they will get there!prof. dr. d.c. van den boom ten overstaan van een door het college voor...

219
UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl) UvA-DARE (Digital Academic Repository) They will get there! Studies on educational performance of immigrant youth in the Netherlands van Welie, E.A.A.M. Link to publication Citation for published version (APA): van Welie, E. A. A. M. (2013). They will get there! Studies on educational performance of immigrant youth in the Netherlands. TIER. General rights It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons). Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible. Download date: 16 Mar 2021

Upload: others

Post on 15-Oct-2020

6 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

They will get there! Studies on educational performance of immigrant youth in theNetherlands

van Welie, E.A.A.M.

Link to publication

Citation for published version (APA):van Welie, E. A. A. M. (2013). They will get there! Studies on educational performance of immigrant youth in theNetherlands. TIER.

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 16 Mar 2021

Page 2: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

Top Institute for Evidence

Based Education Research

TIER Research SeriesUniversity of Amsterdam

They will get there!

Studies on Educational Performance of

Immigrant Youth in the Netherlands

Liesbeth van Welie

The educational performance of youth from immigrant descent has been a key subject in

political and public debate over the past decades. Typically, these debates focus on

underrepresentation in academic tracks in secondary school, and drop-out. The steady

increase, however, in enrolments in academic secondary levels among youth from immigrant

descent, seems almost overshadowed by the emphasis on problematic school careers. This

thesis shows that the average access to higher secondary tracks is slightly lower in largely

stratified neighbourhoods. Further research into three zip code areas with persistent

socioeconomic challenges, however, shows that pupils from Moroccan descent have better

chances to complete secondary school, when they do not switch schools. Furthermore, we

found that up to 90% of pupils in the four major Dutch cities choose another school than

the school nearest to their home. While native Dutch pupils on average prefer a school with

a lower percentage of migrant pupils, pupils from immigrant descent prefer a school with a

higher percentage of migrant pupils than is the case at the nearest school. For pupils from

immigrant descent, chances for upward mobility to a higher track increase slightly, but

significantly, with a lower distance between home and school. Finally, we found that relevant

and manageable existing scientific research, can adequately be matched with key questions

of diverse schools, and support these schools in the closing of the achievement gap.

Liesbeth van Welie graduated in Biology from Radboud University (Nijmegen,

the Netherlands), with a specialization in Aquatic Ecology. She started her career as a teacher

in Biology and was a few years later appointed as Principal of an innovative newly-founded

school for secondary education. Her second appointment as Principal was at a diverse

inner-city school in Amsterdam. Next, she worked as advisor to the Board of the University of

Amsterdam, and participated in several international networks in the field of quality

assessment of higher education, internationalization and organizational change. Her next

step was advisor to the Board of Maastricht University, followed by two years as senior

consultant at a consultancy firm. For several years she held the position of Chief Inspector of

secondary and higher education, where after she was asked to develop a new directorate for

the enhancement of evidence/ information based policymaking at the Ministry of Education,

Culture and Science. She worked on this doctoral thesis at the University of Amsterdam. They w

ill get the

re! Stud

ies o

n Educa

tiona

l Perfo

rma

nce o

f Imm

igra

ntYouth in the

Nethe

rland

s - Liesb

eth

van

We

lie

Page 3: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

  

THEY WILL GET THERE! 

Studies on Educational Performance of Immigrant Youth in the Netherlands 

   

Page 4: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

© Liesbeth van Welie, Amsterdam 2013 

 

All  rights  reserved.  No  part  of  this  publication  may  be  reproduced,  stored  in  a 

retrieval system or transmitted in any form, or by any means, electronic, mechanical, 

photocopying, recording, or otherwise, without the prior permission in writing, from 

the author. 

ISBN 978‐94‐003‐0052‐1 

 

 

Cover design: Raadhuis voor creatieve communicatie, Alkmaar 

 

This book is no. II of the TIER Research Series, a PhD thesis series published by TIER. 

 

Page 5: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

  

 

 

 

 

THEY WILL GET THERE! 

Studies on Educational Performance of Immigrant  

Youth in the Netherlands 

 

 

ACADEMISCH PROEFSCHRIFT  

ter verkrijging van de graad van doctor 

aan de Universiteit van Amsterdam 

op gezag van de Rector Magnificus 

prof. dr. D.C. van den Boom 

ten overstaan van een door het college voor promoties 

ingestelde commissie, 

in het openbaar te verdedigen in de Aula der Universiteit 

op woensdag 3 juli 2013, te 13:00 uur 

 

door 

 

Elisabeth Adriana Aleida Maria van Welie geboren te Appeltern 

 

Page 6: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

  

Promotiecommissie  

Promotor: 

Prof. dr. J. Hartog 

 

Overige Leden: 

Prof. dr. J. Dronkers 

Prof. dr. H. Maassen van den Brink   

Prof. dr. J.G. Mora y Ruiz 

Prof. dr. H.G. van de Werfhorst 

Prof. dr. A.M.L. van Wieringen 

 

Faculteit der Economie en Bedrijfswetenschappen 

Page 7: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

  

 

In memory of JanKarel Gevers 

 

Especially in our country, where the sometimes comical, but often also foolish race to 

the top seems to prevail, de‐specialisation, imagination, overview and synthesis 

might not be such a bad antidote (1996). 

Page 8: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit
Page 9: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

vii 

Acknowledgements 

A sense of the whole 

When  I  started  the endeavour  to write a  thesis, my only  certainty was my 

ambition  to  add  to  a  better  understanding  of  the  school  success  of  pupils  of 

immigrant descent.  I  am  forever  grateful  to  Joop Hartog, who  guided me  in  such 

mysterious ways that only now, in hindsight, I begin to understand how his intuitive, 

thoughtful and wise guidance made me find my own way  in this  journey. Although 

we  appreciate  system  dynamics  and  nonlinear  processes  differently–  as  a  highly 

respected Economist, Joop feels  less  inclined to this way of thinking than I do as an 

Ecologist– I  intend to honour him by saying that he demonstrated system dynamics 

at its best: a strong image of the complexity of the entire process of writing a thesis, 

without being distracted too much by the intricacies of every step I had to take. 

 

Biological time flows unevenly 

Motherhood  is  the  fundamental  fulfilment of my  life  and  largely  surpasses 

any  other  endeavour  I  undertook;  getting  children  is  an  all‐overwhelming  instant 

                                                        All texts in Italics except the last one for my parents are from Gunderson L.H. and Holling C.S. (2002). Panarchy: understanding transformations in human and natural systems. Washington, DC: Island Press. 

Page 10: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

viii 

change of purpose.  My talented and beautiful son and daughter, Felix and Bérénice, 

taught me what transcendental love is, which only has the well‐being and happiness 

of the other  in mind. Through them I could more profoundly direct my work to the 

aim of enhanced equity in educational opportunities for all young people. 

 

Flip into another state 

Both at the scale of creating the research  institute TIER, and similarly  in her 

creativity in finding a way around all sorts of bureaucratic impediments that seemed 

to block my way  to  the project of writing a  thesis,  I admire and  respect Henriette 

Maassen van den Brink  for her perseverance,  faithfulness and  friendship. Noblesse 

comes to mind in her presence. 

 

Innovation undermines  bureaucratic  rule‐making  and  stability,  but  induces  greater 

resilience 

In my last conversation with JanKarel Gevers, before he so sadly passed away, 

he once more demonstrated  in his beautiful eloquent style how deeply he believed 

in academe, and how any bureaucratic impediments simply had to be moved out of 

the way of education and research. Those who have known him will remember that 

typical  inquisitive  look he could have, with his head a bit tilted. He  looked at me  in 

that way when he said: "I do not know exactly how this came about, but great men, 

with  insight and wisdom, always seem to watch over you and guide you; you need 

never  to worry."  Pepe  Gines Mora,  Roel  in't  Veld,  Fons  van Wieringen  and  Jaap 

Dronkers were a source of  inspiration  for the most part of my career  in education. 

Koos  van  der  Steenhoven  always  believed  in  my  endeavour  to  add  to  equal 

opportunities  for  all  pupils;  he  created  the  solid  ground  I  could  stand  on  while 

working on this thesis. He pushed research‐based policy making at the government 

level over  the  tipping point; however complicated such changes may be,  there will 

only be ever more research basis for policies, there is no way back. 

 

Page 11: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

ix 

The shift from one set of scales to another 

Every  once  in  a  while  I  stumbled,  and  doubted  whether  I  could  ever 

successfully complete my research; at all those moments Hessel Oosterbeek not only 

prevented me from falling, but also gave me at first sight casual suggestions, to try 

another approach, read an article that just seemed to come to his mind, or "use less 

words and think deeper". All his advice turned out to be exactly right, at exactly the 

right time.  

 

Understanding slow variables 

Countless researchers, students and politicians are indebted to Hank Levin at 

Columbia University, I among them. Hank dedicated a lifetime of work explaining to 

the world that the return on investment of education is always positive– and should 

be calculated over the  lifetime of  individuals, not by the mere costs of one year of 

schooling per individual.  

Growing numbers of children of  immigrant descent have entered the school 

system over the past decades and there have been times that I was deeply worried 

about the tone of the public debate about migrant pupils. I have always believed in 

the University and its academic members as good sui generis. At this point I want to 

acknowledge  the work of Herman van de Werfhorst and Maurice Crul, who  in  the 

best  tradition  of  academe  described  the  steady  progress  of  pupils  of  migrant 

descent, and herewith opposed  the negative debate on migrant pupils with  sound 

research, tirelessly. They strengthened my hope that indeed they will get there. 

 

Adaptiveness is at the heart of understanding 

No text survives sloppy editing; each time Juli Behrendt sent back my writing, 

her editing skills surpassed my expectations. Apart from her expertise  in matters of 

style,  she  demonstrated  time  and  again  her  profound  understanding  of my work. 

Under her watch  I  learned to design my papers more clearly, and express myself  in 

better English. Most of all I valued that I could count on her at all times, as if no time 

difference existed between Auckland, in New Zealand, and Amsterdam. 

Page 12: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

Adaptation to faster levels 

I  could  never  have  succeeded  in  translating my  research  questions  on  the 

educational  success  of migrant  pupils  into  sound  analyses, without  Ilja  Cornelisz, 

Bregje  Zwaan,  and  Adam  Booij.  With  creativity  and  great  mastery  of  statistical 

analysis they supported me in getting out of the data what indeed lay hidden inside 

them. My  visits  to  Ilja  at  Columbia  University were  very  inspiring  and  great  fun, 

especially our discussions about our  joint research paper  in the  library of Columbia 

University– silent and solemn realms  in other places‐ where a sign at the entrance 

says that debate is explicitly encouraged. As if we needed that encouragement. 

 

Adding another dimension 

  I  devote  an  important  part  of my  life  to music,  literature,  and  poetry;  on 

many  bleak  afternoons  over  the  few  past  years,  however, when  I  got  lost  in  the 

complexities of understanding my own data properly,  it seemed as  if all the Muses 

had  left me.  In  hours  like  these  there was  always  Sebastiene  Postma;  she  could 

inspire me  instantly again with her vast knowledge of  the beauty of  language and 

literature.  

 

Transfer and storage of experience 

I  value  the  experience  that  I  had  to  struggle,  like many  pupils  of migrant 

descent, with bridging two cultures: the world of school practice and of academe. At 

times I felt as  if I was unable to express myself clearly  in either world anymore. Lex 

Borghans  and  Inge  de  Wolf  taught  me  how  to  build  that  bridge;  not  by 

compromising, but with identifying the best of both worlds. 

 

Avete atque valete ("Ave atque vale", "Hail and Farewell", Catullus 84‐54BC) 

My  father  and  mother–  who  both  passed  away  at  the  beginning  of  this 

millennium–  and  I,  had  no  easy  life  together.  Although  I was well  aware  of  the 

tragedies that struck them when they were young children, this did not help me to 

Page 13: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

xi 

have a more understanding view,  in hindsight, on sadness and  loneliness during my 

childhood. 

Maybe writing  a  thesis  forces  you  to  deeper  understanding–  it  seems  the 

most challenging episode in my working life– but for the first time I could think of my 

parents with profound kindness and gratitude: these are my mother's genes, which 

planted the idea that there  is no realm where I would not be entitled to enter, and 

my  father's  genes, which enabled me  to  complete  this  grand  intellectual odyssey.

Page 14: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit
Page 15: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

xiii 

Table of Contents 

1  Introduction .................................................................................................... 1 

1.1  A brief history of educating pupils with an immigrant background  1 

1.1.1  "Black" and "white" schools  3 

1.1.2  Was the integration in Dutch society of youth with a migrant background failing?  4 

1.1.3  Segregation  5 

1.2  Building upon success  6 

1.2.1  A focus on local contexts and distinct groups  8 

1.3  Outline of the thesis  8 

1.3.1  Segregation at zip‐code level  8 

1.3.2  Three socio‐economically challenged zip‐code areas  9 

1.3.3  Distance measures offer new insights into patterns of secondary school choice  9 

1.3.4  Matching relevant research findings with concrete school questions  10 

1.3.5  Scientific and societal contributions  11 

2  The educational gap and neighbourhood composition .................................. 12 

2.1  Introduction  12 

2.2  Previous research  17 

2.3  Methods and data  22 

Page 16: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

xiv 

2.3.1  Selecting the data  22 

2.3.2  Assembling the data set  24 

2.3.3  Data from public files  25 

2.4  Results  26 

2.4.1  Enrolments in academic tracks among Moroccan Dutch, at zip‐code level  26 

2.4.2  Enrolments in academic tracks in zips at risk  29 

2.4.3  The correlation between the total percentage of migrants at zip level and 

enrolments of Moroccan Dutch  31 

2.5  Conclusions and discussion  32 

2.6  Appendices  35 

3  School switching, a concern for migrant pupils .............................................. 41 

3.1  Introduction  41 

3.1.1  Young people with a migrant background in Amsterdam‐West  42 

3.2  Previous Research  45 

3.3  Data and Methods  49 

3.3.1  Single schools and clusters  49 

3.3.2  Migrant students compared to native Dutch peers  49 

3.3.3  Three data sets  50 

3.4  Results  53 

3.4.1  The frequency of switching in the three zip‐code areas  53 

3.4.2  The retention rate of schools  56 

3.4.3  Switching and upward and downward mobility between track levels  57 

3.5  Conclusions and policy implications  61 

3.5.1  The three zip areas  62 

3.5.2  A hidden problem?  63 

3.5.3  Youth in the three zips might benefit from agreements between schools to reduce 

switching  63 

3.5.4  Schools offering all tracks may be favourable for migrant students  64 

3.5.5  Reducing switching may decrease costs  64 

3.6  Appendix  66 

Page 17: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

xv 

4  Patterns in secondary school selection in the context of unlimited choice .... 67 

4.1  Introduction  67 

4.2  Previous research  72 

4.3  Data and Methods  80 

4.3.1  BRON data and additional data sources  80 

4.3.2  Ethnic diverse populations in the four major cities  80 

4.3.3  Distance measurements  82 

4.3.4  SES indices  83 

4.3.5  Indicators of pupils' prior achievement and secondary school quality  83 

4.3.6  Limitation  84 

4.4  Results  84 

4.4.1  A general overview of socio‐economic measurements and distance to school  85 

4.4.2  Distance to school  88 

4.4.3  Choosing the nearest school or not  91 

4.4.4  Selectivity of choice  94 

4.4.5  School choice and segregation  97 

4.4.6  Benefits of school choice?  101 

4.5  Conclusions and discussion  104 

4.5.1  Who travels further to school?  104 

4.5.2  Differential sorting, mobility increases segregation  104 

4.5.3  Upward mobility, an important opportunity for migrant students  105 

4.5.4  Policy implications and further research  105 

4.6  Appendices  107 

4.6.1  Explanation of variables  108 

5  A tailor made transfer of scientific knowledge to school practice ................. 119 

5.1  Introduction  119 

5.2  Previous research  121 

5.3  Methodology  126 

5.3.1  Six participating school leaders  127 

5.3.2  Five separate phases of investigation  129 

   

