, the numbers can vary across cities. In Rio de Janeiro, for example, the proportion of students enrolled in private schools is around 30 per cent. The four cities reported on are the capitals of four large states in the most developed region of the country (Paraná, São Paulo, Rio de Janeiro and Minas Gerais), all lying at the centre of major metropolitan regions. Curitiba, the smallest of the four capitals, has almost 2 million inhabitants; São Paulo, the largest, 12 million; Rio de Janeiro 6.5 million; and Belo Horizonte, 2.5 million. All the data were projected to 2016. Table 4.1 presents the differences in dimensions of each of the four cities’ municipal systems. There is a clear trend in the country for the municipal educational systems to cater for all the children of these grades, in the public ambit, but this is still not verified in all the states and cities. In Rio de Janeiro, the public educational system is almost entirely municipal, while in the other three cities there are both municipal and state schools which are attended by pupils at those initial grades. Table 4.2 presents patterns that are a little different and relatively stable over time, in terms of school segregation measured by colour. In this case, the reference to colour for which segregation was measured was black. The Index indicates the exact proportion of disadvantaged students who would have to change school for there to be no segregation.

Understanding School Segregation

72

Table 4.1  Total number of schools and students for public and private schools in 2016 PUBLIC

PRIVATE

City

No. of schools first to fifth grades

No. of students in first to fifth grades

BH Cu Rio SP

303 184 757 1,187

109,185 83,852 264,186 561,601

No. of schools first No. of students in to fifth grades first to fifth grades 306 174 1,125 1,339

42,977 36,378 139,718 218,319

Note: BH = Belo Horizonte; Cu = Curitiba; Rio = Rio de Janeiro; SP = Sao Paulo. Source: INEP, School Census 2016.

Table 4.2  Segregation Index (GS black students) in primary education (first to fifth grade) for public schools in Rio de Janeiro (Rio), Sao Paulo (SP), Belo Horizonte (BH) and Curitiba (Cu) – 2007–16 Year

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

GS BH GS Cu GS Rio GS SP

34 27.5 18 21

26 26.5 16 15

21 25 15.5 14

21 25.5 15.5 15

19.5 25 16 15.5

20 24 15.5 16

19 22.5 15 16.5

18.5 21 15 17.5

18 22.5 15.5 17.5

17 22.5 15 17

Source: INEP, School Census 2007–16.

The accentuated reduction measured in the segregation, especially between 2007 and 2010, and in particular in the three cities, Curitiba, São Paulo and Belo Horizonte, can be attributed to a reduction in the proportion of missing cases, which produces the statistical effect of an increase in heterogeneity regarding the aspect, skin colour, measured by school. In Belo Horizonte, missing cases fell from 44 per cent to 14 per cent, between 2007 and 2010. In Curitiba, they fell from 29 per cent to 12 per cent, and in São Paulo, from 35 per cent to 12 per cent. Rio de Janeiro, due to possessing an older, more consolidated municipal information collection system, presented a more stable curve, given that, in 2007, it had only 7 per cent missing cases, falling to 4 per cent in 2010. There are many differences among the cities analysed and one main variable that could influence GS is the proportion of disadvantaged students. Table 4.3 presents the proportion of disadvantaged students in each city considering only the public systems. Data suggests that changes in students’ profiles is not highly correlated with changes in GS. The correlation between GS and proportion of disadvantaged students varies across cities: Rio de Janeiro 0.79; Belo Horizonte –0.69; and Sao Paulo –0.21. This is an initial suggestion that other factors, such as students’ allocation policy and territory, could be playing a relevant role in the current patterns of school segregation. The phenomenon becomes more interesting when one seeks a more realistic picture of the school segregation and information is added about private schools found in most cities in Brazil. It is probable that the data presented in Table 4.4 demonstrate a complex phenomenon, whose most precise interpretation exceeds the limits of the analyses conducted for this chapter. One example is the frequent migration of students

Patterns of School Segregation in Brazil

73

Table 4.3  Proportion of Disadvantaged Students (Black) for Public Schools in Rio de Janeiro (Rio), Sao Paulo (SP), Belo Horizonte (BH) and Curitiba (Cu) – 2007–16 Year

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

Proport BH Proport Cu Proport Rio Proport SP

0.06 0.01 0.12 0.04

0.05 0.01 0.11 0.04

0.06 0.01 0.11 0.04

0.07 0.01 0.11 0.04

0.07 0.01 0.1 0.04

0.07 0.01 0.10 0.03

0.07 0.01 0.10 0.03

0.07 0.01 0.10 0.03

0.08 0.01 0.10 0.03

0.08 0.01 0.10 0.03

Source: INEP, School Census 2007–16.

Table 4.4  Segregation Index (GS black students) in primary education (first to fifth grade) for public and private schools in Rio de Janeiro (Rio), Sao Paulo (SP), Belo Horizonte (BH) and Curitiba (Cu) – 2007–16 Year GS BH GS Cu GS Rio GS SP

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

39 32 23 27

33 33.5 26.5 22.5

28 33.5 25.5 21.5

29 30.5 25 22.5

28.5 32.5 25.5 22.5

28.5 32.5 25 23.5

28 33 25.5 24.5

27.5 31 25.5 24

27.5 32.5 26 24

27 39 25.5 25

Source: INEP, School Census 2007–16.

between public and private schools. In Rio de Janeiro, around 6 per cent of all students enrolled in public schools change to a private school every year (Bartholo, 2014). It is reasonable to assume that a similar proportion of students enrolled in private schools migrate to the public system, since the proportion in both systems is fairly stable. These changes over time have the potential to affect GS and make causal inference more difficult. When the data about the students enrolled in the first to fifth grade (primary education) in the municipal, state and private school system are added up, it is possible to observe an increase in the overall levels of school segregation. This was expected, since the students’ profiles in the public and private sector are not similar. Also, in the private schools, the number of missing cases is reduced, but in a much lower proportion than in the municipal systems. The greatest reduction in the absence of information about skin colour occurs in Belo Horizonte between 2007 and 2010, which may explain the greatest fall in the GS in this particular place. However, in 2010, in Curitiba, Rio de Janeiro and Belo Horizonte, 50 per cent of the cases were still not informed. São Paulo had a reduction from 33 per cent to 27 per cent in the period, which may contribute to a quite stable curve. In the opposite direction to the decreasing trend of the GS, by a merely statistical effect of an increase in information available, the period presents another antagonistic trend, which is a certain rise in real average income of a significant portion of the Brazilian population until approximately 2014. With this income rise, there occurred a significant movement on the part of the population towards seeking matriculations in private schools.8 See recent publication in http://www.latitude.org.br/evolucao-­das-matriculas-­na-rede-­publica-e-­ privada-no-­municipio-do-­rio-de-­janeiro.

8

Understanding School Segregation

74

Considering there exists a much greater proportion of black individuals in the lower income bracket, the migration of non-­blacks to private schools may contribute to the rise in the GS by making the schools more homogeneous in this aspect. Rio de Janeiro, in its municipal system, since the beginning of the period, has practiced a policy of school choice for the families, with support from the programme that ensures free public transport for students in the school system. This seems to be one of the reasons for lower segregation levels than in the other cities, as, possibly, the factors related to segregation based on residence act with less intensity. The results presented in Table 4.5 concern an aspect of the segregation phenomenon that is more directly related with the school: age-­grade retention. It is known that the non-­progression of students through their school transitions is strongly affected by their social origin and other associated factors. However, we do not have sufficient information at our disposal on this aspect. Retention was measured considering the same criterion adopted by INEP: children aged 8 or older in the first year are in a situation of age-­grade distortion. It is important to emphasize that the Ministry of Education suggests that children should enter the first year of primary education at the age of 6. Just as in the aspect, skin colour, with regard to retention, Rio de Janeiro and São Paulo present less segregation. It is very likely that part of the difference observed in GS, considering the four cities, is related to the total proportion of disadvantaged students (age-­grade distortion). Table 4.6 presents the proportion of students that have been coded in the dataset as retained. In the case of Rio de Janeiro, the proportion of

Table 4.5  Segregation Index (GS retention) in primary education (first to fifth grade) for public schools in Rio de Janeiro (Rio), Sao Paulo (SP), Belo Horizonte (BH) and Curitiba (Cu) – 2007–16 Year

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

GS BH GS Cu GS Rio GS SP

32 24.5 19.5 22

27.5 21.5 19 21.5

25.5 26 18.5 22.5

23.5 26.5 17 23.5

28.5 28.5 18.5 23.5

31 28.5 18 22.5

32 28 18 21.5

32.5 26 19 20.5

31 29 19.5 19.5

31.5 30.5 19.5 19

Source: INEP, School Census 2007–16.

Table 4.6  Proportion of disadvantaged students (retention) in primary education (first to fifth grade) for public schools in Rio de Janeiro (Rio), Sao Paulo (SP), Belo Horizonte (BH) and Curitiba (Cu) – 2007–16 Year

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

GS BH GS CU GS Rio GS SP

0,11 0,05 0,20 0,05

0,10 0,04 0,18 0,05

0,09 0,05 0,16 0,04

0,08 0,05 0,18 0,04

0,08 0,06 0,21 0,04

0,08 0,06 0,21 0,04

0,07 0,05 0,19 0,04

0,07 0,05 0,20 0,04

0,06 0,05 0,19 0,05

0,06 0,05 0,19 0,05

Source: INEP, School Census 2007–16.

Patterns of School Segregation in Brazil

75

Table 4.7  Segregation Index (GS retention) in primary education (first to fifth grade) for public and private schools in Rio de Janeiro (Rio), Sao Paulo (SP), Belo Horizonte (BH) and Curitiba (Cu) – 2007–16 Year

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

GS BH GS Cu GS Rio GS SP

35.5 40 25.5 28

32.5 27.5 26.5 27.5

31 32 26 29

29 34.5 24.5 29.5

35.5 38.5 27 30.5

37 39.5 27 30.5

38 39.5 27 45.5

37.5 39 28.5 28

39 40.5 28 27.5

36.5 42 27.5 28

Source: INEP, School Census 2007–16.

students in an age-­grade situation, in all the years considered, is at least double those of the other cities. The difference among capitals is very large and can only be explained as a result of educational systems. In Standardized Tests like Prova Brasil, students from Rio de Janeiro public schools present good results compared to other capitals such as São Paulo; however, despite the fact that on average students are learning more or less the same, some public systems present higher rates of retention. This outcome is intriguing and corroborated the theory of ‘culture of failure’ in public and private schools in Brazil (Ribeiro, 1991; Earp, 2009). The correlation between GS and proportion of disadvantaged students for retention varies greatly across cities: Rio de Janeiro 0.12; Belo Horizonte –0.28; Curitiba 0.65; and São Paulo –0.61. Corroborating the results presented for colour, data for retention suggests that other factors, such as educational policy and territory, play relevant roles in the patterns of school segregation. Table  4.7 presents figures for GS retention considering all students enrolled in private and public schools. As expected, the numbers suggest an increase in school segregation when adding data from private schools. Since the proportion of age-­grade distortion is lower in private schools, the segregation rise suggests inequalities in age-­ grade flow for children starting school. It is important to highlight that there is quite a lot of stratification among private schools in Brazil. The distinctive position – in the market – of each school would be related to their reputations as schools of greater or lesser rigour regarding the performance of their students. Thus, the more prestigious schools would have less tolerance to maintain students with lower academic performance in their registers. Less prestigious private schools (with lower fees in their respective markets) would tend to receive, alongside the public schools, students who accumulate problems in their school trajectories.

5.  Impact of a new allocation policy for primary education in Rio de Janeiro Since records began, school enrolment in the Rio de Janeiro municipal educational system, which follows the so-­called fundamental education (students aged 6–14 years

76

Understanding School Segregation

old), has been theoretically free. In other words, there are no restrictions of a geographical nature, as is usual in various other large cities in the country and throughout the world. Hypothetically, the over 1,400 schools that cater for primary-­ level pupils feature open access to any student, provided that the stipulated age limits are observed. In a situation where there is over demand, schools can select their own students, mainly because there are no clear protocols about how school bureaucracy should act (Koslinski and Carvalho, 2015). There are data that suggest that most students study near their residences (Bartholo, 2014). However, due to the large supply of schools, spread throughout almost the whole city, and also given the policy that ensures free public transport for all public school students, in principle, the majority of these students always have alternative schools to choose from. There is, therefore, plenty of potential for a ‘market’ alongside the demand for schooling, especially considering that, given the demographic inversion and immigration retraction, the system is somewhat overdimensioned, catering for around 10,000 fewer students every year. In view of this ample supply, with demand fully met, the stable system in terms of its administrative and teaching staff, with universal rules, and reasonably egalitarian physical conditions of the installations and equipment of the school buildings, one would expect the municipal schools to be quite homogeneous. Nevertheless, as demonstrated in previous works, we are far from this situation (Bartholo, 2013, 2014; Bartholo and Costa, 2014). The variations in terms of school prestige, and consequently demand, are far from being balanced. Some schools tend to be viewed as outstanding in the eyes of the public, as already reported in the international literature. The renown attained bears a circular relation with the public served. Even in the conditions peculiar to a country like Brazil, in which the middle class and the more affluent sectors do not seek education in public schools, the remaining social hierarchies bear an intimate association with the hierarchy of prestige of the public schools. The more prestigious schools tend to be so mainly due to the social background of their students; in turn, they become more sought after because of their reputation. Their selective power is boosted in an atmosphere akin to a ‘student market’. The opposite circumstance is also verified. The information about IDEB and other indices of school quality are public and often publicized in newspapers or television. Recognizing this context, as of 2010, the Rio de Janeiro Municipal Department of Education has experimented with procedures that seek to counterbalance the unequalizing trend described above. There are two major situations in which students need to change school as a block: when they leave preschool to enter the primary school; and when they leave primary school (five initial grades, starting at age 6) on their way to secondary education (sixth to ninth grades). It was already customary in the educational system for parents to be called upon in these transitions to fill a form in triplicate with a list of schools, in order of preference, to which they intended to send their children. These are the so-­called ‘remanagements’ that shift students between schools on a large scale at the turn of the academic year. The reallocation occurs in close meeting and it is kind of a ‘black box’. These processes still occur and remain under the control of the educational bureaucracy, above all, the school boards and the

Patterns of School Segregation in Brazil

77

Regional Educational Committees (eleven altogether in the city). There are no clear rules in these events, except for enrolment priorities that affect a small fraction of the public,9 and we have had the opportunity of studying them (Carvalho, 2014). However, our interest in this new phase of investigation is aimed at a policy seeking to mitigate the trend described above. The first enrolments in the municipal public system, plus those of students returning to the municipal schools are required to be made through the central computerized system that requests up to five school choices, listed in order of preference. The system will always endeavour to match the preferred choice to each individual, who may reapply in a second round, in case, for some reason, his/her matriculation in the designated place is not confirmed in the first round. In the case of oversubscription to a specific school, children will be allocated randomly by the system. The initial hypothesis based on findings in the UK (Gorard et al., 2003) and in our previously studied field would be that, if segregating effects were identified, including in its operational mechanisms, under the former scheme, one could expect a reduction in the observable segregation patterns, since the lottery system would reduce the capacity for control over the process on the part of the school bureaucracy. In order to analyse the potential impact of the ‘lottery’ policy, we used a quasi-­ experimental design of Interrupted Time Series. The introduction of this system in 2010 was the temporal cut-­off point. Our analyses compare the school segregation patterns before and after the introduction of this system in the initial years of primary school. It is important to highlight that since the beginning of its implementation, the ‘lottery’ policy is not mandatory and parents can still choose the former scheme in order to enrol their children. We also made a comparison of the segregation patterns observed in the first year of primary school with those in the other grades of the first segment (second to fifth year of primary school). The reason for the comparison is simple: the proportion of families that adhered to the ‘lottery’ system in the first year of primary school is greater compared to the other first segment series – that is, if, in fact, we observe an impact of the new policy, this should be greater in the first year of primary school. In 2010, for example, 22 per cent of the families with children in the first year of primary school adhered to the new policy. For the other primary school grades, the figures were substantially less: second year 7.7 per cent; third year 5.8 per cent; fourth year 5.7 per cent; fifth year 6.1 per cent. In this specific analysis, we will use data obtained from the Rio de Janeiro Municipal Education Department. This database, created for administrative use, presents more variables than those observed in the Education Census, produced by INEP, and is of better quality (measured here with a smaller proportion of missing data). We shall measure school segregation taking into consideration colour/race and the parents’ schooling level. The parents’ educational level is possibly the best predictor of the child’s school trajectory, with the exception of direct measurement of proficiency. Controlling all the known variables that can affect school performance (including initial proficiency

For example, if the child has a brother or sister in the school.

9

78

Understanding School Segregation

measurements), children with more educated parents have greater chances of success at school (Gorard and See, 2013). In the databases, parents’ education is an ordinal variable with five points: illiterate; did not conclude primary school; concluded primary school; complete secondary school; higher education. In order to devise segregation indicators, the variable was summarized, creating two distinct groups of potentially disadvantaged students: parents who had not finished primary school (EducEF); and parents who had not finished secondary school (EducHS). Figure 4.1 presents the school segregation patterns for all the students enrolled in the first year of primary school. Part of these results had already been calculated by Bartholo (2014), but with a temporal series interrupted in 2010 – exactly in the initial period of the new student allocation policy – ‘lottery’. Now we can compare the implementation of the new policy in the pre (2006, 2007, 2008 and 2009) period with the post (2010, 2011, 2012 and 2013) period. A preliminary analysis of the data suggests a change in the school segregation pattern as of 2011, especially the variable, parents’ education – EducFS and EducHS. Considering the students whose parents had not completed primary school (EducFS), there is an inversion of the declining segregation trend with modest, but constant, increases in school segregation levels. For EducHS (for parents who had not completed secondary school), the initial pattern of inertia is altered as of 2011, with constant modest increases in school segregation. The school segregation patterns for black students is less clear, the school segregation rising and falling without any apparent link to the introduction of the aforementioned randomized policy. A complementary analysis considering only the variable, parents’ schooling, was conducted comparing the segregation patterns of the first year of primary school with the other students in the first segment of primary school (matriculated in the same schools). Our hypothesis is that the lottery effect would be stronger in the first year, taking into account the greater proportion of students allocated at random, when we

Figure 4.1  Segregation Index of students enrolled in the first year of primary school (2006–13) in the Rio de Janeiro Public Municipal Schools. Source: Rio de Janeiro Municipal Education Department.

Patterns of School Segregation in Brazil

79

Figure 4.2  Segregation Index for students matriculated in the first year and the second to fifth year of primary school (2006–13) in the Rio de Janeiro Public Municipal Network. Source: Rio de Janeiro Municipal Education Secretariat.

compare with the other grades in the same segment (second to fifth years). Figure 4.2 presents the temporal series considering the two student groups. The data suggest a rupture in the segregation patterns when comparing the students matriculated in the first year with those matriculated in the first segment of primary school (second, third, fourth and fifth years). School segregation becomes greater considering only the first-­year students with the other pairs in the same segment as of 2011. This difference seems to increase gradually when we consider EducFS (parents who have not concluded primary school). This result refutes the initial hypothesis that had suggested the ‘lottery’ policy would reduce the school segregation levels in the public municipal system. In reality, the data suggest the impact of the ‘lottery’ was on the increase (modest) in school segregation. The results are intriguing and new pieces of research should test two hypotheses: 1) the school-­level bureaucracy kept control on part of the pupils’ intake – they can be strengthening selective mechanisms; 2) the trigger-­effect – families with higher socioeconomic status have more resources (for example, access to information and the internet connection) which allowed them to make better use of the new opportunities created by the lottery scheme. Both hypotheses do not exclude each other, and it is plausible to assume that both could explain the increase in school segregation. The study design, Interrupted Time Series, presents some threats that must be investigated in future studies. Here we stress the two principal ones. The historical threat cannot be disregarded (Cano, 2000). In the case of the Rio de Janeiro municipal public educational system, the introduction of other policies, such as the restructuring of the network, or the Escolas do Amanhã [the Schools of Tomorrow], may alter the social composition of the schools, which would influence the final result calculated in this study. Another threat is related to the non-­random missing cases. Bartholo (2014)

80

Understanding School Segregation

has made a detailed description of the distribution of the missing cases over the years and has observed that there has been an important improvement in the quality of the data as of 2006. However, we cannot completely disregard the effects of the missing data in the calculation of the segregation indicators.

6.  Preliminary conclusions School segregation is a universal and complex phenomenon. There is very little research using large scale data in Brazil and most of the evidence is restricted to research in major cities like Rio de Janeiro and São Paulo. The chapter presents data for four state capital cities when discussing patterns of school segregation from 2007 to 2016. The results suggest that the profile of families (proportion of disadvantaged students in each city) in the public educational system is not correlated with the Segregation Index. This could be interpreted as an indication that other factors are influencing school segregation. Because we know that different states practice different allocation policies10, it is reasonable to assume that at least part of the variation observed across cities is related to educational policy and the specific characteristics of the territory. Future analysis should try to assess the specific impact of state or municipal level allocation policies. In the case of Rio de Janeiro, the chapter presents complementary evidence of the potential impact of a new allocation policy, called the ‘lottery’ system. The initial hypothesis suggested that the introduction of the new policy would diminish the potential of school bureaucracy to select students and, therefore, cause a decline in the overall levels of school segregation in the public schools. However, the initial data suggest the opposite effect. Complementary analysis should be conducted, especially the measuring of school segregation for the following years (2014, 2015 and 2016). Nonetheless, the evidence suggests a possible ‘trigger’ effect. More privileged families, with more access to information and use of the computer and Internet (Costa, Prado and Rosistolato, 2013), could be adjusting more rapidly to the new scheme.

References Alves, F. (2010), Escolhas familiares, estratificação educacional e desempenho escolar: Quais as relações. Dados – Revista de Ciências Sociais, 53 (2): 447–68. Alves, M. T. G. and Franco, C. (2008), ‘A pesquisa em eficácia escolar no Brasil: evidência sobre o efeito das escolas e fatores associados à eficácia escolar’, in N. Brooke and J. F. Soares (orgs.). Pesquisa em eficácia escolar: origem e trajetórias. Belo Horizonte: UFMG, 482–500. Alves, M. T. G. and Soares, J. F. (2007), Efeito-­escola e estratificação escolar: o impacto do uso da habilidade dos alunos na composição de turmas. Educação em Revista, 45 (1): 25–59.

It is possible to find State or municipal administration with two different school admissions codes: school choice (for example, Rio de Janeiro) or catchment area liked to family postcode. There could be small differences for each specific city of State.

10

Patterns of School Segregation in Brazil

81

Alves, M. T. G. and Soares, J. F. (2008), O efeito das escolas no aprendizado dos alunos: um estudo com dados longitudinais no Ensino Fundamental. Educação e Pesquisa, 34 (3): 527–44. Alves, M. T. G. and Soares, J. F. (2009), Medidas de nível socioeconômico em pesquisas sociais: uma aplicação aos dados de uma pesquisa educacional. Opinião Pública, 15 (1): 1–30. Bartholo, T. L. (2013), ‘Measuring Between-School Segregation in an Open Enrolment System: The Case of Rio de Janeiro’, Journal of School Choice, 7: 353–71. Bartholo, T. L. (2014), Segregação escolar na Cidade do Rio de Janeiro: Análise da Movimentação de estudantes. Estudos em Avaliação Educacional, 25 (58): 242–71. Bartholo, T. L. and Costa, M. (2014), Turnos e segregação escolar: discutindo as desigualdades escolares. Cadernos de Pesquisa, 44: 670–92, 2014. Bartholo, T. L. and Costa, M. (2016), ‘Evidence of a School Composition Effect in Rio de Janeiro Public Schools’, Ensaio, 24: 1–24. Brooke, N. and Soares, J. F. (2008), Pesquisa em eficácia escolar: origem e trajetórias. Belo Horizonte: Editora UFMG. Bruel, A. L. (2016), Distribuição de oportunidades educacionais: políticas para a escolha da escola em redes municipais de ensino. Jornal de Políticas Educacionais, 10 (19): 11–23. Cano, I. (2002), Introdução à Avaliação de Programas Sociais. São Paulo: FGV. Carvalho, J. T. C. (2014), Segregação escolar e a burocracia educacional: uma análise da composição do alunado nas escolas municipais do Rio de Janeiro. Dissertação. PPGE/UFRJ. Carvalho, J. T., Almeida, K. R. S., Koslinski, M. C. and Costa, M. (2016), Segmentação socioespacial, oportunidades escolares e patrimonialismo: sobre a construção de hierarquias internas aos sistemas públicos de ensino. Pesquisa e Debate em Educação, 6: 111–30. Consorte, J. G. (1959), A criança favelada e a escola pública. Educação e Ciências Sociais, 5 (11): 45–60. Costa, M. and Bartholo, T. L. (2014), Padrões de segregação escolar no Brasil: um estudo comparativo entre capitais do país. Educação & Sociedade, 35: 1183–1203. Costa, M. and Guedes, R. (2009), Expectativas de Futuro como Efeito Escola – explorando possibilidades. São Paulo em Perspectiva, 23: 1–14. Costa, M. and Koslinski, M. C. (2006), Entre o Mérito e a Sorte: escola, presente e futuro na visão de estudantes do ensino fundamental do Rio de Janeiro. Revista Brasileira de Educação, 11: 133–54. Costa, M. and Koslinski, M. C. (2008), Prestígio escolar e composição de turmas – explorando a hierarquia em redes escolares. Estudos em Avaliação Educacional, 40: 305–30. Costa, M. and Koslinski, M. C. (2011), Quase-­mercado oculto: a disputa por escolas ‘comuns’ no Rio de Janeiro. Cadernos de Pesquisa, 41: 246–66. Costa, M. and Koslinski, M. C. (2012), Escolha, estratégia e competição por escolas públicas. Pró-Posições UNICAMP, 23: 195–213. Costa, M. and Nogueira, M. A. (2009), Desiguais oportunidades de escolarização – antigo tema sob novos olhares na Sociedade Brasileira de Sociologia. Revista Contemporânea de Educação, 4: 228–33. Costa, M., Prado, A. P. and Rosistolato, R. (2013), Talvez se eu tivesse algum conhecimento . . .: caminhos possíveis em um sistema educacional público e estratificado. Interseções, 14: 165–93.