Page 18: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

xvi 

5.4  Portraits of the participating schools and their main questions  132 

5.4.1  School A: An enriched idea of the school's added value for students with a migrant 

family history  133 

5.4.2  School B: the advantage of a professional review of academic literature  135 

5.4.3  School C: The important next step, translating research outcomes to teacher 

practices  137 

5.4.4  School D: Taking time to focus on the complex questions we encounter  138 

5.4.5  School E: The pleasure of taking time for reflection  140 

5.4.6  School F: Time is the most scarce ingredient  141 

5.5  Findings  142 

5.6  Conclusions and Discussion  147 

5.6.1  Principals have questions that can be linked to research outcomes  147 

5.6.2  Comparable questions, different approaches  148 

5.6.3  Actual applications of academic knowledge and impediments  149 

5.6.4  Contributing to the use of existing research in school practice  149 

5.6.5  Contributing to research into the valorisation of scientific knowledge  151 

5.6.6  Limitations and further research  152 

5.7  Appendices  153 

6  Conclusions and Discussion .......................................................................... 161 

6.1  Conclusions  161 

6.2  Policy Implications  166 

6.2.1  A different perspective on segregation  166 

6.2.2  Stable school careers  167 

6.2.3  Validation of academic knowledge  168 

6.2.4  The crucial importance of data  168 

6.3  Limitations and further research  168 

References ........................................................................................................... 171 

Samenvatting ....................................................................................................... 183 

Biography ............................................................................................................ 191 

Page 19: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

xvii 

List of Tables 

Table  2‐1:  Total  average  of  16  year‐olds  per  zip  in  the  four major  cities,  and 

average  number  and  percentage  of  16  year‐old Moroccan  Dutch  among  this 

group  27 

Table  2‐2:  Regression  coefficients  of  the  percentage  enrolments  in  academic 

tracks among 16 year‐old Moroccans on the share of Moroccan 16 year olds at 

zip‐code level  30 

Table 2‐3: Correlation of the percentage of 16 year‐old Moroccans per zip, and 

the percentage among  them  in  secondary academic  tracks,  for all  zips and  for 

zips at risk separately  31 

Table 2‐4: Correlations between total percentage of all  immigrants per zip‐ and 

the  percentage  of  16  year‐old  Moroccans  in  secondary  academic  tracks, 

separately for the four major cities.  32 

Table 3‐1: Composition by ethnicity of all 16‐22 year‐olds living in zip‐code areas 

1061, 1062 and 1063 in the district Amsterdam‐West, at the reference date of 31 

July 2009  53 

Page 20: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

xviii 

Table  3‐2:  The  share  of  16‐22  year‐olds  by  ethnicity,  enrolled  in  academic 

secondary  tracks.  Individuals were  registered  at  least  once  in  the  data  file  as 

being enrolled at this level  54 

Table  3‐3:  The  number  of  secondary  schools  attended  by  16‐22  year‐olds,  by 

ethnicity  55 

Table  3‐4:  The  number  of  secondary  schools  attended  by  16‐22  year‐olds,  by 

ethnicity, measuring school switching that includes cluster switching  55 

Table  3‐5:  Categories  of  school‐exit  among  pupils  of Moroccan  descent, who 

were enrolled in academic secondary tracks (reference date 31 July 2009)  56 

Table 3‐6: School switching among pupils in Amsterdam enrolled in the Year 3 of 

secondary education in 2011  58 

Table  3‐7:  School  switching  among  pupils  living  in  zip‐codes  1061,  1062  and 

1063, enrolled in the Year 3 of secondary education in 2011  58 

Table 3‐8: Cluster switching among pupils  in Amsterdam enrolled  in   Year 3 of 

secondary education in 2011  59 

Table  3‐9:  Cluster  switching  among  pupils  living  in  zip‐codes  1061,  1062  and 

1063 enrolled in Yea 3r of secondary education in 2011  59 

Table 3‐10: Comparison of track level at entry and in Year 3 of secondary school, 

separate for school switchers and non‐switchers, among pupils in Amsterdam  60 

Table 3‐11: Comparison of  the  track  level at entry‐ and  in Year 3 of secondary 

school,  separate  for  school  switchers  and  non‐switchers,  among  pupils  in  zip‐

code areas 1061, 1062 and 1063  60 

Table 3‐12: Comparison of track level at entry‐ and in Year 3 of secondary school, 

separately  for  school  switchers  and non‐switchers,  among pupils of Moroccan 

descent in Amsterdam.  61 

Table 3‐13: Comparison of track level at entry‐ and in Year 3 of secondary school, 

separately  for  school  switchers  and non‐switchers,  among pupils of Moroccan 

descent in zip‐code areas 1061, 1062 and 1063  61 

Table 4‐1: Summary statistics  for pupils  living  in the  four major Dutch cities, at 

the individual, neighbourhood and school level  86 

Page 21: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

xix 

Table 4‐2: Regression  results  (OLS)  for distance  to  secondary  school  in  km, on 

individual‐,  school‐  and  neighbourhood  variables  (standard  errors  in 

parentheses).  89 

Table 4‐3: Summary statistics of the comparison between the group of students 

who  choose  the  nearest  relevant  school  (i.e.  a  school  that  indeed  offers  a 

desired  track)  to  their  home,  and  those  who  do  not  (standard  deviations  in 

parentheses)  92 

Table  4‐4:  Regression  results  (OLS)  for  choosing  the  nearest  school  (i.e. 

comparing pupils who choose the nearest school with those who choose another 

school), with choosing the nearest school = 1  93 

Table 4‐5: Regression results (OLS) for the distance difference (i.e. between the 

actual  and  the  nearest  school),  with  distance  difference  =  0  for  pupils  who 

choose the nearest school  95 

Table 4‐6: Comparisons of nearest  (non‐chosen)  school  and  the  actual  school, 

considering non‐nearest school choosers only  (14247  individuals, 88.7 % of the 

data set)  96 

Table 4‐7: Regression results (OLS) for SES difference (i.e. average SES of chosen 

school minus  average  SES  of  nearest  school),  considering  non‐nearest  school 

choosers only (14197 individuals)  99 

Table 4‐8: Regression results (OLS) for the difference in the percentage of pupils 

with  a  non‐Western migrant  background  (i.e.  percentage  at  the  actual  school 

minus  the  percentage  at  the  nearest  school),  considering  non‐nearest  school 

choosers only (14197 individuals)  100 

Table 4‐9: OLS regressions of individual upward mobility at the entrance of Year 

3, considering only pupils who started at the lower vocational secondary level in 

Year 1 (4343 individuals); non‐nearest school choosers only  103 

Table  5‐1:  All  school  questions  clustered  in  five  domains,  plus matches  with 

selected scientific publications and the nature of every publication.  145 

Page 22: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

xx 

Table 5‐2: Categories of selected research publications, per cluster of questions. 

(Note  that  the  same  publication may  be  suggested  to  several  schools;  e.g.  5 

meta‐review studies were matched 15 times with a school question).  145 

Table 5‐3: Overview of  the outcomes of  the evaluative  interview. Schools A, B, 

and E had an extra intermediate exploratory discussion.  146 

Table  5‐4:  Application  (after  Johnson,  1998)  of  research  outcomes  by  school 

leaders after three months.  147 

 

Page 23: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

xxi 

List of Figures 

Figure 2‐1: Data  scatter,  linear  regression  (OLS) and  LOWESS  regression of  the 

percentage of 16 year‐old Moroccans enrolled in academic secondary tracks per 

zip on  the percentage of Moroccan Dutch 16 year‐olds as a  share of  the  total 

number of 16 year‐olds at zip level in the four major cities in the Netherlands  27 

Figure 2‐2: Data scatter and linear regression (OLS) of the percentage of 16 year‐

old Moroccans enrolled in secondary academic tracks on the share of Moroccan 

Dutch  16  year‐olds  at  zip  level,  separately  for  the  four  major  cities  in  the 

Netherlands  28 

Figure 2‐3: Data  scatter,  linear  regression  (OLS) and  LOWESS  regression of  the 

percentage of 16 year‐old Moroccans enrolled in academic secondary tracks per 

zip at risk on the share of Moroccan Dutch 16 year‐olds at zip  level  in the  four 

major cities in the Netherlands.  30 

Figure  3‐1:  The  retention  rate  in  ranking  order  of  all  secondary  schools 

(considering  separate  schools,  not  clusters)  in  Amsterdam,  defined  as  the 

percentage of pupils who entered in Year 1, and are still enrolled in Year 3  57 

Page 24: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

xxii 

Figure 4‐1: Linearly fitted lines (OLS), one for each of the 5 ethnic groups on the 

correlation  between  absolute  distance  to  secondary  school  and  relative 

neighbourhood SES  91 

 

Page 25: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

xxiii 

List of Appendices 

Appendix 2‐1: Highest completed level of education, by age and ethnicity (2007)  36 

Appendix 2‐2: Percentage of pupils with initial qualification for the labour market 

(ISCED 3 and higher), by ethnicity and age cohort (2007)  37 

Appendix 2‐3: Percentage of pupils with primary education as highest completed 

level, by ethnicity and age cohort (2007)  37 

Appendix 2‐4: Enrolment of 16 year‐olds by ethnicity in senior general education 

(HAVO, ISCED 3) and pre‐university education (VWO, ISCED 3)  38 

Appendix  2‐5:  Percentage  of  pupils  enrolled  in  academic  track  in  Year  3  of 

secondary school (school year 2011/2012), mean neighbourhood SES (scale ‐4 to 

+4), and percentage of pupils eligible  for weighted  student  funding when  they 

were enrolled in elementary education.  39 

Appendix  2‐6:  Data  scatter,  linear  (OLS)  and  LOWESS  regression  of  the 

percentage of 16 year‐old native Dutch pupils on the percentage of 16 year‐olds 

of Moroccan descent at zip level in the four major cities in the Netherlands  40 

Appendix 3‐1: Comparison of the two main data sets.  66 

Appendix 4‐1: Maps of Amsterdam, Rotterdam, Utrecht and The Hague.  110 

Page 26: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

xxiv 

Appendix  4‐2:  LOWESS  regression  (Locally  Weighted  Scatterplot  Smoothing; 

bandwidth = .6) and linear regression (OLS) of the extra distance travelled to the 

preferred school (as compared with the nearest school) and school's mean final 

exam score (national standardized exam score)  115 

Appendix 4‐3: Regression results (OLS) for difference  in the mean relative exam 

score at school level (i.e. the actual school exam score minus exam score at the 

nearest school), considering non‐nearest school choosers only  116 

Appendix  4‐4:  Regression  results  (OLS)  for  difference  in  school  percentage  of 

upward mobility  to  a  higher  track  (i.e.  upward mobility  at  the  actual  school 

minus upward mobility  at  the nearest  school),  considering non‐nearest  school 

choosers only  117 

Appendix 5‐1: General characteristics of the six participating schools.  153 

Appendix 5‐2: Format of the first interview  153 

Appendix 5‐3: Questions for the evaluating interview  155 

Appendix 5‐4: Questions and matches per school  156 

 

Page 27: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

1 Introduction 

A person's true joy and felicity lie solely in his wisdom and 

knowledge of truth, not in being wiser than others or in others' 

being without knowledge of truth, since this does not increase his 

own wisdom which is his true felicity. 

Baruch de Spinoza, 1670 

1.1 A brief history of educating pupils with an immigrant background 

In the last decades of the past century, secondary schools, especially in the larger 

Dutch  cities,  experienced  substantial  changes  in  the  composition  of  their  student 

population.  These  changes were  induced  by  the  arrival,  from  the  1960s  onwards,  of 

large  numbers  of  labour  migrants,  who  were  invited  to  come  and  work  in  the 

Netherlands.  The  majority  of  those  labour  migrants  originated  from  Turkey  and 

Morocco, and were predominantly employed  in  low‐skilled work. In the years to come, 

they were, in most cases, followed by their wives and children. 

In their countries of origin, many Turkish and Moroccan people did not have the 

chance  to  go  to  good  quality  schools,  or  pursue  continued  education;  moreover, 

substantial numbers of them had lived too far from an elementary school to be able to 

go to school at all. Their children, among them increasing numbers who were born in the 

Page 28: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

Netherlands, both literally and figuratively had to come a long way to find their rightful 

place in the different strata of the Dutch educational system. 

In 1987,  I was appointed as Principal of a secondary school  in Amsterdam‐West 

that was in the midst of such significant changes in its student population. Many people 

from Morocco and Turkey came  to  live  in  this part of  the city– and still do– and  their 

children were arriving in secondary school in large numbers. Our students were not only 

very diverse with respect to the country of origin of their parents, but also in their level 

of  schooling:  some  were  born  in  the  Netherlands,  and  had  acquired  a  sufficient 

proficiency  in  the Dutch  language  in elementary school, but others still  struggled with 

language–  albeit  they  were  born  in  the  Netherlands–  for  example  because  another 

language was spoken at home. However, a substantial number of pupils at that time had 

been  born  in  their  country  of  origin,  and  had  therefore  been  enrolled  in  Dutch 

elementary  schools  at  a  later  age  than  most  of  their  classmates.  Obviously,  their 

language proficiency required specialized attention. In order to serve our students well, 

however, we had to re‐consider many other aspects of daily school practice as well.  