82

Understanding School Segregation

Duncan, O. D. and Duncan, B. (1955), ‘A Methodological Analysis of Segregation Indexes’, American Sociological Review, 20 (2): 210–17. Earp, M. L. S. (2009), A cultura da repetência em escolas cariocas. Ensaio, 17 (65): 613–32. European Group for Research on Equity in Educational Systems (2005), Equity in European Educational Systems: a set of indicators, European Educational Research Journal, 4 (2): 1–151. Flores, R. M. V. and Scorzafave, L. G. D. S. (2009), Uma análise dos efeitos da segregação racial sobre a proficiência dos alunos do ensino fundamental brasileiro. ANPEC. Goldhaber, D. (1999), ‘School Choice: An Examination of the Empirical Evidence on Achievement, Parental Decision Making and Equity’, Educational Researcher, 28: 16–25. Gorard, S. (2009), ‘Does the Index of Segregation Matter? The Composition of Secondary Schools in England since 1996’, British Educational Research Journal, 35 (4): 639–52. Gorard, S. and See, B. H. (2013), Overcoming Disadvantage in Education. London: Routledge Falmer. Gorard, S., Taylor, C. and Fitz, J. (2003), Schools, Markets and Choice Policies. London: Routledge Falmer. Gramberg, P. (1998), ‘School Segregation: The Case of Amsterdam’, Urban Studies, 35 (3): 547–64. Haarh, J., Nielsen, T., Hansen, E. and Jakobsen, S. (2005), Explaining Student Performance: Evidence from the International PISA, TIMSS and PIRLS Surveys, Danish Technological Institute. Harris, R. (2010), ‘Local Indices of Segregation with Application to Social Segregation between London’s Secondary Schools’, Environment and Planning, 44: 669–87. Hill, P. T. and Lake, R. J. (2010), ‘The Charter School Catch–22’, Journal of School Choice: Research, Theory, and Reform, 4 (2): 232–35. Koslinski, M. C. and Alves, F. Lange, W. J. (2013), Desigualdades educacionais em contextos urbanos: um estudo da geografia de oportunidades educacionais na cidade do Rio de Janeiro. Educação e Sociedade, 34 (125): 1175–1202. Massey, D., White, M. and Phua, V. (1996), ‘The Dimensions of Segregation Revisited’, Sociological Methods and Research, 21 (2): 281–92. Paiva, A. R. and Burgos, M. B. (2009), A Escola e a Favela. PUC-Rio/Palas. Ribeiro, L. C. Q. and Kaztman, R. (2008), A Cidade contra a Escola? Segregação urbana e desigualdades educacionais em grandes cidades da américa latina. Rio de Janeiro, Letra Capital. Ribeiro, S. C. (1991), A pedagogia da repetência. Estudos em Avaliação Educacional, 5 (12): 7–21. Rosenthal, R. and Jacobson, L. (1968), Pygmalion in the Classroom: Teacher Expectation and Pupils’ Intellectual Development. New York. Holst, Rinehart & Winston. Saporito, S. (2003), ‘Private Choices, Public Consequences: Magnet School Choice and Segregation by Race and Poverty’, Social Problems, 50 (2): 181–203. Sikkink, D. and Emerson, M. (2008), ‘School Choice and Racial Segregation in US Schools: The Role of Parents’ Education’, Ethnic and Racial Studies, 31 (2): 267–93. Soares, J. F. (2009), Avaliação da qualidade da educação brasileira In: Schwartzman, S. Schwartzman, I. F. O sociólogo e as políticas públicas (Org). Rio de Janeiro, Editora FGV, 215–242. Torres, H. and Gomes, S. (2002), Desigualdade educacional e segregação social na região metropolitana de São Paulo. Novos Estudos CEBRAP, número 64.

Part II

School Segregation and Student Performance

5

Refining Measures of Poverty and Their Impact on Student Progress in England Stephen Gorard and Nadia Siddiqui

The authors are currently looking at different ways of estimating disadvantage in schools in England based on existing data sets, creating new variables to encompass individual ‘trajectories’ of disadvantage, and applying these to analyses of the segregation of school intakes, and their outcomes. For example, we take a variable such as whether a pupil is eligible for free school meals (FSM) in any year (or whether data is missing), and collate this for every year the pupil was in compulsory schooling. The results can be used to create new variables, such as how many years a child has been FSM-eligible, for the individual and for those with whom they go to school. We do the same thing with other background variables such as living or going to school in a deprived area, having a special educational need (SEN), having English as an additional language (EAL), and even ethnic classification. We are also linking our new records to other data sets such as the Longitudinal Study of Young People in England (LSYPE) to see how well our new trajectory variables match the richer data, such as parental occupation and income, in such smaller data sets. This chapter looks at improving measures of student poverty in education in order to see what light this casts on substantive issues, such as the purported underachievement of specific groups, schools and regions. It suggests that some policies are being misdirected, and that funding to improve results for poorer students is not being targeted efficiently.

1.  The background Examining the precise nature of school intakes in terms of their social and economic characteristics is an important issue for research. The extent to which pupils with different backgrounds are (or are not) clustered together with others like them can reveal a great deal about a national school system (Gorard and See, 2013). In England, the level of poverty, ethnicity, language and special needs segregation between schools is now monitored annually (Gorard, 2015a). This chapter goes further by looking at one extreme form of residential segregation (North of England compared to South of England), and within England at areas using academic segregation (tracking). Once the ensuing clustering of school intakes by poverty is understood better, using a new

86

Understanding School Segregation

and more refined indicator of poverty, our understanding of relative attainment is improved.

1.1  Use of FSM in policy and practice Eligibility for free school meals (FSM) in England is a long-­standing indicator of the family poverty of any student. In practice, FSM-eligible refers to a student from any family entitled to income support, income-­based jobseekers allowance, child tax credit, the first four weeks of working tax credit following unemployment, the guaranteed element of state pension credit, employment and support allowance, and/or where part VI of the Immigration and Asylum Act 1999 applies. The 15 per cent or so FSMeligible students have lower levels of attainment at school, on average, and are less likely than their peers to make progress over time in terms of attainment (DfE, 2017), continue to post-­compulsory education or training, and to attend HE (Gorard and See, 2013). FSM has many advantages as an indicator of SES (socio-­economic status) background, compared to ethnicity or parental occupational class, for example. It has a clear legal definition in which a child either is or is not FSM-eligible. The chief criterion has not changed for decades, meaning that figures are reasonably comparable over time. Recording and reporting of it is a legal requirement for all state-­funded schools, and the FSM-status of each child is held as part of the National Pupil Database (NPD). FSM has become embedded in school policy in England. FSM-eligibility forms the main basis on which additional pupil premium (PP) funding is allocated to schools in order to help reduce the attainment gap between poor and other students (Gov.UK, 2014a). This money must be spent on activities primarily intended to raise the attainment of these potentially disadvantaged pupils (Gov.UK, 2015). It is used by the school inspection regime OFSTED when inspecting schools, and a pupil premium achievement gap has been formalized as part of their tracking system, ‘Reporting and Analysis for Improvement through school Self-Evaluation’ (OFSTED, 2015). This gap is the simple difference in percentage points in each school between the percentage of PP and non-PP pupils attaining five GCSEs at grade A*–C or their equivalent, including English and Maths (Gov.UK, 2014b). The GCSE is the most common traditional public examination at age 16. This gap is used routinely by schools themselves, their local authorities and sponsor chains, to monitor progress in improving attainment for PP pupils. It is even used to justify giving annual awards to schools with small or narrowing PP gaps (Pupil Premium Awards, 2015). However, FSM-eligibility is only a threshold characteristic. Some individuals will be from families only just below the income threshold, while others will be permanently and far below it. Some will have trajectories of moving in and out of FSM, while others will not. Some of this will be due to differences and changes in personal circumstances. Some will be sensitive to variations in the national or global economy, or to changes in the incentives for registering as FSM-eligible or for schools to collect accurate information about it (Gorard, 2015a). Those pupils previously eligible for FSM but not subsequently are termed a ‘hidden poor’ by Noden and West (2009, p. 4); they are no longer entitled to some benefits but are potentially still suffering the impacts of earlier

Refining Measures of Poverty and Student Progress in England

87

disadvantage. Hence, the Department for Education (DfE) has since used a variable representing whether a student has been eligible for any of the past six years: EverFSM6. However, this leads to new problems. Put simply, some FSM-eligible pupils will be poorer than others and eligible for assistance every year, and some may be at or near the threshold and so moving in and out of FSM-eligibility over time. All will trigger receipt of the pupil premium by their schools, but their absolute level of deprivation may vary considerably in a way that is subsequently linked to their attainment. If so, this would make the PP gap calculation intrinsically unfair, by favouring those schools or regions with more pupils near the threshold and fewer who are FSM-eligible year after year. Those currently eligible for FSM but not previously, or previously but not now, are in many respects more like those not eligible at all than those permanently poor (Crawford et  al., 2014). Those permanently eligible are the most disadvantaged, meaning that using EverFSM6 as a context variable would continue to be unfair to them. Instead using the number of years a student has been known to be FSM-eligible while at school would be a more sensitive metric of relative poverty (Gorard, 2016).

1.2  Segregation between schools It can be important to take into account not just individual poverty (above), but also the extent to which the poorest students are clustered together with others like them – by school, district or economic region (Siddiqui, 2017). It is clear from international evidence that students segregated by disadvantage fare worse than average at school for a number of reasons (Roew and Lubiesnki, 2017). Such settings can make pre-­existing inequalities worse by providing differential opportunities to learn (Schmidt et  al., 2015) or teachers’ responses to children (Strand and Winston, 2008), poorer instruction at school, less qualified teachers and substandard resources for the lower tracks (Harris and Williams, 2012; Kalogrides and Loeb, 2013). It can increase the impact of individual SES and low expectations (Parker et al., 2016), and affect relationships between pupils and teachers (Vieluf et  al., 2015), and between pupil peers, leading to poorer social skills. It can widen the gap between privileged and not so privileged pupils in terms of civic knowledge (Collado et al., 2015), emotional and behavioural problems (Muller and Hofmann, 2014), and even achievement (Goldsmith, 2011; Danhier and Martin, 2014; Yeung and Phuong Nguyen-Hoang, 2016). Taking into account both this form of clustering by poverty (hereafter termed segregation) and a more accurate level of poverty as assessed by the number of years FSMeligible, it is possible to show that the basis of several policy issues in England is flawed.

1.3  Case 1: The North–South divide Since at least the 1990s, it has been relatively common for policymakers, commentators and even academics to use differences in school-­level attainment between geographical regions of the UK to claim that the lower attaining region is somehow underperforming. No allowance is made for differences in school intakes, levels of relative poverty or

88

Understanding School Segregation

usually even the prior attainment of students. When such differences are factored in, or a comparison is made with similar regions in terms of school intakes, the differences tend to disappear (Gorard, 2000; Gorard and See, 2013). But the pattern of abusing raw-­score regional figures continues. This can be distressing for staff, pupils and families. It can lead to wasteful or even harmful attempted solutions. For example, there might be claims that it is somehow the students’ fault in that school or area, because they are not aspirational enough despite clear evidence that high aspirations are relatively common and un-­stratified in schools (Gorard, 2012). Although policy is often directed at a local level, the region, authority or district attended has little or no relevance for student attainment in the UK (Henderson, 2008). A recent example of such a policy issue concerns the apparent North–South divide in attainment in England. Many commentators agree with the former Chancellor George Osborne that ‘There is now overwhelming evidence that attainment at 16 is too low in the north, leaving us lagging behind UK and international competitors’ (The Guardian, 2017). The then Chief Inspector of Schools in England agreed that in northern England, ‘children have less of a chance of educational success than children south of the Wash’ (The Guardian, 2016). According to OFSTED, there are more than twice as many secondary schools judged inadequate in the North and Midlands compared with the South and East. And these views influence policy in and beyond education, such as whether to fund transport links and improve the rail network (Financial Times, 2016). According to the CBI, after ‘an analysis of official statistics’, ensuring that pupils get good GCSE or equivalent qualifications would be the most effective way of tackling productivity differences across the UK, rather than prioritizing faster road and rail links in the Midlands and north of England as the government had planned. Educationally, schools in the NE of England are being asked to improve their results in an NE Challenge akin to the supposedly successful London Challenge (Hutchings et al., 2012), but with the difference that the NE would not receive the extra funding that London schools did. In any case, the London Challenge schools already received more funding per pupil than those in the NE, and started their ‘challenge’ with already higher attainment and a lower poverty gap. It is not clear how successful it really was. The North and South of England are similar in many ways and there are high-­ attaining pupils everywhere, but there are bigger pockets of long-­standing disadvantage in the North – of a kind that is well-­known to be linked with lower average attainment at school. In the same way, two schools may be similar in some respects according to official data such as the percentage of pupils eligible for free school meals (FSM). However, they could be very different in reality – with one school having mostly pupils eligible for that year only and the other having pupils who had been officially classified as poor for their entire school career. This chapter takes such differences seriously.