In  retrospect, our attitude was very goal‐oriented and practical: we wanted  to 

guide all our students to the highest possible  level of secondary education, given their 

individual intelligence and skills; and when the student body changes, teaching needs to 

be  reconsidered  accordingly.  In  only  a  few  years'  time,  our  school  population  had 

changed  from  being  a  student  body  almost  exclusively  of  Dutch  descent,  to  a  vast 

majority of pupils with a migrant background. Therefore,  in a relatively short period of 

time, we had to develop new practices to adapt our school to the needs of our changed 

student body. Of course, we developed programmes for second  language  learners, but 

we  also  acknowledged  the  importance of meaningful  close  cooperation with parents, 

not only for explaining to them the complex system of tracked secondary education, and 

its consequences for tertiary schooling and chances on the labour market in the future. 

Importantly,  some  of  our  colleagues were  skilled  interpreters, who  could  explain  the 

school  system  to  parents,  and  discuss  the  results  and  prospects  of  their  sons  and 

daughters  in their mother tongue. We also had to cope with, as another example, the 

difficult  issue of  inventing new practices  and procedures  that  truly  assessed  a pupil's 

abilities  and  that  "looked  through"  the  façade  of  possibly  insufficient  language 

proficiency, which may  have masked  real  potential. We  acknowledged,  furthermore, 

Page 29: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

1 Introduction 

that we had  to  identify other kinds of  stakeholders, and  reach out  to people with an 

influential  position  in migrant  communities.  Together with  them we mapped  out  the 

route  for our pupils  towards  far higher  levels of  schooling  than had been available  to 

their parents in their country of origin. 

Although we certainly had our share of problems with pupils who did not behave 

as we expected them to do, as in any secondary school, we thought about our students 

as  ambitious  young  people, willing  to  learn,  and  eventually  expecting  to  find  a  good 

position on  the  labour market. Teachers were  in general very dedicated; many among 

them would devote their entire career to teaching migrant pupils and remained loyal to 

their  school  until  the  present  day, more  than  two  decades  later,  even  though  such 

experienced  teachers may have  a  lot of  job opportunities  in other  schools  to  choose 

from.  

Interestingly,  American‐based  studies  on  educational  achievements  among 

migrant students  in general report that, almost as something that goes without saying, 

that migrant students express higher educational aspirations for themselves than is the 

case among American‐born students.  In general,  these higher ambitions are explained 

by the assumption that their parents, who took the far‐reaching decision to  leave their 

family  and  home  country,  may  be  a  select  group  who  are  more  often  ambitious, 

entrepreneurial,  and  willing  to  take  substantial  risks.  Parents  may  even  choose  to 

emigrate specifically for the purpose of finding better schools for their children.  

At  the  time,  in  our  school, we would  have  recognized  these  characteristics  of 

migrant pupils: our students performed remarkably well, and many of them were very 

ambitious, and envisioned a future that would surpass all their families' expectations.  

1.1.1 "Black" and "white" schools 

Increasingly, however, our school and comparable other schools, were classified 

as  "black  schools". And although  in public discourse many discussants and  columnists 

started with some sort of statement  that, of course, "black" was  just an  indication  for 

schools with many migrant students, and did not mean anything disparaging, the reality 

was very different. The designation "black school" became almost equivalent with lower 

results,  lower chances  for pupils, more disciplinary problems and more dropouts. This 

dichotomy  split  the  landscape  of  schools;  some  schools were  even more  determined 

Page 30: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

than  before  to  educate migrant  pupils  to  the  best  of  their  potential. Other  schools, 

however,  developed  strategies  to  attract  mainly  native  Dutch  students,  with  more 

affluent parents. As a result, especially  in the four major cities Amsterdam, Rotterdam, 

Utrecht, and The Hague, nowadays schools are to a large extent segregated (Ladd, Fiske 

and Ruijs, 2009).  

Considering  the  ample  scientific  evidence  for  the  strong  correlation  between 

parental  levels  of  schooling  and  the  educational  prospects  of  their  children,  not 

surprisingly, average exam results tend to be higher  in schools with children from well‐

educated parents. The question  remains,  though, whether a higher mean exam  score 

indicates  the added value of a  school  (Opdenakker and Van Damme, 2001), or  rather 

reflects the pupils' resources at home,  including the presence of role models that may 

lead to a sense of entitlement to high levels of schooling. 

1.1.2 Was the integration in Dutch society of youth with a migrant background failing? 

In  2000  Paul  Scheffer  published  an  essay  in  a  national  newspaper  (NRC 

Handelsblad) entitled  "The Multicultural Drama"  that  rapidly became a  cause  célèbre. 

Scheffer stated that unemployment, dropout and criminality accumulated among ethnic 

minorities, and  that he was not hopeful  for  the  future: his expectation was  that  large 

numbers  of  minority  youths  would  permanently  lag  behind,  and  remain  without 

prospects  for  a meaningful  participation  in  Dutch  society.  Importantly,  the  intent  of 

Scheffer's essay was to  increase opportunities  for youth with a migrant background to 

participate  in  Dutch  society,  his  critique  being  that  policy makers,  and  society  as  a 

whole, had for too long turned a blind eye towards the failing integration of youth with a 

migrant  background.  In  his  view,  this  lenient  attitude  was  an  ineffective  kind  of 

tolerance,  and not  in  the  interest of migrant  students. However,  the  following public 

discourse  initiated  a  second  dichotomy:  those  who  agreed  with  Scheffer  were 

considered  to  be  the  new  realists,  while  those  who  pointed  at  the  continuous 

educational  improvement of migrant pupils were, oddly enough  in a pejorative  sense, 

referred to as multiculturalists– people who were considered to be in a state of serious 

denial.  For more  than  a  decade  to  follow,  fundamental  shifts  in  the  Dutch  political 

landscape  resulted  in  an  almost  categorically  negative  view  of  young  people with  a 

migrant background, notably those of Moroccan descent. As a result, the ever growing 

Page 31: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

1 Introduction 

shares of migrant pupils who were successful in higher levels of schooling, and did so at 

an  increasing  tempo,  were  almost  ignored  because  of  the  strong  emphasis  on  the 

assumed educational and societal failure of migrant pupils.  

Nevertheless,  teachers  and  school  leaders  in  secondary  schools  with  large 

percentages of migrant pupils invested with increasing success in adapting their teaching 

to a very diverse student body, and were not at all of the opinion that the education of 

migrant pupils was failing; on the contrary, they saw increasing numbers of enrolments 

in the highest secondary tracks, and  improving exam results. As the Director of a  large 

secondary  school  recently  shared:  "We  stopped  counting ethnicities and nationalities, 

we consider this to be irrelevant; we want to serve, as best as we can, all pupils who are 

living  in  the district around our school. To  that end, we constantly work on  improving 

our  programmes,  the  quality  of  our  assessments,  and  the  final  exam  results  of  our 

pupils."  

While diverse schools largely invested in adequate education for students with a 

multitude  of  backgrounds,  the  societal  discourse  seemed  to  put  desegregation  of 

schools  and  neighbourhoods  centre  stage,  by  implication  suggesting  that  close‐knit 

ethnic communities and neighbourhood schools were the cause of the assumed societal 

failure of young people with a migrant family history. 

1.1.3 Segregation 

Among politicians and policy makers, pessimistic views on black schools and on 

communities  of  people  with  a  migrant  background  seemed  to  be  mainly  based  on 

research in the context of large metropolitan areas in the U.S. (Dobbie and Fryer, 2009; 

Payne, 2010; Kerbow, Azcoitia and Buell, 2003). However, some major characteristics of 

the Dutch educational system, and characteristics of the larger cities (where the majority 

of pupils with a migrant background live) differ substantially from metropolitan areas in 

the  U.S.  First,  in  the  Netherlands,  free  school  choice  is  universal  and  no  financial 

considerations  restrict  free  choice,  since  all  schools  are  funded  equally  by  the 

government, with  additional  funding  for  all  low  SES  (socio‐economic  situation)  pupils 

and no tuition fees. Second, unlike large metropolitan areas in the U.S., the major Dutch 

cities do not have  large  impoverished areas that are at substantial distance from more 

affluent  areas.  In most  cases,  lower  SES  neighbourhoods may  be  adjacent  to more 

Page 32: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

affluent areas. Third, Dutch cities have a high number of schools that are, furthermore, 

all supervised by the Inspectorate of Education. Any seriously underperforming schools 

(currently two schools in the four major cities) come under intensified supervision by the 

Inspectorate, and typically manage to  improve their results within the space of a year. 

Fourth, teacher salaries (based on the  level of training and teaching) are equal; diverse 

schools do in general, not report other problems with the hiring of teachers than those 

related to the shortage of teachers on the  labour market  in some subjects, and do not 

report a higher  turnover of  teachers.  In discussions,  school  leaders of diverse  schools 

express rather the contrary: the majority of their teachers made a deliberate choice to 

teach pupils who,  in most  cases, are  the  first  in  their  family  to  reach  secondary  track 

levels that qualify them for higher education.  

Obviously, attending a segregated school and  living  in a segregated community, 

may  raise  concerns  considering,  for  example,  the  preparation  of migrant  youth  for  a 

successful  future entry  in the  labour market. The question remains, however, whether 

schools  specialized  in migrant  education–  with  an  emphasis  on  upward  mobility  to 

higher  tracks  (Crul, Schneider and Lelie, 2012)– are  the better option, or  schools with 

mixed populations, possibly supported by housing policies.  

Remarkably,  considering  the  effect  of  segregation  and  desegregation  policies, 

academic  publications  contain  surprisingly  contrasting  findings:  1)  large  scale  housing 

policies  aimed  at  desegregating  impoverished  residential  areas  did  not  result  in  any 

improvement  of  educational  attainment  levels  (Oreopoulos,  2007);  2)  Desegregating 

schools can further burden black pupils (Stuart Wells et al., 2009); 3) School investments 

in cooperation with  local segregated communities have beneficial effects on the school 

results  of  black  children  (Dobbie  and  Fryer,  2009);  and  4)  neighbourhoods with  high 

percentages of  inhabitants  from different  countries of origin have negative effects on 

school results (Dronkers, 2010).  

1.2 Building upon success 

The  increasing school success of pupils with a migrant background has  inspired 

this thesis, which  is based on the assumption that preventing or repairing failure  is not 

simply  the  opposite  of  building  upon  success–  if  only  because  failing  students  and 

successful  students  are different  individuals.  Interestingly, Crul et  al.  (2012)  find,  in  a 

Page 33: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

1 Introduction 

large scale  international comparative survey, that  in the Netherlands the school results 

of migrant pupils are remarkably polarized: being of migrant descent, both considerably 

raises the odds of pupils successfully completing far higher levels of schooling than their 

parents did, but, at the same time, raises the odds of dropping out– with about equal 

percentages. Noguera  (2004) describes similarly, that,  in the U.S.,  immigrants are both 

more  likely  to  succeed and more  likely  to  fail academically.  Importantly,  the  fact  that 

these  two  sides  of  the  distribution  represent  different  individuals,  national  average 

enrolments may not offer an adequate insight into patterns of school success in the case 

of migrant pupils. Interestingly, van de Werfhorst and van Tubergen (2007) demonstrate 

that, after controlling for parental education and occupational class, "ethnic differences 

in achievement vanish, and differences in secondary school type almost disappear. What 

remains  (…)  is  not  an  ethnic  penalty,  but  an  ethnic  advantage:  Turks,  Moroccans, 

Surinamese and Antilleans choose higher types of secondary schooling than natives with 

comparable class backgrounds".  

This thesis aims to add to a better understanding of the successful group among 

immigrant pupils for two main reasons: first, when success is better understood, policies 

to  reinforce  success  factors  may  add  to  cost‐effective  government  policies,  since 

repairing  failure  in  fact comes too  late, and may  imply considerable costs  (e.g. guiding 

dropout students back  to school, often  repeatedly). Second, as mentioned earlier,  the 

success  of  growing  shares  among  migrant  students  seems  almost  obscured  by  the 

predominant  attention  to  failure.  As  a  result,  notably Moroccan  students  are  almost 

categorically portrayed as underperforming. 

This negative image may have consequences for the constantly growing group of 

high achievers: they currently experience hindrances to find an internship or a job, and 

unemployment is substantially higher among young people of migrant descent. 

Admittedly, the effort to add to equity and justice for all pupils has been a major 

driving  force behind my  long career  in education; however,  the point of departure  for 

the current research is the rational consideration that migrant youth in the larger cities 

form half of the future labour force. There are, therefore, obvious economic interests in 

educating them to the best of their potential.  

Page 34: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

1.2.1 A focus on local contexts and distinct groups 

In  order  to  study  the  school  success  of  migrant  pupils,  several  preliminary 

choices have been made. First, for the purpose of this thesis, access to the highest levels 

of secondary education, the academic tracks that grant access to higher education, has 

been defined as "success". Second, a  focus was put on a better understanding of  local 

contexts, individual schools, and separate migrant groups (depending on the country of 

origin  of  their  parents).  To  this  end,  1)  the  zip‐code  level  has  been  used  as  a  local 

geographical  unit;  2)  five  distinct  ethnicities  have  been  taken  into  consideration 

separately:  pupils  of  Dutch, Moroccan,  Turkish  and  Antillean/Surinamese  descent,  as 

well as the combined group of "other migrants", that may, however, largely vary in their 

socio‐economic characteristics; 3) a special focus was put on the school success of one 

specific group, pupils with a Moroccan  family history; 4)  three  specific  zip‐code areas 

were studied separately; and 5) six secondary schools with a large population of migrant 

students were investigated in greater detail.  

1.3 Outline of the thesis 

1.3.1 Segregation at zip‐code level 

In Chapter 2, a basic question underlying this thesis is explored: Is school success, 

defined  as  enrolments  in  academic  secondary  tracks,  among  students  of  Moroccan 

descent, associated with the degree of segregation at the zip‐code level of their area of 

residence? Data on the ethnic composition among 16 year‐olds at zip‐code  level  in the 

four major  cities have been merged with  individual data on  enrolments  in  secondary 

academic  tracks  at  that  same  age,  in  order  to  analyse  the  association  between  the 

residential area and school success. 

As  described  above,  contrasting  outcomes  from  academic  research  have  been 

published concerning the effect of segregation at the neighbourhood level: at the same 

time,  existing  research  points  at  negative  effects  of  segregated  communities  and 

schools,  but  also  demonstrates  both  positive  and  negative  effects  of  desegregation 

policies,  and,  furthermore,  reports  positive  effects  of  strong  (segregated)  migrant 

communities and neighbourhood schools.  