1.4  Case 2: Expanding grammar school places At the time of writing, the UK government is planning to increase the number of pupils attending state-­funded selective grammar schools. They claim that this will assist overall standards, reduce the poverty attainment gap and so aid social mobility. Historically, grammar schools were widespread in the UK, set up as part of a planned

Refining Measures of Poverty and Student Progress in England

89

tri-­partite system after 1944 which actually became a two-­tier system of grammar schools and secondary-­moderns. Children took a series of tests at age 10/11 (the 11+), and those with high scores were selected to attend grammar schools, with the remainder going to secondary-­modern schools. In England, the number of grammar schools peaked at 1,298 in 1964, and then dropped as low as 150 in 1989, before returning to the current level by 2004 (House of Commons Library, 2017). The 163 remaining schools are disproportionately academies, and single-­sex with a sixth form, and their pupil intake has increased since 1980 (Bolton, 2016). These schools are over-­subscribed and popular with many local parents (Lloyds Bank, 2016). The system has largely been abolished in Scotland and Wales for the same reasons as in England (they tended to segregate by excluding poorer children). The system was retained in Northern Ireland, where it was made even more segregated by having sectarian grammar schools for Catholic and Protestant families (Gallagher and Smith, 2000). In September 2016, the UK Prime Minister proposed removing the law banning state-­funded schools in England (other than the 164 grammars in place in 1997) from using academic selection to allocate their pupil places (Foster et  al., 2016). Instead, new schools such as free schools and academies would be able to become new grammar schools, and the existing grammar schools could expand further by opening satellite schools. This was described by the PM and the DfE as a way to provide more good school places through ‘schools that work for everyone’ (https://engage.number10. gov.uk/good-­school-places). They see it as a responsibility of the state to provide especially for pupils who are gifted and talented, and they feel that a comprehensive school environment cannot fully nurture the potential of such pupils. One problem with grammar schools is that in selecting young children by attainment at age 11, they also select by poverty and other background characteristics linked to attainment, leading to very segregated local school systems and all of the problems that this entails (above). This is not like schools in the North and South happening to differ in terms of student disadvantage. This is now deliberate sorting by background because it has long been known that the extent to which pupils are clustered with others like them socially and ethnically is much higher in countries with selective systems (Jenkins et al., 2008; OECD, 2014), creating even further handicaps for the most disadvantaged students (Danhier, 2017). Developed countries with little or no diversity of schooling, such as those in Scandinavia, have low achievement gaps, higher average attainment and also the highest percentage of very skilled students (Alegre and Ferrer, 2010). Grammar schools use examinations to select children aged 10 or 11 who are predicted to do well in subsequent examinations at age 16. They select well, as evidenced by the high raw-­scores outcomes of these pupils five years later. This seems to confuse some commentators, members of the public and even policymakers, who assume that the good results are largely due to what happens in the school rather than the nature of the children selected. This is not a correct interpretation, and a counterfactual is needed to tell us what would have happened to these children if they had attended a different school. This is often attempted by looking not at raw-­score outcomes but at the amount of progress made by each pupil while at the school (the ‘value-­added’ model).

90

Understanding School Segregation

Some value-­added studies comparing grammar schools with comprehensive or other types of school have suggested that the former have better pupil outcomes even once prior attainment is accounted for (Prais, 2001; Levaçić and March, 2007). However, these studies also suggest that the subsequent lower attainment of the much larger number of pupils in the associated secondary-­modern schools at least outweighs any such gains. The system of selection is zero-­sum at best. Intriguingly, these same value-­added results appear if the model is still based on pupils attending grammar schools after age 11, but uses their Key Stage 1 (KS1) results for prior attainment when aged 7, and their KS2 results when aged 11 as the outcome. This cannot be due to attending a grammar school because pupils only move after KS2, and so this odd result suggests that the purported grammar school effect is in fact a form of unmeasured pre-­selection (Manning and Pischke, 2006). If the intakes to grammar schools really are already on a path to success based on their KS1 results, then that subsequent success at KS4 at age 16 must not be mistakenly attributed to having attended a grammar school in the meantime. Grammar schools did not confer a real advantage in the past and in their prime (Halsey and Gardner, 1953), and they do not appear to do so now (Sullivan et al., 2014).

2.  Methods The research presented here is based on the National Pupil Database for England, specifically the 2015 Key Stage 4 cohort, with attainment, school and background information for every year that they have been in compulsory schooling. There are 549,203 pupils with relatively complete records, of whom 75,787 (14 per cent) are listed as eligible to receive free school meals (FSM). The same analyses (below) have also been conducted with other KS4 cohorts with the same substantive results. The original pupil-­level variables involved in the headline analyses are: l l l l l l l l l l

l

l

l

school attended; local authority area; birth month and year – used to compute age in year; sex – girls tend to have better results than boys; ethnic origin or group; English as an additional or second language; special needs with a statement; special needs without a statement; whether the pupil moved to the school in the last two years; FSM-eligibility at KS4 – a flag variable showing whether a pupil is from a home officially classified as below the poverty line; EverFSM6 – whether a pupil has been eligible for FSM in any of the past six years; IDACI score – a measure of average deprivation for the area where the pupil lives or goes to school; KS1 points score – a scaled measure of attainment across all relevant subjects at age 7;

Refining Measures of Poverty and Student Progress in England l

l

91

KS2 points score – a scaled measure of attainment across all relevant subjects at age 11; and KS4 capped points score – a scaled measure of attainment across all relevant subjects at age 16.

Pupil-­level variables derived from the data are: l

l

l l l

the number of years in total a pupil was eligible for FSM up to KS4 – a more accurate measure of enduring poverty; the month of birth within the school year – to distinguish between summer- and winter-­born pupils; whether a pupil goes to school in the NE of England; whether a pupil goes to school in an area with grammar schools; and whether a pupil goes to a grammar school.

A key school-­level variable is the mean number of years pupils in that school have been known to be FSM-eligible (an indication of the segregation or clustering between schools by poverty). The effectiveness of schools in the NE, grammar schools and grammar school areas is assessed via regression models. Each model has the same basic structure. The outcome variable to be explained is the KS4 attainment score for each pupil. The predictors for all models at the first stage are all of the other variables listed above, except for the last three. These include prior attainment and the background characteristics of each pupil. Then, in one model, a variable is entered for being in a selective area/school or not, and in the other model, a variable is entered for attending a school in the NE or not. In this way, the amount of variation explained at each stage, and the coefficients for the explanatory variables, provide an estimate of the impact of attending a grammar school, or in the NE, or not, shorn of the known differences in the intakes to each type of school.

3.  Results As described above, the official measure of student poverty used in policy and practice is whether a child has been FSM-eligible in the past six years. For those with valid figures, most KS4 pupils have never been FSM-eligible in their last six years at school (Table  5.1). The remainder are split between those currently eligible (in their last year of compulsory school) and those no longer eligible. The first group (15 per cent) must include all students permanently living in poverty, and the latter group (11.7 per cent) is indicative of at least some of those pupils from families on the threshold of poverty. Pupils ever eligible for FSM differ, on average, from those not eligible in other ways – such as being more likely to have special educational needs (SEN) or have English as an additional language (Table 5.1). However, the ever eligible indicator used in policy

92

Understanding School Segregation

Table 5.1  Percentage of FSM groups with specified characteristics, England, KS4 FSM group Never FSM FSM previously FSM now

Percentage of cohort

Any SEN

EAL

Mean KS4 points (best 8)

73.3 11.7 15.0

14.5 25.5 32

10 17 20.2

303 230 205

conflates the more disadvantaged currently eligible group with those previously eligible, which form a group midway between the two (Gorard, 2016). As would be expected, the attainment results follow suit, with the ‘FSM now’ pupils having the lowest average KS points (for the best eight GCSE subjects passes, or their equivalent). The gap between the two FSM groups is smaller than that between the two groups combined and those never eligible, but it is still considerable. The duration of poverty matters. The difference that this could make to the pupil premium attainment gap as use in policy is illustrated using two local authority areas. Kensington and Chelsea, and Middlesbrough, are both urban areas and have around the same proportion of pupils who have never been eligible for FSM, which means that they each receive comparable pupil premium payments based on EverFSM6 (Table 5.2). However, Kensington and Chelsea in the South is a relatively rich area of London, while Middlesbrough is a very deprived area in the NE with high unemployment. They therefore differ considerably in terms of the proportions of the kind of FSM-eligible pupils that they contain. In Kensington and Chelsea, the clear majority of pupils who have even been FSMeligible are not now. These include, therefore, a proportion who are near the threshold of FSM rather than among the very poorest in the country. This could affect the level of qualifications obtained. In fact, over 36 per cent of pupils in Kensington and Chelsea are missing any data on FSM-eligibility, confirming that a large number of residents use private fee-­paying schools. This will remove some of the highest-­attaining or richest pupils from attendance at local state-­maintained schools. Because of the well-­ established correlation between socio-­economic status and attainment, this would then tend to reduce the overall level of attainment in local state-­funded schools. However, it would also reduce the likely gap between the poorest and the majority of those pupils remaining in state-­funded schools. This is the kind of factor never considered by those promoting the apparent success of the London Challenge. Any assessment of the pupil premium attainment gap must take these two factors into account. In Kensington and Chelsea, most pupils receiving the pupil premium are not currently FSM-eligible, and a large proportion of pupils go to school outside the Table 5.2  Percentage of each FSM group in Middlesbrough, and Kensington and Chelsea FSM group Never FSM FSM previously FSM now

Middlesbrough

Kensington and Chelsea

53 14 33

53 27 19

Refining Measures of Poverty and Student Progress in England

93

state system and are not included in the figures here. On average, the pupil premium attainment gap is lower in Kensington and Chelsea than in England overall. This is to be expected because some of the highest attaining pupils are missing (not in maintained schools), and more importantly because it has fewer permanently deprived pupils than the other areas. Therefore, as well getting as much extra PP funding as Middlesbrough, the schools in Kensington and Chelsea are more likely to be praised by OFSTED and rewarded in PP awards for having a low poverty gradient. This is unfair. The situation in the deprived authority of Middlesbrough is very different. Here only 4.7 per cent of pupils are missing data on FSM eligibility, which is around the same as the national average. This confirms that few pupils attend private fee-­paying schools. Almost all pupils are in the state-­funded system and so contributing to the pupil premium attainment gap there. Unlike in Kensington and Chelsea, the clear majority of pupils who have ever been FSM-eligible still are. They are likely to include many of those from families permanently receiving other benefits or on low incomes. It should also be expected that these two factors would both tend to increase the pupil premium attainment gap (irrespective of what actually goes on in schools or how the PP is used). The situation for FSM-eligible students in grammar schools is similar. Not taking account of the duration or level of poverty may be badly misleading policy and funding in England.

3.1  A new approach: FSM years Looking at the number of years a pupil has been eligible for FSM instead clarifies the picture (Table 5.3). The longer-­term poorer pupils are more likely to be from an ethnic minority, and to be labelled as having a special educational need. There are clear differences in the characteristics of FSM students, as well as between them and the majority. These differences then translate into attainment. Pupils who have never been known to be eligible for FSM have a much higher level of KS4 attainment than any pupil with even one year of FSM-eligibility. This is well-­known. Less well-­known is the fact that, on average, KS4 attainment declines with every year of FSM-eligibility (Table  5.4). This matters because the prevalence of the long-­term poor differs between North and South, and the kinds of FSM students attending grammar schools or not. Grammar schools are not only taking only a fraction of their fair share of FSM pupils, the few FSM pupils they do take are disproportionately those towards the left of Table 5.4, who are likely to do better whatever school they attend. This means that the other schools in selective areas are not only taking more than their fair share of FSM pupils, but are also

Table 5.3  Percentage of pupils with specified characteristics by number of years FSM, England SEN Non-­white

Never FSM

1

2

3

4

5

6

7

8

9

10

12 16

19 26

20 28

22 28

24 29

24 28

26 28

27 30

29 32

32 32

34 32

Understanding School Segregation

94

Table 5.4  Mean attainment scores of pupils by number of years FSM, England Never FSM KS1 points KS2 points KS4 entries KS4 points Best 8 VA

1

2

3

4

5

6

7

16.0 14.5 14.3 14.0 13.9 13.6 13.5 13.2 21.1 19.0 18.7 18.6 18.2 17.9 17.7 17.7 9.54 8.78 8.64 8.59 8.46 8.35 8.24 8.12 332 288 281 277 272 266 261 256 +8.7 −12.4 −16.0 −18.8 −22.8 −25.1 −30.3 −32.0

8

9

10

12.9 12.5 12.4 17.3 16.7 16.5 8.00 7.70 7.71 251 237 238 −36.0 −44.6 −42.5

disproportionately dealing with the more chronically poor in their areas. A similar, but naturally occurring, situation distinguishes North and South. It is shocking that the only group with positive value-­added scores on average at KS4 is those never eligible for FSM, otherwise the value-­added score declines in a clear progression with every year of eligibility (Best 8 VA). Poorer children start school with lower attainment than their peers, and then continue to lose ground over time, and the poorer they are the more they fall behind (as also noted in DfE, 2017). The long-­term poor are far less likely to go to grammar school (7 per cent as likely as never FSM students and only 22 per cent as likely as temporarily poor students) and more likely to live in a poor area (with 230 per cent of the IDACI score of never FSM students, and 144 per cent of the IDACI score of temporarily poor students), and they attend schools with more long-­term poor like themselves (in a local system segregated by chronic poverty). What difference does this make to the two cases presented earlier?