Page 35: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

1 Introduction 

1.3.2 Three socio‐economically challenged zip‐code areas 

The  aim  of Chapter  3  is  to  study  in more  detail  the  school  histories  of  young 

people with a migrant background  in  three  residential areas with a  largely segregated 

population, mainly of Moroccan descent,  in  the district of Amsterdam‐West. All  three 

areas  have  had  to  cope  with  serious  socio‐economic  disadvantages,  and  have  been 

depicted as notoriously problematic  in the media over the past decade. This analysis  is 

based on a local data set, covering the school history (all exact exit and entrance dates) 

of all 16‐22 year‐olds at  the  reference date of 31  July 2009. The study starts with  the 

open question: Does the analysis of municipal data on school histories of the youths  in 

this specific area, reveal local aspects of education that probably may foster the success 

of some pupils, but as yet do not mitigate the problematic school careers of others. 

Even at first sight, the database showed a surprisingly high incidence of recurrent 

departure (without a diploma) and re‐enrolment in secondary school among pupils with 

a migrant background. 

The  reviewed  literature  reports disruptive effects of  school  switching, both  for 

positive  and  for  negative  reasons,  and  finds  that  switching  raises  the  odds  of  later 

dropout.  School  switching  appears  to  be  a  relatively  under‐studied  subject  in  the 

Netherlands.  In  order  to  be  able  to  compare  the  frequency  of  switching  among 

Moroccan youth living in the three investigated areas, in this chapter a second analysis is 

carried  out with  the  use  of  up‐to‐date  national  educational  data. With  this  data  set, 

switching rates for Amsterdam as a whole, and differences among ethnic groups could 

be  demonstrated,  measured  in  Year  3  of  secondary  schools.  The  national  data  set 

confirms the findings based on the local data. 

1.3.3 Distance measures offer new insights into patterns of secondary school choice 

Chapter  4  is  built  around  geographical  distance  measurements,  notably  the 

distance  to  the  school  concerned and  the difference  in distance between  the nearest 

and  the more distant preferred school,  in  the search  for underlying patterns of school 

choice between different migrant groups‐ and native Dutch pupils. Additionally, distance 

measurements  allow  for  controlling  for  population  density,  the  number  of  schools 

available within 5 km, and "urbanicity" (a measure for economic activity in a given area). 

Patterns  of  school  choice may  reveal  the  preferences  of  parents  and  their  children, 

Page 36: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

10 

notably when a school other than the nearest relevant school is chosen. The underlying 

idea was that, in terms of travel time and possible travel costs, choosing a more distant 

school requires an extra effort. Can patterns of school choice reveal the motivation for 

this extra effort? We considered the extra effort of selecting a more distant school to be 

a proxy for the selectivity of choice. Importantly, the choice for the nearest school may 

also  be  a  deliberate  choice:  parents  in  affluent  neighbourhoods,  for  example, might 

prefer  the  nearest  school  because  the  average  socio‐economic  level  of  the  school 

population may reflect the affluence of  the residential area.  In  this case, however, we 

cannot know the selectivity of their choice; others may opt for the nearest school to the 

home‐address, without making a deliberate choice. For the purpose of this study, a rich 

government data  set has been used,  containing  all educational data  and many  socio‐

economic data of the cohort that was enrolled in the last grade of elementary education 

in  2008.  The  data  set  has  been merged with  additional  socio‐economic  data  on  the 

neighbourhood  level,  and  qualitative  data  on  school  performance  provided  by  the 

Inspectorate of Education. 

1.3.4 Matching relevant research findings with concrete school questions 

Chapter  5  presents  an  in‐depth  study  of  the  transfer  or  the  lack  thereof–  of 

applicable outcomes from academic research to actual school questions and practice. Six 

Principals  of  diverse  schools  in  Amsterdam  actively  participated  in  this  project.  This 

study has been  inspired by, on  the one hand,  the abundance of high‐quality,  recently 

published studies on migrant education, and, on  the other hand,  the observation  that 

school  leaders experience difficulties  in  selecting a knowledge base  for current  school 

developments and innovative ambitions. Some issues at stake appear to be the timelines 

and accessibility of research publications by school  leaders and teachers, and the often 

contradictory outcomes concerning the same educational theme. 

On  the basis of semi‐structured  in‐depth  individual  interviews,  the main school 

questions were identified. Then, an extensive literature review was carried out to match 

these questions  to up‐to‐date  scientific publications  that offered opportunities  for  an 

effective translation to school practices.  

Page 37: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

1 Introduction 

11 

1.3.5 Scientific and societal contributions  

In Chapter 6, general  conclusions, based on  the  four  studies  in  this paper, are 

presented.  Furthermore,  possible  implications  for  school  practice  and municipal  and 

national government policies are considered. Finally, the ways  in which this thesis may 

contribute to the field of scientific knowledge is reflected upon. 

 

Page 38: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

12 

2 The educational gap and neighbourhood composition 

…no disadvantages stem from such freedom (i.e. liberty of 

judgement) […] I do not need to go far to find instances of this. 

Amsterdam is a fine example of a city which enjoys the fruits of 

this liberty,… 

Baruch de Spinoza, 1670 

2.1 Introduction 

The educational achievement gap between native Dutch students and pupils with 

a migrant background, as measured by enrolments in the academic tracks of secondary 

education,  is substantial:  in 2008 43.9% of native Dutch students1 were enrolled at this 

level–  that  qualifies  for  access  to  higher  education–  against  19%  of  youngsters  of 

Moroccan descent.  

                                                       1 Crul et al. (2012), rightly states that the vast majority of pupils with a migrant background are born in the Netherlands and have Dutch citizenship. As a consequence, Crul et al. (2012) criticize the habit to call students with a Dutch family history "native Dutch" and prefers to refer to this group as the "comparison group". Nonetheless, in this paper they are still referred to as native Dutch, for considerations of readability. 

Page 39: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

13 

In  the  current  chapter,  the  definitions  of  immigrant  background  used  by  the 

National  Statistics  Office  CBS  have  been  used2.  A  main  distinction  has  been  made 

between  native Dutch  and migrant  background, with  a  further  focus  on  a Moroccan 

migrant background. The vast majority of pupils  in the data set that has been used for 

measurements  in the current research are themselves not  immigrants, they were born 

in  the  Netherlands;  migrant  background  refers  to  the  history  of  their  parents  and 

grandparents.  

At  present,  about  half  of  all  12‐year‐olds  in  the  four  largest  cities  in  the 

Netherlands  have  a migrant  background; we  are,  therefore,  looking  at  a majority  of 

minorities (Crul et al. 2012), rather than minorities. The vast majority of students with a 

migrant  background  currently  enrolled  in  secondary  education  were  born  in  the 

Netherlands, but, by definition, have  at  least one parent who was born  abroad;  they 

form what  is  called  the  second generation. While  the achievement gap  indicates  that 

migrant pupils  still as yet  lag behind  in enrolments  in  the highest  tracks  in  secondary 

education,  their  constantly  increasing  enrolment  levels  over  the  past  decades 

demonstrate,  however,  considerable  educational  success  and  intergenerational 

improvement.  The  parents  and  grandparents  of  pupils, who  are  currently  enrolled  in 

secondary  schools,  have  on  average  lower  levels  of  education.  Especially  among  the 

older  generation,  illiteracy may occur  relatively  frequently, because  in  their  region of 

origin, schools were not always available, and elementary education may not have been 

compulsory  in  their  early  youth. Considering  the  strong  correlation between parental 

education and the  level of schooling of their children, this makes the  intergenerational 

educational improvement of the current generation of young people of migrant descent 

even more remarkable.  

Nevertheless,  public  and  political  discourse  over  the  past  decade  has  been 

almost categorically negative about people of all ages with a migrant background and on 

the educational achievements of their children, notably those of Moroccan descent. This 

                                                       2 First generation immigrant: someone born abroad with at least one parent born abroad. Second generation immigrant: someone born in the Netherlands with at least one parent born abroad. Foreign background: someone with at least one parent born abroad. Western immigrant: someone originating from a country in Europe (exclusive of Turkey), North America, Oceania, Indonesia or Japan. Non‐western immigrant: someone originating from Africa, South America, Asia (exclusive of Indonesia and Japan) or Turkey. 

Page 40: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

14 

negative  discourse  seems  to  have  been  inspired  by  notions  of  a multicultural  drama 

(Scheffer, 2000). In political debate, moreover, the supposed failure of the multicultural 

society  has  been  a  recurrent  topic.  The  discourse  seemed  to  be  dominated  by  this 

supposed  educational  and  societal  failure  of  former  immigrants  and  their  children. 

However,  publicly  accessible  national  statistics  on  education  have  recorded  ever 

increasing  attainment  levels  among  pupils  with  a  migrant  background;  these  easily 

available statistics did not, however, apparently influence the negative image of migrant 

groups in society. 

This negative debate focused especially on youth with a Moroccan family history: 

in 2008, 620 articles on Moroccan youth appeared in Dutch national newspapers, more 

than half had a negative tone. In many articles a strongly negative picture was presented 

of those neighbourhoods where many people of Moroccan descent live. However, while 

the  population  of  people  with  a  Turkish  background  is  roughly  equal  in  size,  by 

comparison, only 41 articles on Turkish youth were published in the same period, among 

which  these were about 20 with a negative  tendency  (source: Ministry of Education). 

The  ever‐increasing  average  levels  of  schooling  of  the  majority  of  Dutch  Moroccan 

pupils, referred to as active agents of change in cities by Crul et al. (2012), were virtually 

unreported  in public discourse. The desegregation of  schools  and neighbourhoods  (in 

public discourse called black schools and black neighbourhoods‐ "black" referring not to 

race,  but  to  a  migrant  background)  became  an  important  policy  goal,  since,  the 

residential concentration of ethnic groups was considered to be a major hindrance  for 

full participation in Dutch society. 

Some  characteristics  of  the  Dutch  system  for  secondary  education may  be  in 

favour of pupils with a low SES migrant background, who may be the first in their family 

to be educated at academic levels: 1) in the Netherlands, all schools are funded equally 

by the government; schools are, therefore, independent of local taxes for their funding; 

2)  there  are  no  tuition  fees;  3)  children  from  all  low  SES  parents  (including migrant 

parents) receive extra funding through a voucher system  in elementary education, and 

extra funding based on overall school SES composition in secondary education; 4) school 

choice has been  free  for almost a century; no  financial or other barriers  limit parents 

and children in their choice.  

Page 41: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

15 

Furthermore, the distances between poorer and more affluent neighbourhoods 

are  relatively  small  in  Dutch  cities;  poorer  areas  are  not  geographically  distant  or 

isolated  from  high  SES  areas.  Residential  areas  where many  people  with  a migrant 

background  live  are,  therefore,  never  very  geographically  distant  from more  affluent 

areas. 

The  current  chapter  sets  out  to  present  and  describe  actual  educational 

enrolment  levels  of  16  year‐old  youth with  a migrant  background  (reference  date  1 

October  2008).  In  order  to  better  understand  the  effect  of  stratification  at  the 

neighbourhood  level,  the association between  the ethnic  composition at  the  zip‐code 

level and enrolment in academic tracks has been considered, with a focus on enrolments 

of youths with a Moroccan background. Several choices have been made  to delineate 

the research target:  

First, the analyses were restricted to pupils in secondary schools, who were living 

in  the  four  major  cities  Amsterdam,  Rotterdam,  Utrecht  and  The  Hague,  since  the 

majority of migrant pupils live in these four cities.  

Second, because  this  research  considers educational  success  in particular  (and 

not  the  full  range  of  educational  outcomes),  only  pupils  enrolled  in  the  two  highest 

tracks  of  secondary  education  have  been  taken  into  account.  The  Dutch  system  of 

secondary  education  distinguishes  three  main  track  levels:  pre‐vocational  education 

(further  subdivided  in  four  distinct  levels),  senior  general  education  (HAVO,  Dutch 

acronym)  and  pre‐university  education  (VWO,  idem).  In  the  context  of  this  research, 

"success" has been defined as access to the two highest  levels of secondary education: 

senior general education and pre‐university education,  in combination here referred to 

as academic tracks. Both academic tracks grant access to higher education.3 The choice 

                                                       3 The following translations of the Dutch educational system in the ISCED (International Standard Classification of Education by UNESCO, update 1997) classification are used: 

VMBO: pre‐vocational secondary education, 4 years, ISCED 2, qualifying for senior secondary vocational education. 

HAVO: senior general education, 5 years, grade 1‐3 ISCED 2, grade 4‐5 ISCED 3, qualifying for higher education. 

VWO: pre‐university education, 6 years, grade 1‐3 ISCED 2, grade 4‐6 ISCED 3, qualifying for higher education. 

MBO: senior secondary vocational education, level 1 ISCED 2, level 2‐4 ISCED 3, level 4 qualifying for higher education.  

HBO: universities for applied sciences, ISCED 5B. 

WO: research universities, ISCED 5A. 

Page 42: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

16 

to  focus  on  the  association  between  neighbourhood  composition  and  access  to 

academic  tracks  has  been  motivated,  furthermore,  by  the  2010  report  of  The 

Netherlands  Institute for Social Research that states, contrary to assumptions  in public 

and political discourse: "Dutch  research does not  indicate a clear association between 

the  ethnic  composition  of  a  neighbourhood  and  the  socio‐economic  position  of 

individuals. When considering chances of unemployment, dependency on social welfare 

and poverty,  individual characteristics, notably educational attainment, seem  far more 

important" (translated from Dutch by the author). Granted the importance of individual 

educational  attainment,  neighbourhood  composition may,  however,  have  an  indirect 

effect on education. 

Third,  zip‐code  level has been used  to define  the  residential area of  individual 

pupils; and fourth, analyses are based on all 16 year‐olds living in the four major cities at 

the reference date of 1 October 2008. Age has been considered to be a more accurate 

selection than school grade, since pupils in the same grade may vary considerably in age, 

especially  in  the  case  of migrant  pupils, who may more  often  follow  a  longer  route 

through successive secondary tracks.  

This resulted in the following research questions 

1. How is ethnic composition at zip‐code level, associated with the access among 16 

year‐old  Moroccan  Dutch  pupils  to  academic  tracks  of  secondary  education 

(HAVO  and  VWO),  in  the  four  major  cities  in  the  Netherlands  (Amsterdam, 

Rotterdam, Utrecht, and The Hague)? 

2. Considering only zips with severe socio‐economic challenges (zips at risk), how is 

living in these areas associated with access to academic secondary tracks? 

3. Considering the total share of pupils with a migrant background in zip areas– as 

opposed to measuring the share of Moroccan pupils specifically in the case of the 

first  two  research  question–  how  is  the  total  ethnic  composition  of  zip  areas 

related to access to secondary academic levels?  