3.2  Case 1: results for North–South comparison As envisaged above, it is clear that the number of years any student has been FSMeligible varies considerably between areas in the North and South of England, on average (Table 5.5). It must be recalled that any student ever eligible will attract PP funding, and be taken into account in official performance figures. But such students are likely to have been recorded as poor for over six times as long in Middlesbrough (NE) as Wokingham (SW). This suggests that official accounts of failure in the NE could be at least partly due to such a big difference in levels of poverty. The first regression model is used to assess this. Putting all of the background variables in the first step explains or predicts around 67 per cent of the variation in KS4 outcomes (Table  5.6). However, then adding knowledge of whether a pupil was at school in the North East of England or not added nothing to the explanatory power of the model. This strongly suggests that any surface differences between NE and other regions disappear once the differences in school intakes are taken into account. There is greater poverty in the NE, but there is no additional educational problem in terms of school or teacher quality for policy to address. Table  5.7 lists the coefficients from the model. As would be expected, prior attainment at KS2 is the best single predictor of attainment at KS4. Aside from prior attainment, the most important predictor is the number of years any student has been

Refining Measures of Poverty and Student Progress in England

95

Table 5.5  Mean number of years FSM by local authority, England Local authority

Mean years FSM by KS4

Wokingham Buckinghamshire West Sussex Middlesbrough Manchester

0.5 0.6 0.7 3.1 3.4

Table 5.6  ‘Effect’ sizes from multi-­stage regression models predicting total capped KS4 points Background predictors Whether in North East

0.82 0.82

Table 5.7  Standardized coefficients from multi-­stage regression models predicting capped KS4 points KS2 average points (prior attainment) KS1 average points (prior attainment) Sex of pupil Month in year (summer born) Number of years known to be FSM-eligible IDACI scores (deprivation) Special need (SEN) Mean number of years FSM-eligible, school Joined school in last two years (mobility) English as an additional language Non-­white UK (ethnic minority) Schooled in North East

0.57 0.10 0.08 0.04 −0.10 −0.05 −0.10 −0.05 −0.08 0.08 0.06 0.01

eligible for FSM. Adding context, and using this more refined measure of poverty, leads to a different substantive finding to that of the CBI and others (above). In fact, the coefficient for NE yields a negligibly small benefit for attending a school in the NE (‘effect’ size 0.01).

3.3  Case 2: results for grammar schools Almost exactly the same picture emerges for grammar schools. In England, the still largely comprehensive system means that social, racial and economic segregation between schools is relatively low in international terms (Gorard, 2015b). However, there is considerable variation between areas such as local authorities. The few authorities that have retained selection and grammar schools have the highest level of SES segregation in England. The correlation between FSM segregation and the number of grammar schools in any area is 0.62 (Gorard et al., 2013). In areas that have grammar schools, those living in the most disadvantaged parts are less likely to attend a grammar even where they have high prior attainment scores (Cribb et  al., 2013). Grammar schools take only 28 per cent of their ‘fair’ share of poorer children (DfE,

96

Understanding School Segregation

2017). Even where those attending grammar schools are FSM-eligible, they will have been so for fewer years. Of course, some of these differences could be due to the kind of pupil populations in areas where the minority of 163 grammar schools remain, which could differ from the rest of the country. To assess this, Tables  5.8 and 5.9 compare those attending grammar schools with only those pupils in areas with grammar schools. These make it clear that the differences are not produced by the geography of where grammar schools still exist. In fact, pupils in grammar schools are even less representative of their local areas than they are of pupils in England as a whole. In particular, even the few FSMeligible pupils in grammar schools have been eligible for noticeably fewer years than in the rest of the school system. Children aged 10 or 11 are put forward and then tested for entry to grammar schools on the basis of their ability, prior attainment and motivation. However, this selection process indirectly also selects for a wide range of other characteristics most of which should not be relevant. It is understandable that pupils with serious learning challenges will be less likely to pass an 11+ test of ability or attainment, even assuming it is a fair test in that respect. It is also understandable that children for whom English is not their first language will tend to do worse. But it is harder to see why the family income, ethnic origin and precise area of residence should be so stratified. For the present, the key point is that those who go to grammar schools differ from the rest of the schools in England by far more than their talent as tested by the 11+. Therefore, grammar schools cannot be said to be obtaining better results with equivalent pupils

Table 5.8  Characteristics of grammar school pupils compared to pupils in selective areas, 2015 Mean IDACI score FSM years by KS4, all pupils FSM years by KS4, FSM pupils

Grammar

Not grammar

0.1 0.3 5.1

0.2 1.6 6.8

Table 5.9  Characteristics of grammar school pupils compared to pupils in selective areas, percentages, 2015 Non-­white UK White, any Black, any Pakistani/Bangladeshi Chinese EAL students SEN students with statement SEN students no statement EverFSM6 FSM2015

Grammar

Not grammar

25 75 3 15 2 11 0 5 5 2

18 82 4 9 0 13 4 13 26 14

Refining Measures of Poverty and Student Progress in England

97

even once that prior ability and attainment, as demonstrated in the 11+ test, is taken into account. This also means that grammar schools and the areas they are in are much more segregated by any of these indicators than the rest of England is. However good grammar schools are (or not) this must be set against the real dangers from such a deliberate policy of socio-­economic segregation between schools (above). The differential effectiveness of grammar schools is estimated via two regression models: one based on all schools, and the other only in areas that contain grammar schools (Table  5.10). The R value using all of the pupil background was 0.82 for both models (just the same as for the NE model, of course). Additional background variables could improve this somewhat, but there will always be an error term due to missing data, measurement errors and so on (Gorard, 2010). The first model in each category is slightly better than it would be traditionally, due to knowledge of exactly how many years each pupil has been FSM-eligible rather than using the traditional binary classification. If the model is run with ‘current’ FSM-eligibility at KS4 as is standard rather than the number of years eligible, the R for the first model drops to 0.77. Although this difference is small, it does suggest that the new measure of chronic poverty is picking up variation that neither current FSM nor EverFSM6 does. The first model does not improve at all when knowledge is added of whether a pupil goes to school in a selective area or not. This means that if grammar schools are at all differentially effective, their effect is indeed zero-­sum and wiped out by exactly equivalent harm done to the rest of the nearby school system. However, adding knowledge of whether a pupil goes to a grammar school also improves neither model at all. With only 163 grammar schools, it could be argued that they would not be expected to add much to the full model for all pupils in England. However, they do not add anything to the smaller model restricted only to areas with grammar schools either. On this basis, grammar schools appear to be no more or less effective than other schools, once their clear difference in intake has been taken into account. Table  5.11 shows the standardized coefficients for each variable in the models. Again, prior attainment at KS2 is by far the best predictor of KS4 outcomes, following by prior attainment at KS1, the number of years eligible for FSM, and whether a pupil has any kind of special need. The least important variables, with almost negligible ‘effect’ sizes, are the date of birth in the school year, the level of deprivation in the area, whether the school is in a selective area, and whether it is a grammar school. However, the level of deprivation would increase in importance if FSM-eligibility were not available because the two are correlated. Similarly, the month of birth in year appears far less important than in reality because the prior attainment scores are acting as proxy to a great extent (younger pupils do less well at KS1, KS2 and KS4). Table 5.10  R from multi-­stage regression models predicting capped KS4 points, England Background predictors Whether in selective area Whether in grammar school

All pupils

Pupils in selective area

0.82 0.82 0.82

0.82 – 0.82

98

Understanding School Segregation

Table 5.11  Standardized coefficients from multi-­stage regression models predicting capped KS4 points, England KS2 average points (prior attainment) KS1 average points (prior attainment) Sex of pupil Month in year (summer born) Number of years known to be FSM-eligible IDACI scores (deprivation) Special need (SEN) Mean number of years FSM-eligible, school Joined school in last two years (mobility) English as an additional language Non-­white UK (ethnic minority) Schooled in a selective area Attends grammar school

All pupils

Pupils in selective area

0.57 0.10 0.08 0.04 −0.10 −0.05 −0.10 −0.05 −0.08 0.08 0.06 −0.01 0.04

0.55 0.11 0.07 0.04 −0.10 −0.05 −0.10 −0.07 −0.07 0.06 0.05 – 0.07

4.  Conclusions This chapter suggests that using the more sensitive measure of the duration of poverty has a lot to recommend it. Using the number of years a student has been eligible for FSM, and how segregated a school system is by poverty and other indicators of disadvantage, it is possible to explain substantive differences such as the apparently superior attainment of schools in the South of England compared to the North, and grammar schools compared to comprehensives. Both appear to be due more to local and sector-­related segregation between schools in the national school system. Because there is greater relative poverty in the North, one would expect the average school outcomes to be slightly lower there. And because grammar schools inadvertently select by poverty as well as attainment at age 11 one would expect their outcomes to be slightly higher than predicted by prior attainment alone. The finer measure of poverty introduced in this analysis reveals that this is so, in a way that the EverFSM measure cannot. This means that any policies predicated on the raw differences between North and South or grammar and non-­selective are being misdirected. Improving schools in the NE may not be as effective as investment in infrastructure. Poorer families may be helped more by abolition of the existing grammar schools than the creation of new ones, and inevitably a much larger number of schools around them that cannot be comprehensive in intake and tend to receive less funding (Levaçić and March, 2007). The selective system is a clear driver towards increased social and economic segregation between schools, and all of the dangers that this entails – such as lower self-­esteem and aspiration, poorer role models, poorer relationships and distorted sense of justice (Gorard and See, 2013). The potential implications for policies and practices based on calculating a pupil premium attainment gap are substantial. The findings mean that when policymakers,

Refining Measures of Poverty and Student Progress in England

99

advocates of the success of the London Challenge, OFSTED, RAISE, the pupil premium Champion, awards committees and others use the pupil premium gap as a measure of success they are probably and unwittingly being very unfair. It has already been suggested that there is a problem for all such calculations caused by missing data, and because they take no account of the proportion of local residents using private schools – with both currently ignored in the calculation of any pupil premium attainment gap (and, as shown above, both influencing the calculation by their absence). What this chapter shows more importantly is that they are unfair because they do not take account of the threshold nature of FSM-eligibility. They are ignoring the variation within that category stratified by prior educational challenges like SEN and EAL, and then again by the qualification outcomes used to calculate the gap. Almost as importantly, the analysis shows that different areas have different proportions of the three FSM pupil groups. Heavily disadvantaged areas are likely to have more of the always FSM-eligible pupils, and this makes any comparison with other areas based on the pupil premium gap intrinsically invalid. This is in no way an argument against the pupil premium policy itself, but it does suggest that the impact of the policy needs a rather more robust evaluation than simply measuring changes in the pupil premium attainment gap. Perhaps just as importantly, this chapter has implications for the delivery of the pupil premium itself. Currently these extra resources are given to schools on the basis of the number of pupils in that school who have ever been eligible for free school meals (for the previous six years). This means that schools not only miss out the extra money when data is missing, but that those schools taking the most disadvantaged pupils (likely to attain the lowest at KS4) get the same per capita as those who take the pupils moving in and out of eligibility. Currently, until all else is resolved it would make more sense to allocate the pupil premium primarily on the basis of pupils eligible for FSM at the time of allocation, and then to update this every year throughout their school career. This would mean money going to the schools of those most in need, while they are most in need.

Acknowledgements This work is funded by the ESRC: grant number ES/N012046/1.

References Alegre, M. and Ferrer, G. (2010), ‘School Regimes and Education Equity: Some Insights Based on PISA 2006’, British Educational Research Journal, 36 (3): 433–61. Atkinson, A., Gregg, P. and McConnell, B. (2006), The Results of 11+ Selection, CMPO Working. Bolton, P. (2016), Grammar School Statistics, House of Commons Library Briefing Paper 1398. Collado, D., Lomos, C. and Nicaise, I. (2015), ‘The Effects of Classroom Socioeconomic Composition on Students’ Civic Knowledge in Chile’, School Effectiveness and School Improvement, 26 (3): 415–40.

100

Understanding School Segregation

Crawford, C., MacMillan, L. and Vignoles, A. (2014), Progress Made by High-­attaining Children from Disadvantaged Backgrounds, London: Social Mobility and Child Poverty Commission, http://dera.ioe.ac.uk/20433/1/High_attainers_progress_report_final.pdf. Cribb, J., Sibieta, L. and Vignoles, A. (2013), Entry into Grammar Schools in England, Institute for Fiscal Studies, https://www.ifs.org.uk/docs/Grammar_Schools2013.pdf. Danhier, J. (2017), ‘How Big is the Handicap for Disadvantaged Pupils in Segregated School Settings?’, British Journal of Educational Studies, doi: 10.1080/00071005.2017.1322682. Danhier, J. and Martin, E. (2014), ‘Comparing Compositional Effects in Two Education Systems: The Case of the Belgian Communities’, British Journal of Educational Studies, 62 (2): 171–89. DfE (2017), Analysing Family Circumstances and Education, April 2017, file:///C:/Users/ nbuser/Downloads/Technical%20consultation&20document%20(1).pdf. Financial Times (2016), ‘Better School Results are More Important than Improved Transport Links to Narrow the North-South Divide’. Available online: https://www.ft. com/content/0c06cec2-b6f4–11e6–961e-­a1acd97f622d. Foster, D., Long, R. and Roberts, N. (2016), Grammar Schools in England, House of Commons Library Briefing Paper 7070. Gallagher, T. and Smith, A. (2000), The Effects of the Selective System of Secondary Education In Northern Ireland, Bangor: Department of Education. Available online: http://www.education.gg/CHttpHandler.ashx?id=97491&p=0. Goldsmith, P. (2011), ‘Coleman Revisited: School Segregation, Peers, and Frog Ponds’, American Educational Research Journal, 48 (3): 508–35. Gorard, S. (2000), ‘ “Underachievement” is Still an Ugly Word: Reconsidering the Relative Effectiveness of Schools in England and Wales’, Journal of Education Policy, 15 (5): 559–73. Gorard, S. (2010), ‘Serious Doubts about School Effectiveness’, British Educational Research Journal, 36 (5): 735–66. Gorard, S. (2012), ‘Querying the Causal Role of Attitudes in Educational Attainment’, ISRN Education, Volume 2012, Article ID 501589, 13 pages, doi:10.5402/2012/501589. Gorard, S. (2015a), ‘The Complex Determinants of School Intake Characteristics, England 1989 to 2014’, Cambridge Journal of Education, 46 (1): 131–46. Gorard, S. (2015b), ‘The Uncertain Future of Comprehensive Schooling in England’, European Educational Research Journal, 14 (3–4): 257–68. Gorard, S. (2016), ‘A Cautionary Note on Measuring the Pupil Premium Attainment Gap in England’, British Journal of Education, Society and Behavioural Sciences, 14 (2): 1–8. Gorard, S. and See, B.H. (2013), Overcoming Disadvantage in Education, London: Routledge. Gorard, S., Hordosy, R. and See, B. H. (2013), ‘Narrowing the Determinants of Segregation between Schools 1996–2011’, Journal of School Choice, 7 (2): 182–195. Gov.UK (2014a), ‘Pupil Premium: Funding for Schools and Alternative Provision’. Available online: https://www.gov.uk/pupil-­premium-information-­for-schools-­andalternative-­provision-settings. Gov.UK (2014b), ‘The Pupil Premium – An Update’. Available online: http://www.ofsted. gov.uk/sites/default/files/documents/surveys-­and-good-­practice/t/The%20pupil%20 premium%20-%20an%20update.pdf. Gov.UK (2015), ‘Pupil Premium Reviews’. Available online: https://www.gov.uk/pupil-­ premium-reviews.

Refining Measures of Poverty and Student Progress in England

101

Halsey, A. and Gardner, L. (1953), ‘Selection for Secondary Education and Achievement in Four Grammar Schools’, The British Journal of Sociology, 4 (1): 60–75. Harris, D. and Williams, J. (2012), ‘The Association of Classroom Interactions, Year Group and Social Class’, British Educational Research Journal, 38 (3): 373–97. Henderson, B. (2008), ‘The Importance of Districts’, School Effectiveness and School Improvement, 19 (3): 261–74. House of Commons Library (2017), Grammar School Statistics, Briefing Paper #1398. Available online: http://researchbriefings.files.parliament.uk/ documents/SN01398/SN01398.pdf. Hutchings, M., Greenwood, C., Hollingworth, S., Mansaray, A. and Rose, S. with Minty, S. and Glass, K. (2012), Evaluation of the City Challenge Programme, DFE Research Report DFE-RR215. Available online: https://www.gov.uk/government/uploads/ system/uploads/attachment_data/file/184093/DFE-RR215.pdf. Jenkins, S., Micklewright, J. and Schnepf, S. (2008), ‘Social Segregation in Secondary Schools: How Does England Compare with Other Countries?’, Oxford Review of Education, 34 (1): 21–37. Kalogrides, D. and Loeb, S. (2013), ‘Different Teachers, Different Peers’, Educational Researcher, 42 (6): 304–16. Levaçić, R. and Marsh, A. (2007), ‘Secondary Modern Schools: Are Their Pupils Disadvantaged?’, British Educational Research Journal, 33 (2): 1469–1518. Lloyds Bank (2016), ‘Parents Willing to Pay £53,000 More to Live Near a Top School’. Available online: http://www.lloydsbankinggroup.com/Media/Press-Releases/2016press-­releases/lloyds-­bank/house-­prices-near-­schools. Manning, A. and Pischke, J. (2006), Comprehensive Versus Selective Schooling in England and Wales: What Do We Know?, IZA Discussion Paper 2072. Available online: http://ssrn.com/abstract=898567. Muller, M. and Hofmann, V. (2016), ‘Does Being Assigned to a Low School Track Negatively Affect Psychological Adjustment?’, School Effectiveness and School Improvement, 27 (2): 95–115. Noden, P. and West, A. (2009), Attainment Gaps between the Most Deprived and Advantaged Schools, London: The Sutton Trust. OECD (2014), Education at a Glance 2014: OECD Indicators, Paris: OECD Publishing. OFSTED (2015), ‘Analyse School Performance (RAISE)’. Available online: https://www. raiseonline.org/login.aspx?ReturnUrl=%2f. Parker, P., Jerrim, J., Schoon, I. and Marsh, H. (2016), ‘A Multination Study of Socioeconomic Inequality in Expectations for Progression to Higher Education’, American Educational Research Journal, 53 (1): 6–32. Prais, S. (2001), ‘Grammar Schools’ Achievements and the DfEE’s Measures of Value-­added: An Attempt at Clarification’, Oxford Review of Education, 27 (1): 69–73. Pupil Premium Awards (2015), http://www.pupilpremiumawards.co.uk. Roew, E. and Lubienski, C. (2017), ‘Shopping for Schools or Shopping for Peers’, Journal of Education Policy, 32 (3): 340–56. Schmidt, W., Burroughs, N., Zoido, P. and Houang, R. (2015), ‘The Role of Schooling in Perpetuating Educational Inequality: An International Perspective’, Educational Researcher, 44 (7): 371–86. Siddiqui, N. (2017), ‘Socio-­economic Segregation of Disadvantaged Children between Schools in Pakistan: Comparing the State and Private Sector’, Educational Studies, 43 (4): 391–409.

102

Understanding School Segregation

Strand, S., and Winston, J. (2008), ‘Educational Aspirations in Inner City Schools’, Educational Studies, 34 (4): 249–67. Sullivan, A., Parsons, S., Wiggins, R., Heath, A. and Green, F. (2014), ‘Social Origins, School Type and Higher Education Destinations’, Oxford Review of Education, 40 (6): 739–63. The Guardian (2016), ‘Ofsted Chief Calls for Radical Shakeup to Close Widening Skills Gap’. Available online: https://www.theguardian.com/education/2016/dec/01/england-­ faces-widening-­skills-gap-­says-outgoing-­ofsted-chief. The Guardian (2017), ‘George Osborne: North-South Divide in Schools Needs Urgent Attention’. Available online: https://www.theguardian.com/society/2017/feb/03/ george-­osborne-england-­north-south-­divide-schools-­needs-urgent-­attentioneducation-­gap. Vieluf, S., Hochweber, J., Klieme, E. and Kunter, M. (2015), ‘Who Has a Good Relationship with Teachers?’, Oxford Review of Education, 41 (1): 3–25. Yeung, R. and Phuong Nguyen-Hoang, P. (2016), ‘Endogenous Peer Effects: Fact or Fiction?’, The Journal of Educational Research, 109 (1): 37–49.

6

An Evaluation of the Intensity and Impacts of Socio-­economic School Segregation in Argentina Natalia Krüger

1.  Introduction An education system’s performance is generally judged in light of several essential goals, such as the development of socially meaningful learning, the provision of equal opportunities and the redistribution of sociocultural capital to foster social equality and cohesion. Accordingly, educational inclusion implies not only striving for universal coverage, but also offering schooling conditions that may compensate for the impact of initial disparities, in order to promote general excellence in achievement and democratic interactions at school. Like other Latin American countries, Argentina has made significant progress during the past decades in terms of increasing access to all levels of schooling. Social sectors that had historically been excluded from formal education have gradually gained entry. However, the country remains far from offering high-­quality and equal educational opportunities to all of its citizens. National and international evaluations of academic performance have revealed that the skills acquired by Argentine students are relatively low on average and strongly associated with SES (socio-­economic status). Additionally, there is an evident horizontal differentiation of schools. As stated by Sen and Kliksberg (2007), even within the education system there are circumstances of exclusion that respond to ‘inclusion in uneven conditions’. Socio-­economic segregation is one of the many forms of such unequal inclusion and, according to the scarce available evidence, it currently represents a great problem in Latin America. A school system is said to be segregated when its students are unevenly distributed across schools according to some social characteristics (Jenkins et  al., 2008), so that schooling experiences develop between peers of similar social status. This phenomenon is intrinsically relevant because it hinders the contribution of education to social cohesion. School may be thought of as a privileged place to promote socialization between different sectors, building a democratic citizenship and fostering future upward mobility and personal development. These opportunities

104

Understanding School Segregation

are lost when social interaction is substantially reduced (Gorard, 2009; Valenzuela et al., 2013). Furthermore, socio-­economic segregation causes concern because of its potential impact on inequality of educational outcomes. Although there is still no consensus on the matter, the balance of evidence supports the existence of compositional effects (Rumberger and Palardy, 2005; Willms, 2006; Schindler Rangvid, 2007; Lauder et al., 2010; Van Ewijk and Sleegers, 2010; Benito et  al., 2014). This means that pupils’ aggregate characteristics may affect achievement levels even after individual background has been taken into account (Dumay and Dupriez, 2008), through different direct and indirect mechanisms. Directly, school composition may have an influence through a channel known as peer effects, which acknowledges that interaction among students might create a context that either facilitates or hampers learning. This responds to the fact that attitudes, aspirations and social norms, which are likely associated with socio-­economic background, tend to be transmitted from one student to another (Palardy, 2013). Indirectly, the student body composition may serve as a proxy for, or even interact with, other school covariates that affect achievement. Schools with high-SES intakes tend to have other advantages: superior managerial and pedagogical practices; more or better resources; higher teacher expectations and engagement; greater support from parents; healthier school climate, etc. (Willms, 2006; Alegre, 2010; Lauder et al., 2010; Van Ewijk and Sleegers, 2010; Krüger, 2011; Palardy, 2013). Despite a growing interest in these topics, empirical studies focusing on the intensity, causes or consequences of socio-­economic school segregation in Latin America have been relatively limited. The aim of this chapter is to present an overview for the Argentine secondary school system, beginning to fill the gap in the literature and also in the hope that studying this particular case contributes to reaching a better understanding of the global phenomenon. I employ PISA (Programme for International Student Assessment) 2012 data1 to analyse the magnitude and main characteristics of the problem, as well as its impact on the inequality of student achievements. Following a review of the local literature, I estimate several indices in order to assess total segregation and how it may be decomposed along school sectors and alternative SES groups. In line with previous studies, the findings suggest that the degree of socio-­ economic segregation in Argentina is significant, less so than in other Latin American countries but much higher than in most OECD (Organisation for Economic Co-­ operation and Development) countries. In such a context, if significant school compositional effects exist, they will explain an important proportion of the variance in student achievement. To evaluate this situation, I estimate a series of multilevel models and find that, as expected, the process of social segregation is reinforcing the association between educational success and socio-­economic background. Thus, it should be considered a key educational policy issue. Due to a minor re-­structuring of secondary schools, the quality of the PISA 2015 sample for Argentina has been called into question (http://www.oecd.org/pisa/data/2015database). However, no problems have been detected regarding the 2012 survey.