 

                                                                                                                                                                  The completion of ISCED level 3 is the internationally agreed initial (or basic) qualification for the labour market. School leavers without an ISCED 3 qualification are regarded as early school leavers or drop outs. ISCED 4 is not used in this text. 

Page 43: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

17 

In Section 2.2 previous research is reviewed; methods and data are described in 

Section 2.3, the results in Section 2.4, and the conclusions and discussion are presented 

in  Section  2.5.  The  results  point  to  a  negative  association  between  neighbourhood 

ethnic composition and enrolment  in academic secondary tracks. The negative slope of 

the  regression  seems,  however,  to  be mainly  the  result  of  the  presence  of  a  limited 

number of zip‐code areas which are burdened by persistent and severe socio‐economic 

challenges.  In  most  zip‐code  areas  enrolments  are  at  the  level  of  national  average 

enrolments  in  academic  secondary  levels,  and  in  many  zip  areas  enrolments  are 

considerably higher. 

2.2 Previous research  

A stubborn problem with severe consequences 

The achievement gap between students with a migrant background and native 

students presents a persistent major challenge to education (Organization for Economic 

Cooperation and Development, OECD, 2010). "Migrant students" is a general term used 

in academic  literature, and may  include different countries of origin, different motives 

for  emigration  and  different  countries  of  destination.  In  the  Netherlands,  the 

educational achievement gap between native Dutch pupils and youths whose parents 

immigrated  from  Turkey  and Morocco  is  an  important  concern.  The  vast majority  of 

these students from Turkish or Moroccan background belong to the second generation: 

they were themselves born in the Netherlands, but at least one parent was born in the 

country of origin. Obviously, the educational achievement gap may have serious further 

consequences, both for the individual and for society at large, as is extensively described 

in the  literature  (e.g. Banks, 2007; Groot and Maassen van den Brink, 2003; Heckman, 

2008;  Belfield  and  Levin,  2007).  The  Country  Report  for  the  Netherlands within  the 

framework  of  the  international  OECD  Reviews  of  Migrant  Education  (OECD,  2010), 

furthermore, states explicitly  (among other  recommendations)  that efforts  to  increase 

enrolments  in high  levels of education among migrant  students  should be  intensified, 

and  career  support,  notably  for  students  in  vocational  programmes,  should  be 

prioritized. 

Assumedly owing to the complex nature of the gap, finding causal evidence for 

effective  strategies on  the  system  level, appears  to be a  tough endeavour:  the  causal 

Page 44: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

18 

effect  of,  for  example,  increasing  per  capita  funding  of  pupils,  reducing  class  size, 

extending  school  hours,  compulsory  summer  schools  or merit  pay  for  teachers  and 

Principals,  cannot  be  proven  (McKinsey &  Company,  2007;  Dobbie  and  Fryer,  2009). 

Although  the current paper does not provide causal explanations  for  the achievement 

gap  either,  this  study  aims  to  contribute  to  clarifying  the  position  of  high‐achieving 

migrant pupils, since the majority of studies in the Netherlands, mainly seem to consider 

those migrant  pupils who  lag  behind.  In  addition,  surprisingly,  Heineck  and  Riphahn 

(2007),  in  their  study  on  intergenerational  transmission  of  educational  attainment  of 

birth cohorts in Germany from 1929 through 1978, find that: "In spite of massive public 

policy  interventions and education reforms our results yield no significant reduction  in 

the role of parental background for child outcomes over the past decades".  

However,  low  educational  attainment  levels  of migrant  parents may  possibly 

have a different cause when compared  to native‐born  low SES parents: many migrant 

parents, notably the older generation of immigrants, often did not have a chance to go 

to school at all, because there was no school at a reasonable distances  in their country 

of origin. Moreover, elementary education may not have been compulsory at the time. 

Taking  this  thought  further,  low attainment  levels of migrant parents may  to a  lesser 

degree point to  lower capacities for  learning to a  lesser degree than could be the case 

for  low  SES  parents who  actually  did  have  access  to  schooling  over  the  generations. 

Interestingly,  considering  the  intergenerational  transmission of  IQ, Anger and Heineck 

(2009)  find  that  the  cognitive  abilities  of  parents  are  positively  correlated with  their 

children's  cognitive  skills,  also  after  controlling  for  educational  attainment  levels  and 

family background.  

Segregation 

A  key  element  in  the  current  political  discourse  on  the  participation  in Dutch 

society of  former  labour migrants and  their children  is  the assumed negative effect of 

segregated  residential  areas  and  schools.  The  Dutch  Study  Centre  for Mixed  Schools 

(Landelijk Kenniscentrum Gemengde Scholen)– an organization funded by the Ministry of 

Education,  Culture  and  Science–  is  dedicated  to  the  further  ethnical  integration  of 

schools  and  the  reduction  of  the  number  of  schools with  a  vast majority  of migrant 

pupils  or  native  Dutch  pupils.  Although  this  centre  strongly  believes  in  the  positive 

Page 45: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

19 

effects  of  mixing  school  populations,  it  recognizes,  nonetheless,  that  international 

research publications hardly present any  findings that demonstrate a positive effect of 

mixed  schools  on  learning  outcomes  (Bakker,  2012).  Interestingly,  Margolis  (2010) 

suggests that we may need a broader understanding of local communities– not just the 

level  of  segregation–  in  order  to  better  understand  the  stubbornness  of  the 

achievement gap.  

According  to  Ladd  et  al.  (2010),  segregation  levels  of  schools  are  high  in  the 

Netherlands compared to the U.S., and schools tend to be even more segregated than 

their  surrounding  neighbourhoods.  They  explore  the  question  whether  unlimited 

freedom of school choice in the Netherlands, actually increases school segregation.  

A further complicating factor in establishing the effects of school segregation on 

learning  outcomes  may  be  that  research  publications  on  the  negative  effects  of 

segregated schools and residential areas are often based on research findings based on 

impoverished neighbourhoods in metropolitan areas in the United States. Circumstances 

of  extreme  poverty,  a  high  incidence  of  teenage  motherhood,  widespread 

unemployment,  and  criminality  in  these  urban  areas  appear  to  be,  however,  hardly 

comparable to conditions in cities in Europe or Canada according to Oreopoulos (2007). 

Another  important  difference  between Dutch  and  other  European  cities,  on  the  one 

hand,  and  metropolitan  areas  in  the  U.S.,  on  the  other  hand,  may  be  that  the 

geographical  distances  between  low  SES  and  affluent  neighbourhoods  are  in  general 

considerably smaller in European cities. Interestingly, recent research publications show 

that  estimated  positive  effects  of  proximity  to  affluent  neighbourhoods  and  good 

schools  may  be  larger  than  estimated  negative  effects  of  low  average  SES  at 

neighbourhood level per se (Durlauf, 2004; Glaeser, Resseger and Tobio, 2008; Massey, 

2008).  

In addition, Hartog and Zorlu (2009) find that segregation in the Netherlands may 

be lower than the public discourse suggests: 95 % of the Moroccan Dutch population, for 

example, live in a neighbourhood with less than 30 % Moroccan Dutch residents.  

Finally, Dronkers  (2010)  finds  that  the diversity at  school  level, as expressed  in 

the  number  of  different  ethnic  groups,  has  a  stronger  negative  effect  on  attainment 

levels and  school outcomes  than  the  total percentage of migrants per  se. His  findings 

would  imply  that  a high percentage of one  specific ethnic  group would have no or  a 

Page 46: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

20 

limited negative effect on educational outcomes. A possible explanation  for this effect 

could be the "connectedness"4 at neighbourhood level (e.g. contacts among inhabitants, 

between  parents  and  school,  between  inhabitants  and  neighbourhood  services) 

resulting in close‐knit communities of people sharing the same background (e.g. Dobbie 

and Fryer, 2009). 

Desegregation 

Extensive  research  into  the  effects  of  the  desegregation  of  schools  and 

neighbourhoods  sometimes  leads  to  ambiguous  or  unexpected  results.  A  large‐scale, 

intensively  investigated  programme  for  desegregating  severely  disadvantaged 

neighbourhoods, called the Moving to Opportunities programme5 (MTO), did not result 

in any measurable effects on schooling (Oreopoulos, 2007). Furthermore, the enactment 

of the 1964 Civil Rights Act in the United States, that barred the use of federal money for 

segregated  schools,  initiated  a  synchronous  desegregation  of  schools,  thus  creating 

another  large scale natural experiment. A study  into the effects on black students who 

were sent to former white schools (far more numerous than vice versa) after the 1964 

Act, particularly  shed  a  light on  the negative effects of  losing  familiar neighbourhood 

schools, where teachers were part of the black community (Stuart Wells, Jellison Holme, 

Tijerina Revilla and Korantemaa Atanda, 2009). The causal effect on  learning outcomes 

of  a  comprehensive  school  and  community  approach,  based  on  the  idea  that  good 

schools should be brought to disadvantaged children (as opposed to bussing children to 

                                                       4 "Connectedness" is a term used in Ecology, referring to the density and specialization of niches (indicating a highly developed ecosystem) and the subsequent equilibrium of the system. The term is often borrowed to describe sociological systems, as well as the term "resilience" that describes the equilibrium between the capacity to change and the capacity to resist too much disturbance from the environment. 5 The Moving to Opportunity programme started in 1994 in Baltimore, Boston, Chicago, Los Angeles and New York. Originally the 1970 Housing and Urban development Act of 1970 allowed for the introduction of the Federal Experimental Housing Allowance programme. 

Page 47: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

21 

high quality schools elsewhere) in the Harlem Children's Zone (HCZ)6 programme, could, 

however,  indeed  be  proven  by Dobbie  and  Fryer  (2009),  notwithstanding  the  largely 

stratified nature of the area. 

Interestingly, as has been mentioned above, parents and children are not in any 

way  restricted  in  their  choice  of  school  in  the Netherlands, which  is  completely  free 

under  constitutional  law;  nor  are  they  in  any way  limited  in  their  choice  by  financial 

constraints,  since  there  is  no  tuition  fee  and  all  schools  are  equally  funded  by  the 

government. Migrant parents and children, however, overall tend to prefer schools with 

larger  groups  of migrant  pupils,  both  in  elementary  and  in  secondary  education.  By 

implication,  this  would mean  that  compulsory  desegregation  policies  would  restrain 

parental  choice,  while  we  as  yet  do  not  sufficiently  understand  possible  beneficial 

effects of specialized, albeit stratified neighbourhood schools. 

Actually lived lives of individuals 

Crul et al. (2012) started a bold endeavour to compare the actually lived lives of 

people with  an  immigrant  family  history,  but  born  in  the  country  of  destination  (the 

second  generation),  in  six  European  countries  (the  Netherlands,  Belgium,  Sweden, 

Austria, Switzerland, and Germany). The aim of  their study was  the assessment of  the 

relations between  "the  integration  context", notably  the educational  system,  and  the 

position  of  second  generation  migrants  in  the  six  countries.  To  this  end,  they 

interviewed,  by  means  of  a  questionnaire,  500  individuals  belonging  to  the  second 

generation and 500  individuals from the comparison group (of native‐born parentage), 

who were  between  18  and  35  years  old  in  2007/2008.  Interestingly,  considering  the 

second generation of respondents with a Turkish family history in the Netherlands, Crul 

et al. (ibid.) find that the odds of dropping out without a diploma are roughly equal to 

                                                       6 In the early 1990s, HCZ ran a pilot project that brought a range of support services to a single block. The idea was to address all the problems that poor families were facing: from crumbling apartments to failing schools, from violent crime to chronic health problems. HCZ created a 10‐year business plan, to ensure its best‐practice programmes were operating as planned. HCZ was in the vanguard of non‐profit organizations that began carefully evaluating and tracking the results of their work. Those evaluation results enabled staff to see whether programmes were achieving their objectives, and to take corrective actions if they were not. In 1997, the agency began a network of programmes for a 24‐block area: the Harlem Children's Zone Project. In 2007, the Zone Project grew to almost 100 blocks. Today the Children's Zone® serves more than 8,000 children and 6,000 adults. Overall, the organization serves more than 10,000 children and more than 7,400 adults.  The FY 2010 budget for the agency overall is over $75 million (text retrieved from HCZ website).  

Page 48: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

22 

the  odds  of  being  exceptionally  successful  and  achieving  far  higher  attainment  levels 

than their parents. They identified– among other factors– two specific characteristics of 

the Dutch  educational  system  that work  out  positively  for  pupils  of migrant  descent: 

opportunities  for  upstreaming  and  downstreaming  in  secondary  schools,  and  longer 

alternative  routes  up  the  vocational  ladder,  eventually  granting  access  to  higher 

education.  The  authors  report  that  "between  half  to  two  thirds  of  Turkish  second 

generation occupy a stable lower‐ to upper‐middle‐class position…Some members have 

taken a spectacular step in just one generation". 

2.3 Methods and data 

2.3.1 Selecting the data  

The main goal of this study is a better understanding of the success of Moroccan 

pupils; for this reason, access to the highest levels of secondary education at age 16, has 

been defined as the measure of success for the purpose of this research. Typically, pupils 

pass  their  final  exam  in  these  two  secondary  academic  tracks  at  age  17  and  18, 

respectively. At age 16,  therefore,  the number of pupils who have already passed  the 

final exam  is negligible. The  two highest secondary  tracks have been combined  in  this 

research since both tracks qualify the pupils for access to higher education; they will be 

further referred to as academic tracks. At age 16 pupils are typically enrolled in Grade 4 

of  secondary  schools.  For  this  research,  selecting  pupils  at  age  16  was  preferred, 

however, and not enrolment in the fourth year of secondary school. At grade level pupils 

may be of a different age because of grade repetition, and as the result of an age‐range 

of almost 12  to almost 13 when entering  secondary education;  this would complicate 

measurements of the proportion of the population of 16 year‐olds enrolled in academic 

tracks at the zip‐code level.  

This  chapter  focuses,  in  particular,  on  pupils  with  a  Moroccan  immigrant 

background.  The  interpretation  of  research  outcomes  on  educational  achievement  in 

different ethnic groups  is often complicated by  the  frequent use of combined migrant 

groups, notably the indication Non‐western and Western immigrants. These widely used 

combination  of  ethnic  groups  in  research  publications  and  national  statistics,  neglect 

considerable  differences  per  country  of  origin,  notably  the  educational  system,  and 

different  reasons  for  emigration  by  distinct  groups  (de  Heus  and  Dronkers;  2008, 

Page 49: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

23 

Noguera, 2000). The vast majority of Moroccan Dutch 16 year‐olds belong to the second 

generation;  the measurements  presented  in  this  chapter  do  not  distinguish  between 

first and second generation.  In  the Netherlands different groups  in  the population are 

indicated by ethnicity rather  than  race, as  is  for  instance more common  in  the United 

States.  

The four major cities (Amsterdam, Rotterdam, Utrecht and The Hague) by 4‐digit 

zip‐code have been  chosen  as  the  spatial entity,  for methodological  reasons:  the  zip‐

code, although not without  its own  limitations,  is  internationally comparable and more 

strictly defined than neighbourhood, district, community, or borough.  

In  middle‐sized  and  small  cities,  the  number  of Moroccan  Dutch  (and  other 

ethnicities) is often too small at zip level for reliable analyses; the majority of Moroccan 

Dutch pupils live in the four major cities. Therefore, analyses have been carried out only 

for the four major cities. Also to avoid too small numbers in some zips in the four major 

cities, only zips with more than five Moroccan Dutch pupils were selected for analyses, 

unless clearly indicated otherwise.  

Separate measurements have been carried out for zip areas that qualify for extra 

financial  support  by  the  Dutch  Ministry  of  Housing,  Spatial  Planning  and  the 

Environment7;  these socio‐economically severely disadvantaged zips are  referred  to as 

zips at risk in this text. 

Apart  from the main data set that was especially assembled  for the purpose of 

this  research,  also  publicly  accessible  data  published  by  Statistics  Netherlands  (the 

National Statistical Office) and  the Central Financial  Institution  (CFI, part of  the Dutch 

Ministry of Education, Culture and Science) have been reworked. These tables, based on 

these public data bases, are presented in the Appendices (2.6).  

                                                       7 In 2007 the Ministry of Housing, Communities and Integration (Tweede Kamer der Staten Generaal, (2007) started a programme to support, with extra funding, neighbourhoods with an accumulation of problems. Qualification for extra funding is based on a list of possible disadvantages. Among the neighbourhoods that participate in this programme are many zips with large numbers of inhabitants with a migrant background. A list of 340 zip‐codes (out of around 4800 in total in the Netherlands) that qualified for extra support was selected by the Ministry based on four groups of indicators:  

Socio‐economic indicators like average income, employment and educational attainment; 

Infrastructural indicators like size, age, condition and price of houses; 

Social problems like vandalism, harassment and perceived unsafe situations; 

Physical annoyances like noise, traffic, street litter and pollution; The Ministry gave priority to a selection of 40 communities out of these 340, covering in total 83 zip‐codes, among these 59 in the four major cities. At least five Moroccan pupils of 16 years old live in 47 out of these 59 zip‐codes in the major cities; these are used for measurements in this paper.  

Page 50: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

24 

2.3.2 Assembling the data set 

The relevant data necessary for the measurement of the correlation between the 

percentage of 16 year‐old Dutch Moroccans per zip‐code, and their access to academic 

tracks  in  secondary  education,  are  managed  by  two  different  institutions,  CFI  and 

Statistics  Netherlands:  all  address  (zip)  information  that  can  be  traced  back  to  the 

individual  is  collected  by  Statistics  Netherlands  whereas  all  educational  data  are 

registered in what is called the BRON8 database by CFI9.  

The data  set CFI made available  (reference date 1 October 2008)  contains  the 

following information: The data set covers all 16 year‐olds by ethnicity, who are enrolled 

in academic  tracks  in  secondary education, by 4‐digit  zip‐code  level  in  the  four major 

cities.  Nationwide  the  group  of  16  year‐olds  consists  of  around  315.000  individuals; 

around 28.000 live in the four major cities. 

The above data cannot be used to measure the percentage of Moroccan Dutch 

pupils enrolled in academic levels, as a share of their own ethnic group at zip level, since 

the data do not cover the total number of 16 year‐old Moroccan Dutch per zip‐code.  

To measure  the  share  that  enrols  in  secondary  academic  tracks,  the  CFI  data 

were linked at the 4‐digit zip level to a tailor‐made data set made available by Statistics 

Netherlands, with the following description: 

The data set covers all 16 year‐olds (not only those in academic tracks) by 4‐digit 

zip‐code and ethnicity.  

Individuals  are  regarded  as migrant pupils when  at  least one parent was born 

abroad (definition Statistics Netherlands). 

The file contains the year and the month of birth, but not the day. Therefore all 

individuals born on 1 October 1990, up to and including 30 September 1990, are 

included in this file. 

The reference date is 26 September 2008 

                                                       8 BRON: Dutch acronym: Basis Register Onderwijs Nummer. In the new millennium the Dutch government started the phased introduction of one single registration number for all citizens, thus combining for instance registration numbers for education, family related data, income and social security, further training, medical and police records.  9 During the current research, CFI has been in the process of merging, which has resulted in a new agency, called DUO (Dutch acronym, Dienst Uitvoering Onderwijs). DUO is the office for all financial affairs in education, the organization of final exams, and the management of student funding) and is part of the Ministry of Education, Culture and Science. 

Page 51: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

25 

By merging these two data sets, the share of Moroccan youth at zip‐code  level 

could be calculated, as well as the percentage of them enrolled on academic tracks. The 

data set contains 2947 16 year‐old Moroccan Dutch living in the four major cities: 1290 

in Amsterdam, 719 in Rotterdam, 501 in Utrecht, 437 in The Hague.  

2.3.3 Data from public files 

CFI and Statistics Netherlands also allow public access to parts of their databases, 

and  regularly  publish  updated  statistics  considering,  for  example,  education, 

demography, economic development and the labour market. The publicly accessible site 

Statline by Statistics Netherlands offers  tools  to adapt public data  to  specific  research 

questions. Some of these tables and figures were reworked as a reference for findings in 

this  research. Figures based on publicly available data are  included  in  the Appendices 

and  show:  1)  rising  levels  of  educational  attainment  among  people with  a Moroccan 

immigrant  background  age  25  to  65–  2)  decreasing  percentages  of  people  with 

elementary  education  only–  3)  rising  numbers  of  people  with  at  least  an  initial 

qualification for the labour market10; and 4) raising enrolment levels in academic tracks 

among 16 year‐old pupils with a migrant background between 2003‐2009. The tables in 

the  Appendices  show  a  sharp  rise  in  educational  attainment  levels  for  all  ethnicities 

when the cohorts 55‐65 year‐olds are compared to 25‐35 year‐olds (Appendices 2‐1, 2‐2, 

2‐3, and 2‐4). This increase is, however, different for distinct ethnic groups. 40% among 

native‐born  in  the  cohort  55‐65  year‐olds,  did  not  accomplish  ISCED  3  level,  the 

minimum  requirement  for  the  labour market;  among  the  25‐35  year‐old  native‐born, 

16% did not reach this level, a decrease of 24 percentage points (Appendix 2‐1). People 

with  a  migrant  background  improved  average  levels  of  schooling  far  stronger:  80% 

among Moroccan Dutch  55‐65  year‐olds  against  34%  among  25‐35  year‐olds  did  not 

accomplish  ISCED  3,  a  decrease  of  46  percentage  points  (Appendix  2‐1). 

Correspondingly,  the  percentage  with  only  elementary  education  as  the  highest 

                                                       10 Initial qualification for the labour market is a standard introduced by the OECD (Organization for Economic Co‐operation and Development) and is defined as the completion of at least the ISCED 3 level. The ISCED qualification (UNESCO 1997) is precisely introduced to make international agreements possible on required educational attainment levels, thus facilitating the international labour market and stimulating student mobility. The OECD member states agreed that ISCED 3 is the basic qualification for entering the labour market. Of course ISCED 3 refers to different nomenclature for educational tracks in the member states; in the Netherlands it refers to a diploma at pre‐university (VWO), senior general (HAVO) level or advanced levels of senior secondary vocational education (MBO). 

Page 52: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

26 

completed level decreases strongly for all groups‐ but mostly for Turkish and Moroccan 

Dutch  (Appendix  2‐3).  Furthermore,  among  Moroccan  Dutch  25‐35  year‐olds,  16% 

obtained a bachelor or master's degree or a PhD  (38% native‐born  in the same cohort 

reached this level); this was 0% among 55‐65 year‐olds Moroccan immigrants (Appendix 

2‐1).  

2.4 Results  

The  tables  and  figures  presented  in  this  Section  follow  the  three  research 

questions: 

Question 1, which considers  the association between ethnic composition at zip 

level  and  access  to  academic  secondary  tracks:  Table  2‐1,  as  an  introduction, 

presents an overview of  the number of 16 year‐olds at  zip‐code  level, and  the 

share of them who are of Moroccan descent. Figure 2‐1 depicts the regression of 

enrolments among youth with a Moroccan background  in academic tracks, and 

ethnic  composition  at  the  zip‐code  level,  in  the  four major  cities  combined; 

Figure 2‐2 shows regressions for enrolments separately for the four major cities.  

Question  2,  which  looks  specifically  at  zip  areas  at  risk,  with  severe  socio‐

economic challenges: Figure 2‐3  illustrates enrolments  in  zips at  risk; Table 2‐3 

compares all zips, and zips at risk. Regression coefficients are listed in Table 2‐2. 

Question 3, which considers  the  total percentage of  immigrants at zip  level  (all 

ethnicities)  and  access  to  academic  secondary  tracks:  Table  2‐4  presents  the 

correlation  between  enrolments  in  academic  tracks  among  Moroccan  Dutch 

youth,  and  the  total  ethnic  composition  (including  youths  with  a  Turkish, 

Antillean/Surinamese  or  other  immigrant  background)  of  the  zip‐code  area 

where they live. 

2.4.1 Enrolments in academic tracks among Moroccan Dutch, at zip‐code level 

Table 2‐1, as an overview, shows the average number of 16 year‐olds per zip, and 

the average number and percentage of them who are of Moroccan descent. Amsterdam 

has the highest number of youngsters of Moroccan descent, Utrecht the highest average 

percentage at zip‐code level. 

Page 53: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

27 

Table 2‐1: Total average of 16 year‐olds per zip in the four major cities, and average number and percentage of 16 year‐old Moroccan Dutch among this group 

  Amsterdam  Rotterdam  Utrecht  The Hague  Total 

# zips  42  37  22  20  121 

Total average of 16 year‐olds at zip level  125  135  70  133  111 

Average # Moroccan 16 year‐olds at zip level  26  18  21  19  19 

% Moroccan among 16 year‐olds at zip level  21%  13%  30%  14%  17% 

Notes: Only zips with > 5 Moroccan 16 year‐olds included. 

Figure 2‐1: Data scatter, linear regression (OLS) and LOWESS regression of the percentage of 16 year‐old Moroccans enrolled in academic secondary tracks per zip on the percentage of Moroccan Dutch 16 year‐olds as a share of the total number of 16 year‐olds at zip level in the four major cities in the Netherlands 

 

Note: Only zips > 5 Moroccans of 16 years old have been included. 

Figure  2‐1  shows  the  regression  (LOWESS  and  linear)  of  the  percentage  of  16 

year‐old Moroccan Dutch pupils enrolled  in academic secondary tracks on the share of 

Moroccan Dutch among 16 year‐olds at zip‐code  level. LOWESS weights the data point 

by point within a moving band  containing 40%  (in  this  case) of  the data  set, and has 

been added to check that the linear fitted line was not biased because of the relatively 

small number of data points  (zips). Figure 2‐1  shows  that  the association between  zip 

020

4060

80%

in a

cade

mic

trac

ks

0 20 40 60% Moroccan 16 year olds per ZIP

Page 54: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

28 

composition and enrolments  in academic  tracks  is negative  (coefficient  ‐0.21**, Table 

2‐2). Measurements show a relatively high dispersion of data points in zips where up to 

20% of 16 year‐old Moroccans live. As a reference for these findings, an overview of the 

national  dynamics  in  enrolment  levels  among  16  year‐old  students  with  a  migrant 

background (2003‐2009) has been included in Appendix 2‐4. 

Figure 2‐2: Data scatter and linear regression (OLS) of the percentage of 16 year‐old Moroccans enrolled in secondary academic tracks on the share of Moroccan Dutch 16 year‐olds at zip level, separately for the four major cities in the Netherlands 

 

Enrolments in academic tracks among students of Moroccan descent, separately 

for  the  four  major  cities  in  Figure  2‐2,  indicate  higher  access  in  Amsterdam  and 

Rotterdam, as compared to Utrecht and The Hague. An explanation for these differences 

in access between the four major cities  is beyond the scope of this research, since the 

data  do  not  allow  for  further  socio‐economic  analyses  that  might  explain  these 

differences.  Appendix  2‐5,  based  on  new  educational  data  on  current  enrolments  in 

secondary  Year  3, has  been  added, however, because  this new data  set does  indeed 

cover  socio‐economic  indicators: Appendix 2‐5  confirms  that enrolments  in  secondary 

020

4060

80%

in a

cade

mic

trac

ks

0 20 40 60% Moroccan 16 year olds per ZIP

Amsterdam RotterdamUtrecht The Hague

Page 55: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

29 

tracks are higher  in Amsterdam, but does not  indicate a higher average SES at zip  level 

as  a  possible  explanation;  neither  does  the  relatively  low  average  SES  at  zip  level  in 

Rotterdam  in  Appendix  2‐5,  offer more  insight  into  the  higher  enrolment  levels  for 

Rotterdam  in  Figure  2‐2.  Finally,  as  a  reference,  Appendix  2‐6  shows  the  association 

between ethnic composition at zip  level and enrolments among native Dutch students. 

Although  their  average  enrolments  are  overall  higher  than  in  the  case  of  migrant 

students,  enrolments  among  native  Dutch  decline  with  increasing  percentages  of 

migrant youths at  zip  level. A  further analysis of  socio‐economic  indicators  that might 

explain these findings is presented in Chapter 4 of this thesis. 

2.4.2 Enrolments in academic tracks in zips at risk 

In  2007  the  Ministry  of  Housing,  Communities  and  Integration  earmarked  a 

budget  for  the  improvement of  living conditions  in neighbourhoods with accumulated 

socio‐economic  problems.  Priority  was  given  to  a  selection  of  40  neighbourhoods, 

covering in total 83 zip‐codes, among which 59 were in the four major cities. When only 

zip‐codes with more than five 16 year‐old Moroccan Dutch are considered, 52 zips at risk 

remain  for measurements  in  this  Section.  Young  people  in  these  zips  are  supposedly 

extra challenged to  find their way to academic tracks  in secondary education, because 

average  levels of schooling among parents  in zips at risk are  low, household  income  is 

below  average,  and  unemployment  is  high.  General  theory  identifies  these 

circumstances as negative predictors for educational attainment levels. 