1

Intensity and Impacts of Socio-economic School Segregation in Argentina

105

2.  Background information Argentina has a federal system of government and a decentralized education system, so that the funding and operation of public schools is a responsibility of each Province. Attendance is compulsory from the age of 4 to 18 and, depending on the Province, secondary education might begin at the age of 12 or 13. Besides public provision, educational services are offered by private actors under state regulations, and parents may freely choose the type of institution for their children. Private schools are more autonomous and can select their teachers and authorities, define their own pedagogical programmes and freely assign their resources, which are generally higher in quantity and quality. Moreover, they may approve or refuse admission to students according to their own criteria. It is worth noting that within the private sector, around 65 per cent of the schools receive government subsidies – destined to cover teacher salaries partially or totally – while the rest are independent. The former group of institutions possess relatively less autonomy regarding the curriculum and charge lower fees to families, since the amount they are allowed to charge decreases as the proportion of government funding rises (Krüger, 2011, 2014). The current configuration of the education system is the result of multiple transformations it has undergone since its foundation near the end of the nineteenth century. During the last few decades, in particular, there have been two radical waves of reforms,2 aimed at increasing access, quality and equity in basic education. They included further movements towards decentralization, compromises to raise public expenditure, several extensions of the compulsory school attendance period and the introduction of organizational and curricular innovations. Meanwhile, the political and socio-­economic context has experienced periods of rise and decline, including a deep crisis at the end of the twentieth century. In spite of recent recovery, serious problems persist, such as a high degree of labour informality, and substantial income inequality and poverty levels (Rivas et al., 2010). In fact, Argentina has recently lost its position amongst the countries with the lowest levels of social inequality in Latin America (Gasparini and Cruces, 2010). The social structure has become progressively fragmented, and the decline in social cohesion has made itself evident, for example, through the significant growth in residential segregation (Veleda et al., 2011). As a consequence of these changes, considerable progress has been made regarding some educational outcomes, such as access, while results have been disappointing in terms of quality and equality (Krüger et al., 2015).

3.  Previous research Although the problem of socio-­economic school segregation has been recognized for decades in Latin America, very few studies have systematically attempted to measure it.

National Laws No. 24.049 (1991), 24.195 (1993), 24.521 (1995), 26.058, 26.075 (2005) and 26.206 (2006).

2

106

Understanding School Segregation

Among those studies are the works by Vázquez (2011),Valenzuela et al. (2013), Arcidiácono et al. (2014), Tapia and Valenti (2016) and Murillo and Martínez-Garrido (2017). They employ a variety of sources, such as the LLECE (Latin American Laboratory for Assessment of the Quality of Education) surveys for primary schools, the PISA surveys for secondary schools, and National Household Surveys or special educational surveys in each country. In Argentina, Gasparini et al. (2011) use the National Household Survey, while Krüger (2014) uses PISA 2000 and 2009 data. In general, the evidence points to: i) high degrees of socio-­economic segregation in the region, greater than those found in more developed systems; ii) rising levels during the past two decades; iii) important gaps between countries and an intermediate position for Argentina; and iv) relevant segregation between the public and private sectors, but also within each network. Comprehending the causes of this phenomenon is a complex task, since it presents multiple determinants and different dynamics in each country, responding to factors that are endogenous and exogenous to the education system. While it is a relevant problem in many developed countries, it is particularly so in Latin America due to the socio-­economic and cultural fragmentation that characterizes our societies. In such unequal social contexts, the expansion of the education systems seems to have strengthened different forms of marginalization, reproducing external inequalities within the school sector. For instance, school segregation closely reflects the broader decline in social interaction in public spaces, manifested in residential segregation3 or the growing privatization of public services (Rivas et al., 2010). Research on these issues in Argentina dates back to the 1980s – some recommended texts are those by Braslavsky (1985), Fiszbein (2001), Tiramonti (2004), Llach (2006), Veleda (2009, 2014) and Narodowski et  al. (2016). These authors argue that during the past decades, structural aspects of our society and institutional characteristics of the school sector have combined to determine the current levels of segregation. It is often stressed that the gradual withdrawal of the state from the provision of public services has prompted the development of quasi-­market mechanisms, which neglect the goal of improving equality. Simultaneously, as Tiramonti (2004) or Del Cueto (2004) explain, social polarization has promoted the isolation of middle and upper classes, who seem to seek spaces of socialization ‘among equals’ in order to preserve their social position. Moreover, differences in the quality of educational services between schools are being increasingly perceived. As a result, the demand for schooling has become more selective, taking into account academic standards as well as schools’ ability to transmit sociocultural capital. Thus, while families implement strategies to choose centres according to their preferences and purchasing power, schools develop competitive interdependencies and try to select their student population (similar trends have been observed in European countries by Maroy and Van Zanten, 2009; Alegre, 2010; and Alegre and Ferrer, 2010). These practices reinforce the differentiation between schools even though, as Narodowski et  al. (2016) point out, there are no

Although the available evidence is scarce, studies such as Salvia and De Grande (2008) or Groisman (2010) show how socio-­spatial divisions in cities are a main channel for the reproduction of poverty in the country.

3

Intensity and Impacts of Socio-economic School Segregation in Argentina

107

school choice mechanisms like demand-­side subsidies, voucher programmes or charter schools. One of the elements usually considered is the participation of the private sector, with its particular rules and conditions. Currently, around 28 per cent of Argentine students in secondary education attend private schools, and the size of the network has been steadily growing since the 1960s, owing to several factors. First, support from the government in the form of supply-­side subsidies to cover teacher salaries and progressive deregulation is significant. As Narodowski et  al. (2016) explain, although there are no explicit public–private partnerships, the state is increasingly relying on private provision to reduce the costs of including a growing student population. Second, there seems to be a latent demand for private schooling, so that as socio-­economic conditions improve, families leave the traditional public sector. This tendency is related to the gradual loss of prestige by public education and the fact that, for many families, selecting the group of peers seems to be a way to ensure the reproduction of their social position (Rivas et  al., 2010; Krüger, 2014; Narodowski et al., 2016). Private schools naturally attract students from a more privileged socio-­economic background for several reasons: they are usually located in wealthier neighbourhoods; they are able to offer more diversified educational services; fees act as direct assignment mechanisms; and school authorities control admission requirements. However, especially since the 1990s, the middle classes and even some low-­income sectors have been migrating from the public to the private sector. To describe it simply: free state-­ run schools seem to be absorbing the newcomers from vulnerable social sectors, while the middle classes attend the more affordable subsidized private schools and the privileged population increases its isolation within elite or independent schools. Thus, although the relative size of both networks has remained fairly constant, their social composition is continuously changing (Krüger, 2014). The other instance of social segregation takes place within the public sector, where different mechanisms are at play. At the end of the 1980s, parents were given the freedom to choose a school for their children. However, some restrictions apply to this choice, since each public school must respect an order of priority during admission: first, students are guaranteed a place within the same school as they move on to upper levels of education; second, siblings of current students and children of the staff are admitted; then, the proximity of the school to the home or the parents’ workplace is considered; finally, remaining places are assigned through a lottery. Thus, the official rules directly link residential segregation to school segregation. Additionally, more informal mechanisms operate to reinforce segregation: different authors have revealed that the admission process frequently lacks transparency, so that decisions are made discretionarily by school authorities while families apply different strategies to avoid regulations (Fiszbein, 2001; Rivas et al., 2010; Veleda, 2014; Narodowski et al., 2016). On the one hand, Maroy and Van Zanten’s (2009) analysis for six European localities fully applies to the Argentine case. They claim that schools tend to engage in a first- and second-­order competition for students: first, enlarging the student population is important in order to increase the number of teachers and the amount of resources received; then, the characteristics of the students are selected according to their

108

Understanding School Segregation

perceived impact on working conditions, school prestige and teacher satisfaction.4 Some school authorities avoid opening a lunchroom to give free meals because it may attract a more vulnerable population, others review academic records or are influenced by personal contacts during admission, and many try to transfer undisciplined or low-­ achieving students to other centres during the school year. This is favoured by the insufficient control applied by school supervisors. School choice by families, on the other hand, is strongly influenced by their sociocultural capital, since it conditions their means of seeking and analysing information, their motivations and expectations, and their ability to avoid official regulations. For example, some families use personal contacts and professional prestige to negotiate with school authorities, others choose or even alter their home addresses considering the schools in the proximity, and many begin choosing the best school circuits during preschool. Regarding the possible impact of segregation on educational equality, the evidence available for the Latin American region is, again, scarce. Nevertheless, several studies claim that school composition is a relevant determinant of student achievement in primary and secondary school (Willms and Somers, 2001; Somers et al., 2004; Treviño et  al., 2010; Cervini, 2012; Murillo and Martínez-Garrido, 2017, among others). In Argentina, authors like Cervini (2003), Formichella (2011) and Marchionni et  al. (2013) have included school composition in general studies of the explanatory factors of educational outcomes, agreeing on its significance. More specific analyses of peer effects may be found in Cervini (2005), Lugo (2011) or Krüger (2013), who arrive at a similar conclusion.

4.  Data and variables This study draws on data from the PISA project, which targets 15-year-­old students attending an educational institution in grade seven or higher in OECD and partner countries (OECD, 2014). The 2012 wave of the survey focuses on mathematics and covers eight Latin American countries. The Argentine sample consists of 226 schools and 5,9085 students who, once weighted, represent about 80 per cent of the 15-year-­old population. It should be acknowledged that, as it occurs with other Latin American participants, this relatively low coverage rate may introduce an estimation bias, because excluded students are more likely to be low achievers or belong to vulnerable social sectors (Hanushek and Woessmann, 2011; Gamboa and Waltenberg, 2012). Consequently, average scores might be overestimated and performance inequality and social segregation underestimated for this more homogeneous and higher-SES population.6

So far, no incentive mechanisms linked to student achievement, such as the publication of school rankings or teacher pay-­for-performance, have been applied in the country. 5 After dropping six schools with fewer than five students, the sample finally consists of 5,882 students. 6 Also, since the PISA study targets students and not schools, the estimates of segregation between centres might lose precision. However, they are valid for cross-­country comparisons and for showing general trends. Moreover, this is the only reliable source of information in Argentina that allows the study of the socio-­economic composition of high schools. 4

Intensity and Impacts of Socio-economic School Segregation in Argentina

109

The variables were chosen pursuing the following objectives: i) measuring the level and effects of socio-­economic school segregation; ii) estimating compositional effects, including the usual controls in the educational production function (Somers et  al., 2004; Calero and Escardíbul, 2007; Van Ewijk and Sleegers, 2010; Alegre and Ferrer, 2010); and iii) reducing the potential selection bias. In a segregated scenario, the ‘double process of selection’ between schools and families might be a source of endogeneity. In order to mitigate this problem, a reasonable option is to include a rich set of controls, trying to capture some variables which may account for this process (Somers et  al., 2004; Schindler Rangvid, 2007; Andersen and Thomsen, 2011, etc.). Following this approach, it should be emphasized that selection bias remains possible, so that results may not be interpreted in terms of causality.

4.1  Student-­level variables l

l l

Educational outcome: PISA mathematics test performance. As in the past, Argentina presents poor outcomes in 2012, with average results (388 points) far below the OECD mean of 500 points, about two-­thirds of its students performing below a minimum proficiency level (OECD, 2013) and a relevant inequality of results. Personal traits: age, gender, preschool attendance, modal grade attendance. Family traits: economic, social and cultural status, measured through the PISA ESCS index, which summarizes information regarding parental education and occupational status and the amount of wealth, cultural and educational resources available at home. This index presents relatively low levels in Latin America and a high degree of dispersion due to socio-­economic inequality.

4.2  School-­level variables l

l

l

Socio-­economic composition: mean school ESCS index; percentage of students in different quartiles of the index. General school traits and teaching–learning conditions: proportion of girls; school size; school type (public, private independent and private government-­ dependent7); PISA indices of quality of schools’ educational resources, quality of physical infrastructure, teacher shortage, creative extracurricular activities at school and school responsibility for curriculum and assessment. Determinants of the double process of selection: tracking or ability grouping within schools; likelihood of students being transferred to other schools due to bad behaviour or low achievement; parental pressure; competition with other schools in the area; index of academic selectivity during admission.

Government-­dependent private schools receive more than 50 per cent of their funding from public sources.

7

110

Understanding School Segregation

5.  Estimating the degree of socio-­economic school segregation Two dimensions of segregation, among those identified by Massey and Denton (1988), are especially relevant in the educational context: evenness and exposure. Evenness indicates the uniformity in the distribution of each group of students, while exposure indicates the degree to which a group comes in contact with others at school. I estimate three indices: the popular Index of Dissimilarity (Duncan and Duncan, 1955) and the Mutual Information Index (Theil, 1971; Frankel and Volik, 2011), to measure evenness; and the Index of Isolation (Massey and Denton, 1988), to measure exposure.

5.1  Index of Dissimilarity After dividing the student population into a minority and a majority, this index may be interpreted as the proportion of students from the minority who should be transferred to more advantaged schools, so that all centres include an equal share of this group. Maximum segregation corresponds to a value of one and minimum segregation to a value of zero. Massey and Denton (1993), in Cutler et al. (1999), maintain that segregation may be considered low when the value is below 0.3, moderate when it lies between 0.3 and 0.6, and high when it exceeds 0.6. According to Allen and Vignoles (2007), this index meets most of the criteria required of a ‘good index of segregation’. Its main drawback, however, is that it relies on dichotomous characterizations of the population.