Figure  2‐3  shows  a  negative  slope  of  the  regression  line  (coefficient  ‐0.25**, 

Table 2‐2), much like the measurements on all zips in Figure 2‐1, with a flattening out of 

the  LOWESS  regression  line  for  zips with around 20‐40% of 16 year‐olds of Moroccan 

descent. Zips with  lower percentages of Moroccan Dutch 16 year‐olds, do better  than 

the national average for this group (see Appendix 2‐4), while those who live in zips with 

more  than  40%  of  Moroccan  Dutch  16  year‐olds  fall  behind.  The  socio‐economic 

characteristics  for  zips  at  risk mentioned  above  are  negatively  correlated  to  average 

levels of schooling  in many studies (e.g. De Witte, 2010). Although causal explanations 

for  the enrolments  in academic  secondary  tracks among Moroccan pupils are beyond 

the  scope of  this  research,  the  results do not  indicate  that  students  are  less  likely  to 

Page 56: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

30 

enrol in academic tracks in zips at risk when compared to the average enrolments in all 

zips combined (Figure 2‐1). 

Figure 2‐3: Data scatter, linear regression (OLS) and LOWESS regression of the percentage of 16 year‐old Moroccans enrolled in academic secondary tracks per zip at risk on the share of Moroccan Dutch 16 year‐olds at zip level in the four major cities in the Netherlands. 

 

Table 2‐2: Regression coefficients of the percentage enrolments in academic tracks among 16 year‐old Moroccans on the share of Moroccan 16 year olds at zip‐code level  

  All zips  At risk  Amsterdam  Rotterdam  Utrecht  The Hague 

% Moroccans  ‐0.21**  ‐0.25**  ‐0.23  ‐0.26  ‐0.22  ‐0.14 

  (0.09)  (0.10)  (0.18)  (0.26)  (0.15)  (0.27) 

Constant  25.16***  27.52***  28.28***  26.64***  22.43***  20.23*** 

  (2.28)  (3.09)  (4.91)  (4.70)  (4.91)  (4.47) 

N  121  52  42  37  22  20 

Mean  20.7  20.9  22.8  22.5  15.9  18.2 

Notes: Standard errors in parentheses. */**/*** denote significance at the 10/5/1% confidence level. 

020

4060

80%

in a

cade

mic

trac

ks

0 20 40 60% Moroccan 16 year olds per ZIP at risk

Page 57: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

31 

 Table 2‐3: Correlation of the percentage of 16 year‐old Moroccans per zip, and the percentage among them in secondary academic tracks, for all zips and for zips at risk separately 

City  Correlation  P‐value  #zips 

  All zips  Zips at risk  All zips  Zips at risk  All zips  Zips at risk 

Amsterdam  ‐0.194  0.032  0.219  0.910  42  15 

Rotterdam  ‐0.168  ‐0.284  0.321  0.200  37  22 

Utrecht  ‐0.313  ‐0.429  0.156  0.289  22  8 

The Hague  ‐0.122  ‐0.392  0.607  0.385  20  7 

Total  ‐0.204  ‐0.322  0.025  0.020  121  52 

Notes: Significant correlations p≤ 0.05 in bold. 

In addition to the regression analyses, the correlation between the percentage of 

16 year‐old Moroccan Dutch and their enrolments in academic tracks was measured as 

well (Table 2‐3). 

Caution  is  needed  because  measurements  in  Table  2‐3  are  based  on  small 

numbers of zips, notably the numbers of zips at risk. Note that in the case of Rotterdam, 

22  out  of  37  zip  areas  qualified  for  extra  government  funding  because  of  persistent 

socio‐economic  disadvantages. No  significant  correlations were  found  for  16  year‐old 

Moroccans and enrolments among them in academic tracks, both in all zips– and in zips 

at risk, when considering the four cities separately; a significant negative correlation was 

found, however,  in total counts of the four cities combined, probably as a result of the 

larger total number of zips (and a smaller standard error). 

2.4.3 The  correlation  between  the  total  percentage  of  migrants  at  zip  level  and enrolments of Moroccan Dutch 

Table 2‐4 presents correlations between enrolments of Moroccan Dutch and the 

total percentage of 16  year‐old migrants at  zip  level  (all  immigrant groups  combined, 

Moroccan, Turkish, Surinamese/Antillean and other). Note  that  in Table 2‐4, unlike all 

the other main measurements,  all  zips were  included where  at  least  one  16  year‐old 

individual  of  Moroccan  descent  lives  (in  other  measurements  only  zips  with  >  5 

Moroccan  16  year‐olds were  included). Only  in Rotterdam  is  the  correlation negative 

and  significant  (‐0.416, P‐value of 0.002). For  the other  three  cities,  the  correlation  is 

smaller and not significant.  

Page 58: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

32 

Table 2‐4: Correlations between total percentage of all immigrants per zip‐ and the percentage of 16 year‐old Moroccans in secondary academic tracks, separately for the four major cities.  

City  Correlation P‐value  # zips

Amsterdam  ‐0.074  0.579  59 

Rotterdam  ‐0.416  0.002  52 

Utrecht  ‐0.237  0.178  34 

The Hague  0.150  0.338  43 

Total  ‐0.086  0.241  188 

Notes: Significant correlation p≤0.05 in bold 

2.5 Conclusions and discussion 

Three questions have been explored  in  the current chapter: 1)  Is enrolment  in 

academic secondary tracks of 16 year‐olds with a Moroccan background, related to the 

percentage at zip  level of Moroccan 16 year‐old youngsters; 2) When considering only 

zips with persistent socio‐economic challenges, do the results differ from measurements 

in  all  zips;  and  3)  Is  de  total  percentage  of  youths  with  an  immigrant  background 

associated with  enrolments  at  academic  secondary  levels  of  youths with  a Moroccan 

background. 

A  major  assumption  underlying  policymaking  at  the  government  and  the 

municipal  level  is  the  assumed  negative  effect  of  segregated  neighbourhoods  on 

educational  outcomes.  The  results  presented  in  this  chapter  nuance  this  assumption: 

enrolment percentages at zip‐code  level show a  large dispersion,  leading to a question 

for further research: Why do zip areas with the same percentages of 16 year‐old youths 

of Moroccan descent, have quite different enrolment rates. In zip areas with up to 20% 

16  year‐old  youngsters of Moroccan descent,  average enrolments  are on  the  level of 

average  national  enrolments  for  their  group  (20.3%  in  2009).  Overall,  however,  the 

association is negative, with a regression coefficient of ‐0.21** (all zips considered): for 

every 10%  increase of Moroccan 16 year‐olds per  zip, enrolments  in  secondary  tracks 

decrease by 2.1 percentage points. The decrease  is 2.5 percentage points  considering 

zips at risk. 

Earlier  analyses  by  Hartog  and  Zorlu  (2009)  show  that  95%  of  people  of 

Moroccan descent  live  in a neighbourhood with  ≤ 30% of people who have  this  same 

background,  indicating  that  segregation  is  more  moderate  than  public  and  political 

debate  suggests.  Considering  the  zips with  ≤  30% Moroccan  youths  at  zip  level,  no 

Page 59: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

33 

convincing association was found between zip composition and enrolments in high level 

tracks.  

The  findings  in  the current  study do not  ignore  there may be valid  reasons  for 

school  desegregation:  learning  and  living  together  in  mixed  schools  may  add  in 

important ways to skills and valuable experiences that enhance the quality of the future 

labour  force  in a multicultural society. However,  in order to be able to make balanced 

policy  decisions,  further  research  should  also  elucidate  the  effect  of  specific  school 

policies  in  schools with  a majority  of migrant  pupils  (e.g.  specialized  further  teacher 

training, cooperatives with migrant parents, and joint community‐based programmes) as 

the  effect  of  desegregation.  Furthermore,  unavoidably,  desegregation  policies,  like 

centralized school assignment, would limit free school choice in the process. Finally, the 

added value of specialized, albeit segregated, schools may change over  time: whereas 

societal  groups  that  are  in  the  emancipatory  process  of  substantially  increasing  their 

educational attainment  levels may benefit  from  specialized  schools, over  time a more 

balanced school composition could become more important.  

Extrapolating the  index of the  increase  in enrolments  in academic tracks among 

Moroccan pupils between 2003 and 2009  (31%:  see Appendix 2‐4), and assuming  this 

speed of  increase could be sustained, Moroccan youths could be expected to catch up 

with current enrolments among native‐born students around  the year 2025. However, 

attainment  levels of such a  large extent affect both the prospects of the  individual and 

the potential of the workforce as a whole, so further research into strategies that could 

speed up the closing of the gap remain crucially important. 

While  the  overall  ethnic  composition  at  zip  level  (all  ethnicities,  including 

Moroccan background) does not show a significant correlation with secondary academic 

enrolments,  there  is  one  exception:  in  Rotterdam  the  correlation  is  significant  and 

negative  (‐0.416,  Table  2‐4).  Interestingly,  this  finding  corresponds  to  other  research 

which demonstrates that Rotterdam seems to face extra challenges: for example, early 

school  leaving  (De Witte, 2010). At  first  sight,  the  findings  for Rotterdam  in Table 2‐4 

seem to be in contrast to the findings in Figure 2‐2: in Figure 2‐2 enrolments in academic 

tracks were  regressed  on  the  percentage  of Moroccan  youths  at  zip‐code  level,  and 

show  higher  enrolments  both  in  Amsterdam  and  Rotterdam  in  comparison with  the 

other  two  cities. However,  in Table 2‐4, unlike  in  Figure 2‐2,  the  correlation between 

Page 60: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

34 

enrolments  and  the  total  percentage  of  youths  of migrant  descent  is  considered.  A 

possible  explanation  for  these  divergent  outcomes  may  be  the  different  ethnic 

composition at zip‐code level in Rotterdam: Moroccan youngsters in Rotterdam tend to 

live  in neighbourhoods with on  average higher percentages of people with  a migrant 

background (all ethnicities considered) than is the case in the other three cities, but with 

a lower average percentage of Moroccan 16 year‐olds than is the case in the other three 

cities.  Furthermore,  interestingly, updated BRON data  (listing enrolments  in  academic 

tracks  in Year 3 of secondary school  in 2011/2012) show a considerably  lower average 

SES  index  for  neighbourhoods  in  Rotterdam  (Appendix  2‐5).  Their  enrolments  in 

academic  tracks,  as measured  at  the  entrance  of  Year  3  in  secondary  schools,  seem 

substantially lower (16%) than is, for example, the case for Amsterdam (27%), although 

Utrecht shows the same low percentage of enrolments as Rotterdam.  

In  summary,  most  Moroccan  pupils  live  in  neighbourhoods  with  many 

inhabitants  of  migrant  descent.  Whether  Moroccan  pupils  would  have  been  more 

successful  in qualifying  for high  secondary  tracks, as a  result of desegregation policies 

through restrictions of free school choice or housing policies, is a very intricate question. 

Such  policies would  be  expected  to  have  far‐reaching  consequences  for  families  and 

individuals. Moreover, neither can we estimate adequately whether these interventions 

would  indeed speed up  the closing of  the educational achievement gap. Although  this 

study demonstrates that an increase in the ethnic segregation of Moroccan youths at zip 

level significantly reduces enrolments in academic tracks by around 2 percentage points 

per 10% increase of Moroccan youths, the size of the effect may not counterbalance the 

impact of desegregation policies of neighbourhoods and schools, especially not vis‐à‐vis 

the deeply valued right to the free choice of where to live and where to go to school in 

the Netherlands.  

The  findings  presented  in  the  current  chapter  have  inspired  further  research 

projects, which  are  presented  in  the  following  chapters  of  this  thesis. One  project  is 

based on  a  detailed  analysis of  individual  school histories  of  youth  living  in  three  zip 

areas in the district Amsterdam‐West, which face substantial socio‐economic challenges 

(Chapter  3).  Furthermore,  the  analysis  of  the  association  between  the  ethnic 

composition of the area of residence and school choice also inspired an investigation of 

the distance  travelled  to  school,  controlled  for a  rich  set of  socio‐economic  indicators 

Page 61: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

35 

(among which ethnic background). Since school choice is free and not restricted by any 

financial  barriers  in  the  Netherlands,  distance  to  school  may  reveal  interesting 

differences in patterns of school choice (Chapter 4). 

Finally,  the  current  research  raises  some  important  qualitative  questions,  for 

example, on  the motivations of migrant  students  to opt  for  academic  tracks,  and, on 

availability  of  information  for  parents  on  the  consequences  of  school  choice  with 

reference to future success on the labour market of their children. 

2.6 Appendices 

The  figures  presented  below may  offer  relevant  additional  information  to  the 

main  findings presented  in Section 2.4  (Results). These  statistics are not based on  the 

tailor‐made data set used for main measurements in this chapter, but either on the total 

BRON database, or the publicly accessible data source Statline11, managed by Statistics 

Netherlands.  

Increasing educational attainment  

The research questions in this paper put a focus on 16 year‐old Moroccan Dutch 

in academic tracks in secondary education. Moroccan labour immigration started in the 

1960s. Most Moroccan  immigrants  came  to  the Netherlands  to be employed,  at  that 

time,  in unskilled  labour. Their average educational attainment  levels were  (very)  low, 

and so were the levels of schooling of their wives and children who followed in the years 

to come. The  table  in Appendix 2‐1 shows how attainment  levels have  increased over 

time. The completion of ISCED 3 is the minimum requirement or initial qualification for 

the labour market, a standard agreed among OECD countries. 

As mentioned above, the completion of  ISCED 3  is an  important boundary  level 

for access to the labour market. Between Moroccan Dutch aged 55‐65 and aged 25‐35, 

the  initial  qualification  for  the  labour market  has  increased  by  47  percentage  points 

(Appendix 2‐2). The percentage of Moroccan Dutch with only elementary education as 

the highest completed level has decreased accordingly, by 48 % points, as is illustrated in 

Appendix 2‐3. 

                                                       11 www.cbs.nl 

Page 62: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

36 

Appendix 2‐4  shows  that  in 2008  the national  average enrolment  in  academic 

tracks was 19% among Moroccan Dutch, against 43.9% among native‐born pupils. The 

index for the relative increase in enrolments in academic tracks between 2003 and 2009 

is, however, different for native‐born and Moroccan migrants: the relative  increase for 

native‐born over  this period  is 17%  (100 x  (45.1  : 38.5) = 117), and 31%  for Moroccan 

youths (100 x (20.5 : 15.6)= 131). 