5.2  Mutual Information Index This index represents an ordering that satisfies most substantive axioms considered in the literature (Frankel and Volij, 2011). Moreover, it is suitable for the multigroup context, estimating both the overall level of segregation and the segregation between each possible pair of groups (Leckie and Goldstein, 2015). According to Frankel and Volij (2011) or Mora and Ruiz-Castillo (2011), this index is related to the concept of entropy: in a given school system, if one were to draw a student at random, the uncertainty about her or his SES would be measured by the entropy of the system’s socio-­economic distribution; the higher the level of segregation, the higher the reduction in uncertainty after identifying the school the student attends. The Mutual Information Index is thus given by this difference in entropy, averaged over the schools in the system. Its value may be interpreted as the difference between the socio-­economic diversity of the school system and the weighted average diversity of individual schools (Leckie and Goldstein, 2015). Its lower bound is zero, representing complete integration, and its upper bound is the entropy of the system’s socio-­economic distribution, indicating complete segregation (Mora and Ruiz-Castillo, 2011). It also has the advantage of being additively decomposable along groups and along units. In our case, this means that overall segregation may be expressed either as the sum of a between-­groups term and a within-­groups term – combining different socio-­ economic groups into supergroups – or as the sum of a between-­units and a within-­units term – combining schools into superunits, such as the public and private school sectors.

Intensity and Impacts of Socio-economic School Segregation in Argentina

111

5.3  Index of Isolation This index indicates the probability that a minority student drawn at random shares a school with another member of the minority. Thus, it is a suitable complement for the indices that measure evenness. Its lower bound is the proportion of the total population represented by the minority, while its upper bound is one. The next decision after selecting the segregation indices concerns the indicator of SES, in this case the PISA ESCS index, and its division into groups. For the estimation of the Dissimilarity and Isolation indices, I considered two alternative thresholds: the disadvantaged minority (bottom-SES or Q1) includes students whose ESCS index is below the twenty-­fifth percentile; while the privileged minority (top-SES or Q4) includes students with an ESCS index above the seventy-­fifth percentile. For the Mutual Information Index, initial calculations describe the overall segregation between all four SES-quartile groups (Q1, Q2, Q3 and Q4). Then, I combined them into two supergroups, where the low-SES group consists of Q1 and Q2 students, and the highSES group consists of Q3 and Q4 students. This allowed me to decompose the index into segregation between and within each of these supergroups. Additionally, I partitioned total segregation into between and within school-­sectors components. I estimated these indices for the Argentine school system and the other PISA participants, clustered into regions: Latin America, OECD and non-OECD (Tables 6.1 and 6.2). In the case of the Dissimilarity Index, I evaluated the significance of the gaps between each region and Argentina using Ransom’s (2000) statistic, built from bootstrapped standard errors. Also, in all cases I carried out robustness checks redefining the SES groups, finding no significant variations in results. First, the estimations of the Dissimilarity Index suggest that segregation levels in Argentina are moderate to high, significantly lower than the Latin American average, but significantly higher than the OECD and non-OECD averages. Considering the bottom-SES minority, it would be necessary to transfer 45 per cent of these students to more privileged schools in order to reach an even distribution of the population. At this point, the low coverage rates of the PISA samples in Latin America should be recalled, as evidence of previous forms of educational exclusion that have probably Table 6.1  Overall socio-­economic school segregation levels Dissimilarity Index Bottom-SES minority Argentina Latin American average OECD average Non-OECD average

0.453 (0.014) 0.481 (0.011)**

Mutual Top-SES minority Information Index 0.480 (0.015) 0.502 (0.012)*

0.377 (0.013)*** 0.394 (0.013)*** 0.389 (0.016)*** 0.418 (0.016)***

Isolation Index Bottom-SES minority

Top-SES minority

0.270 0.309

0.387 0.403

0.499 0.522

0.176 0.193

0.353 0.360

0.386 0.402

Notes: bootstrapped SE in parentheses. ***, **, * the estimated values of D for the region and Argentina significantly differ at the 1 per cent, 5 per cent or 10 per cent level. Source: author’s analysis based on data from PISA 2012.

Understanding School Segregation

112

Table 6.2  M decompositions Along SES supergroups

Argentina Latin American average

Along school-­sector superunits

M between low-SES and high-SES groups

M within low-SES group

M within high-SES group

0.171 (63.3%) 0.181 (58.7%)

0.084 (15.5%) 0.052 (17.4%)

0.114 (21.2%) 0.073 (23.9%)

M between M within M within public and public private private sectors sector sector 0.063 (23.4%) 0.088 (28.5%)

0.188 (44.2%) 0.189 (50.0%)

0.240 (32.4%) 0.243 (21.5%)

Note: in parentheses are the proportions of segregation once each partition has been weighted by its demographic importance. Source: author’s analysis based on data from PISA 2012.

selected a socially advantaged student population. Thus, these figures are likely underestimating the true levels of segregation. An additional result is that the distribution of privileged students seems to be more uneven than that of their low-­background counterparts, in Argentina as well as in the other regions. Moving on to the Mutual Information Index that avoids the arbitrary definition of a minority group, the estimations confirm the relative positions of Argentina and the Latin American region. Yet more interesting are the decompositions of the index (Table 6.2). First, considering SES supergroups, the most relevant proportion of total segregation is due to the separation between the low-SES (Q1 and Q2) and high-SES (Q3 and Q4) groups. In addition, a significant share of segregation is explained by the distribution within each of these groups, with the separation of Q4 students from those who belong to the adjacent category, Q3, being especially relevant. Second, the decomposition along school sectors shows that a relatively small proportion (23 per cent) is due to segregation between the public and private school networks, so that the lion’s share of total segregation responds to inequalities within each of these sectors. As expected, segregation is higher inside the private sector, where fees constitute the main assignment mechanism.8 However, segregation within the public sector is also quite significant, confirming that other informal or non-­explicit mechanisms are at work. Finally, the Isolation Index again confirms that the greatest levels of social exclusion correspond to the Latin American countries. In Argentina, if one were to draw at random a student from the bottom-SES minority, she or he would have a 39 per cent probability of meeting another minority student at school; this probability rises to

Furthermore, within the private sector only 5 per cent of the overall level of segregation corresponds to segregation between the independent and the government-­dependent sub-­sectors; the highest proportion of segregation (79 per cent) takes place within the government-­dependent sector and the remaining 16 per cent is due to segregation within the independent sector. This is due both to the greater size of the government-­dependent sub-­sector (it serves about 80 per cent of the student population attending private schools) and to the more unequal distribution of its students. (This last result contradicts past research findings based on PISA 2000 and 2009 data, and it should be confirmed by other sources of information before venturing an explanation.)

8

Intensity and Impacts of Socio-economic School Segregation in Argentina

113

50 per cent if the student belongs to the top-SES minority. Thus, the experience of segregation for Argentine students is quite high, considering that each minority group only represents 25 per cent of the total student population. In other words, the likelihood of being in contact with students from different socio-­economic groups at school is low, reducing the potential promotion of social cohesion. Especially relevant is the isolation of the privileged minorities, which may prevent their development of empathy and aversion to inequality of opportunities.

6.  Socio-­economic segregation and inequality of outcomes In addition to its intrinsic importance, the hypothesis is that socio-­economic segregation in the Argentine secondary school system is reinforcing the effect of individual SES on educational results. In order to test this, the determinants of student performance on the PISA math test are evaluated through the estimation of an education production function,9 in which a key explanatory factor is the socio-­ economic composition of the school. If it has a positive effect, controlling for individual background, and it helps to explain a significant amount of the variance in results, then the hypothesis is supported. Following the mainstream literature in educational research, I estimate the compositional effects through a multilevel analysis (Opdenakker and Van Damme, 2001; Somers et al., 2004; Rumberger and Palardy, 2005; Willms, 2006; Lauder et al., 2010; Benito et  al., 2014). This methodology is usually recommended when the structure of the data is hierarchical, because it leads to more efficient estimates and provides a better representation of complex effects compared to the traditional Ordinary Least Squares (OLS) method (Goldstein, 1995; Hox, 2002; Calero and Escardíbul, 2007; etc.). Given the structure of the PISA samples, I estimated two-­level hierarchical models, with information regarding the students and the schools they attend. The first step usually involves the estimation of a null or empty model in order to decompose the variance of outcomes: the variance of the student-­level error term represents the variation in achievement verified between students within-­schools, while the variance of the school-­level error term represents variation in achievement between-­schools. Then, the intra-­class correlation coefficient (ICC) is calculated as the proportion of total variance due to variation between-­schools. After the estimation of the null model, I introduced the explanatory variables sequentially (Tables 6.3 and 6.4), following Bryk and Raudenbush (1992) and Dumay and Dupriez (2008): Model 1 includes the student-­level indicators; then, Model 2a adds only the school composition variable and Model 2b adds all of the school variables

It should be acknowledged that educational outcomes are multiple and complex, so that reducing them to scores on a standardized test is an oversimplification. Additionally, the PISA database does not allow the consideration of the cumulative effects of explanatory factors, since they are measured contemporaneously with achievement. Thus, the impact of segregation on equality of educational results is probably underestimated.

9

114

Understanding School Segregation

Table 6.3  Multilevel models to estimate the school composition effect on maths scores – random effects Null model

Model 1

Model 2a

Model 2b

Model 3

Within-­school variance

3285.6

2768.1

2769.8

2747.3

2748.4

Between-­school variance

2678.3

1429.8

831.3

768.8

612.6

Total variance

5963.9

4197.9

3601.1

3516.1

3361.1

Within-­school variance explained (%)

15.7

15.7

16.4

16.3

Between-­school variance explained (%)

46.6

69.0

71.3

77.1

Total variance explained (%)

29.6

39.6

41.0

43.6

ICC

N Deviance

0.449

5882 5973024.0

5638 5623034.0

5638 5613101.5

4831 4617152.2

4673 4614084.6

Source: author’s analysis based on data from PISA 2012.

except for social composition; finally, Model 3 is the full model that includes the entire set of controls mentioned in Section 4. I also considered the fact that the peer effect may be non-­linear: it might be heterogeneous across students with different individual characteristics; there might be increasing or decreasing marginal effects; or the relevant compositional variable might be the proportion of students from a particular SES category. These alternatives are tested with the same methodology, including a square term of the mean-SES variable, interactions with individual SES or the proportion of students in each category (following Willms, 2006, or Lugo, 2011, among others). Additionally, I evaluated the robustness of results to different specifications. The final model was selected considering the value of the deviance, or likelihood ratio statistic, which decreases as the fit of the model improves (Hox, 2002). The first meaningful result is derived from the value of the ICC: 45 per cent of total variance lies at the school-­level, a proportion higher than that found for the OECD (37 per cent) and even several Latin American countries, confirming that horizontal segmentation is relevant in Argentina. Additionally, since the ICC significantly differs from zero, the existence of intra-­cluster correlation supports the estimation of a multilevel model instead of an OLS model (Hox, 2002). Certainly, owing to the fact that individual SES is an important predictor of scores, the substantial levels of social segregation in the country could explain the significant intra-­class correlation, even if there were no compositional effects. In fact, after incorporating the student-­level variables, the unexplained variation of results between schools was largely reduced (Model 1). This confirms the diagnosis of a relevant degree of social school segregation. Overall, this model accounts for about 30 per cent of total variance and 47 per cent of between-­school variance.

Intensity and Impacts of Socio-economic School Segregation in Argentina

115

Table 6.4:  Multilevel models to estimate the school composition effect on maths scores – fixed effects Variables

Constant Student level Age Girl Preschool attendance SES Modal grade attendance School level Mean SES

Coefficients Null model

Model 1

Model 2a

Model 2b

Model 3

384.09 (4.05)***

339.92 (4.80)***

343.44 (4.45)***

333.06 (9.33)***

342.74 (8.56)***

0.82 (3.19) 1.05 (3.19) 0.16 (3.32) –19.35 –19.66 –19.09 (1.93)*** (1.91)*** (2.16)*** 27.42 27.04 23.11 (3.81)*** (3.80)*** (3.72)*** 7.74 6.30 7.26 (0.97)*** (0.91)*** (0.91)*** 49.72 48.38 49.93 (2.18)*** (2.12)*** (2.40)***

Proportion of girls School size Private independent Private government-­ dependent10 Schools’ educational resources Schools’ physical infrastructure Teacher shortage Responsibility for curriculum Extracurricular activities Tracking or ability grouping Student transfers Parental pressure Competition between schools Academic selectivity

38.86 (4.98)***

Stand. Coeff. Model 3

0.46 (3.31) –19.21 (2.16)*** 22.88 (3.68)*** 6.31 (0.98)*** 49.67 (2.37)***

0.002 – 0.251***

28.90 (5.90)*** –34.94 (23.43) –26.72 (20.76) –0.01 (0.01) –0.01 (0.01)* 35.73 14.35 (13.81)*** (12.04) 33.63 18.20 (7.39)*** (8.94)** 3.85 (3.04) –0.17 (2.83)

0.252***

4.10 (2.67)

2.58 (2.56)

0.298*** 0.092*** 0.648***

–0.041 –0.017* 0.187 0.238** –0.002 0.042

2.85 (2.45) 1.73 (2.18) 12.21 (5.80)* 12.34 (5.02)** 0.18 (5.81) 1.60 (5.21)

0.023 0.106** 0.021

–4.65 (7.35) –2.82 (6.12)

–0.037

–7.90 (5.54) –1.71 (5.05) –7.35 (8.86) –10.00 (7.54) 14.53 (7.29)** 5.67 (6.68)

–0.022 –0.130

1.37 (7.27) –6.14 (5.74)

–0.080

0.074

Notes: robust standard errors in parentheses. Stand. Coeff. = standardized coefficients. *p