Appendix 2‐1: Highest completed level of education, by age and ethnicity (2007) 

Ethnicity  Level of education  Age 

    25‐35  35‐45  45‐55  55‐65 

Native Dutch  Primary education  3%  4%  8%  12% 

  ISCED 2  13%  18%  22%  28% 

  ISCED 3  45%  47%  41%  35% 

  Bachelor  25%  19%  19%  16% 

  Masters, PhD  13%  11%  9%  8% 

  Unknown  1%  1%  0%  0% 

Moroccan  Primary education  12%  24%  50%  60% 

  ISCED 2  22%  25%  20%  20% 

  ISCED 3  46%  36%  30%  20% 

  Bachelor  11%  3%  0%  0% 

  Masters, PhD  5%  7%  0%  0% 

  Unknown  5%  5%  0%  0% 

Turkish  Primary education  15%  27%  39%  64% 

  ISCED 2  29%  22%  21%  18% 

  ISCED 3  42%  40%  25%  18% 

  Bachelor  8%  3%  0%  0% 

  Masters, PhD  5%  5%  7%  0% 

  Unknown  0%  3%  7%  0% 

Notes:   ISCED 2: VMBO pre‐vocational secondary education; ISCED 3: MBO, level 4 (senior general vocational education); HAVO (senior general education); VWO (pre‐university education). 

Source: Statistics Netherlands, Statline, based on the EBB questionnaire to the labour force. 

   

Page 63: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

37 

Appendix 2‐2: Percentage of pupils with initial qualification for the labour market (ISCED 3 and higher), by ethnicity and age cohort (2007) 

  

Source: Statistics Netherlands, treated. 

Appendix 2‐3: Percentage of pupils with primary education as highest completed level, by ethnicity and age cohort (2007) 

  

Source: Statistics Netherlands, treated. 

   

0

20

40

60

80

100

Native Dutch Moroccan Turkish

% with basic qualification

Ethnicity

25-35

35-45

45-55

55-65

0

20

40

60

80

100

Native Dutch Moroccan Turkish

% with primary education only

Ethnicity

25-35

35-45

45-55

55-65

Page 64: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

38 

Appendix 2‐4: Enrolment of 16 year‐olds by ethnicity in senior general education (HAVO, ISCED 3) and pre‐university education (VWO, ISCED 3) 

    2003  2004  2005  2006  2007  2008  2009 

Total  Senior general  20.1%  20.4%  21.1%  21.3%  22.0%  22.3%  22.7% 

  Pre‐university  16.1%  16.9%  17.5%  18.0%  18.8%  19.1%  19.5% 

  combined  36.2%  37.3%  38.7%  39.3%  40.8%  41.3%  42.3% 

Native‐born  Senior general  21.4%  21.7%  22.4%  22.6%  23.4%  23.5%  24.2% 

  Pre‐university  17.1%  18.0%  18.7%  19.2%  20.0%  20.4%  20.9% 

  combined  38.5%  39.7%  41.1%  41.8%  43.4%  43.9%  45.1% 

Moroccan  Senior general  10.9%  10.8%  12.2%  12.2%  12.0%  13.4%  13.5% 

  Pre‐university  4.7%  5.1%  5.5%  5.5%  6.5%  5.6%  7.0% 

  combined  15.6%  15.9%  17.7%  17.7%  18.5%  19.0%  20.5% 

Turkish  Senior general  11.8%  11.9%  12.7%  12.3%  12.8%  13.4%  13.9% 

  Pre‐university  4.9%  6.0%  5.6%  6.5%  6.8%  6.7%  6.7% 

  combined  16.7%  17.8%  18.3%  18.7%  19.6%  20.0%  20.6% 

Notes:  The two secondary tracks combined are referred to in the text as academic tracks, both qualifying pupils for access to higher education. 

Source: CFI, BRON data. I am deeply grateful to Gert Korteweg of the Ministry of Education, Culture and Science for his technical help in assembling this table and his collegial advice. 

   

Page 65: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

2 The educational gap and neighbourhood composition 

39 

Appendix 2‐5: Percentage of pupils enrolled in academic track in Year 3 of secondary school (school year 2011/2012), mean neighbourhood SES (scale ‐4 to +4), and percentage of pupils eligible for weighted student funding when they were enrolled in elementary education. 

Amsterdam 

Ethnicity  % in academic track in  Neighbourhood SES  % weighted  N 

Native‐born  62  0.45  2  1737 

Moroccan   27  ‐0.89  60  968 

Turkish  28  ‐0.83  56  538 

Surinamese/Antillean  23  ‐0.52  31  762 

Other immigrant  44  ‐0.30  33  1054 

Total  42  ‐0.24  30  5059 

Rotterdam 

Ethnicity  % in academic track  Neighbourhood SES  % weighted  N 

Native‐born  47  0.38  4  1777 

Moroccan   16  ‐1.36  56  592 

Turkish  19  ‐1.27  52  659 

Surinamese/Antillean  19  ‐0.84  39  846 

Other immigrant  37  ‐0.62  39  900 

Total  32  ‐0.47  30  4774 

The Hague 

Ethnicity  % in academic track  Neighbourhood SES  % weighted  N 

Native‐born  56  1.47  2  1668 

Moroccan   22  ‐1.01  56  454 

Turkish  16  ‐0.87  48  569 

Surinamese/Antillean  33  ‐0.00  30  528 

Other immigrant  43  0.51  24  699 

Total  41  0.47  23  3918 

Utrecht 

Ethnicity  % in academic track  Neighbourhood SES  % weighted  N 

Native‐born  53  1.29  2  1107 

Moroccan   16  ‐0.16  57  419 

Turkish  26  ‐0.11  52  205 

Surinamese/Antillean  34  0.82  20  91 

Other immigrant  56  0.91  21  219 

Total  42  0.79  21  2041 

Source: BRON data. 

   

Page 66: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

40 

Appendix 2‐6: Data scatter, linear (OLS) and LOWESS regression of the percentage of 16 year‐old native Dutch pupils on the percentage of 16 year‐olds of Moroccan descent at zip level in the four major cities in the Netherlands 

 

020

4060

80 %

in a

cade

mic

trac

ks a

mon

g N

ativ

e D

utch

20 40 60 80 100% migrants per ZIP

Page 67: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

41 

3 School switching, a concern for migrant pupils 

Justice and injustice are also called equity and inequity… 

Baruch de Spinoza, 1670 

3.1 Introduction 

Enhancing continuity and stability  in secondary school careers by the reduction 

of switching between schools might, according to the literature reviewed in this chapter, 

be  a  promising  approach  to  speed  up  the  reduction  of  the  current  educational 

achievement  gap  between  native  Dutch  pupils  and  pupils  with  an  immigrant 

background. Switching, however,  seems  to be a  relatively understudied  subject  in  the 

Netherlands.  This  explorative  study  sets  out  to  take  stock  of  patterns  of  switching 

among  different  ethnic  groups,  notably  among  pupils  with  a  Moroccan  immigrant 

background, in three zip‐code areas in the district of Amsterdam‐West, compared to the 

rest of Amsterdam.  

During  1969s  and  1979s,  large  numbers  of  Turkish  and Moroccan  immigrants 

entered  the  Dutch  labour market, mostly  to  be  employed  in  unskilled  work.  In  the 

following decades, most of  these workers were  followed by  their wives  and  children. 

Therefore,  the  vast majority  of  their  children  and  grandchildren,  who  are  currently 

Page 68: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

42 

enrolled  in  secondary  education,  belong  to  the  second‐generation  immigrants12. 

Enrolments  of migrant  pupils  in  the  two  highest  levels  of  secondary  education  (the 

academic  tracks,  specified  further  below)  have  continuously  increased  over  the  past 

decades, but so did enrolment percentages of native‐born pupils. The width of the gap, 

therefore,  remains  substantial:  in 2009, nationally, 45.1% of 16  year‐old native Dutch 

pupils were  enrolled  in  secondary  academic  tracks,  against  20.5%  among pupils  form 

Moroccan  descent  and  20.6%  of  pupils  from  Turkish  descent  (Ministry  of  Education, 

Culture and Science). 

In general, in this research references to a migrant background indicate second‐ 

generation  immigrants; according  to  the definition used by Statistics Netherlands,  this 

means that a person was born in the Netherlands, but has at least one parent who was 

born abroad. The focus group in this study are pupils belonging to the second generation 

Moroccan  immigrants. Two  levels of school organization are distinguished  in  this  text: 

the  individual  school,  on  the  one  hand,  and  the  cluster  level  of  schools  that  share  a 

common governing board, on the other hand. This distinction has been made in order to 

be  able  to measure  all  transfers  between  schools,  as  well  as  school  switching  that 

includes  transfer across  cluster borders. We will also  consider upward and downward 

mobility  (referring  to  moving  to  another  track  within  the  multi‐tracked  system  for 

secondary education in the Netherlands) and its association with switching. Students can 

change  to another  track within  the same school, or do  this after switching  to another 

school. When a pupil moves up to a higher track in the same school, this is not counted 

as switching; switching implies the transfer to another school 

3.1.1 Young people with a migrant background in Amsterdam‐West 

In the three zip‐code areas studied in this chapter,  inhabitants with a Moroccan 

background  form  the  largest  group.  In  the  words  of  Crul  and  Holdaway  (2009): 

                                                       12 In this text the following definitions used by Statistics Netherlands (the National Statistics Office, www.cbs.nl) are used: 

First generation immigrant: a person born abroad with at least one parent born abroad. 

Second generation immigrant: a person born in the Netherlands with at least one parent born abroad. 

Foreign background: a person with at least one parent born abroad. 

Western immigrant: a person originating from a country in Europe (exclusive of Turkey), North America, Oceania, Indonesia or Japan. 

Non‐western immigrant: someone originating from Africa, South America, Asia (exclusive of Indonesia and Japan) or Turkey. 

Page 69: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

43 

"Amsterdam‐West is the 'capital' of Dutch Moroccans in the Netherlands (…) the second 

largest national origin group  in  the  city.  It  is also  the  fastest growing group  (…) more 

than 40% are younger than 19 years old (…)" 

The  achievement  gap  between  native‐born  and  migrant  pupils  in  the 

Netherlands, and more  in general between children from  low SES parents and children 

from more affluent families in many countries all over the world, has become a complex 

and persistent problem. Moreover, the consequences of the gap are serious and costly, 

both for the individual and for society at large, hence the abundance of research into the 

nature  and  possible  solutions  of  this  problem.  As  will  be  further  described  below, 

because  of  this  complexity,  the  causal  effect  of widely  implemented  interventions  to 

narrow  the  gap–  for example, by  reducing  class  size,  increasing per  capita  funding of 

pupils, merit  pay  for  teachers  and  Principals,  intensifying  testing  and making  schools 

more accountable– could not unequivocally be proved. Moreover, research publications 

that actually do describe  causal effects on  children's  chances of  school  success, often 

identify  circumstances  that  are  outside  the  span  of  control  of  schools,  such  as  low‐ 

educated  parents,  or  impoverished  residential  areas  with  an  accumulation  of  socio‐

economic problems; such circumstances,  in most cases, have to be regarded as a given 

by school leaders. As a further illustration of the stubbornness of the problem, the width 

of the achievement gap between native‐born and migrant pupils in secondary academic 

tracks in the Netherlands, apparently so far could not have been reduced sufficiently by 

large  scale  interventions  by  the Dutch  government,  e.g.  extra  funding  for  all  low  SES 

pupils,  equal  funding  of  all  schools  by  the  government,  and  completely  free  school 

choice  (according  to  the  constitutional  law  of  1917).  Poor  neighbourhoods  are  not 

further held back by underfunded schools, since  these are not dependent on  local  tax 

revenues or tuition fees. 

The  underrepresentation  in  secondary  academic  tracks  of  the  two  largest 

migrant groups in the Netherlands– people of Turkish and Moroccan descent– affects a 

growing number of young people: nationally,  in 2008 16% of all  individuals aged 0‐ 20 

years  had  a  migrant  background;  in  the  four  major  cities  (Amsterdam,  Rotterdam, 

Utrecht and The Hague), around 50% of the young people of that age are of immigrant 

descent. Most Turkish and Moroccan families have a weaker socio‐economic position, a 

considerable lower average family income compared to the native Dutch, and live more 

Page 70: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

44 

often below the poverty line (Herweijer, 2009). Their opportunities to fully participate in 

society  in meaningful ways, seem  further challenged,  in  recent years by  the polarizing 

nature of the public debate on a culturally diverse society. According to Schinkel (2008), 

in his critique on this polarizing debate, migrants tend to be regarded as being "outside" 

Dutch society.  

Schools in the major cities in the Netherlands are fully aware of their crucial role 

in providing  the best opportunities  for migrant pupils  through high‐quality education. 

However, as mentioned above, schools are confronted with evidence‐based results from 

academic research  into factors that seem to explain the gap, which are mostly outside 

the  influence of school management and policies. Interestingly, however, research  into 

the effect of instability in school careers, caused by repeated switching of pupils to other 

schools, may offer promising new opportunities: reducing student mobility lies to a large 

extent within the span of control of individual schools– with the exception, of course, of 

changing schools because parents move to another area. While the currently available 

data  for  this  research  cannot  identify  statistical  causal effects,  the  findings may offer 

further practicable insights into the nature and pattern of switching.  

This chapter proceeds to answer the following questions: 

What  is the  frequency of school switching among 16‐22 year‐olds, with a  focus 

on pupils of Moroccan descent, living in three socio‐economically disadvantaged 

zip  areas  in  Amsterdam,  considering  separately  the  transfer  between  single 

schools, and across clustered school boundaries?  

What  is  the  average  frequency  of  school  transfer,  considering  all  secondary 

schools in Amsterdam, measured as the school retention rate in Year 3? 

Is switching associated with upward and downward mobility between secondary 

tracks, comparing the three zips with the rest of Amsterdam? 

 

Previous research is reviewed in Section 3.2. Data and methods are described in 

Section 3.3; results are presented in Section 3.4. Conclusions and suggestions for further 

research  are  presented  in  Section  3.5.  The  results  indicate  that  living  in  the  three 

investigated zip‐code areas and being of Moroccan descent are factors associated with 

an  increase  in  the  frequency of  switching. Switching  is associated with a  reduction of 

Page 71: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

45 

upward mobility,  and  an  increase  in  downward mobility  between  secondary  tracks, 

especially for Moroccan pupils, and more so in the three zip areas. 

3.2 Previous Research 

The achievement gap and switching schools 

The complexity of reducing the achievement gap  is  illustrated by the difficulties 

researchers  encounter  in  their many  attempts  to  prove  the  causal  effect  of  widely 

implemented  interventions  aimed  at  improving  the  educational  outcomes  of 

underrepresented groups. For example, Dobbie and Fryer  (2009),  in their  investigation 

into the effects of the Harlem Children's Zone initiative13 on results in Maths and English 

Learning Arts, could not prove the causal and sustainable effect of investments in school 

quality  alone.  Interestingly,  they  could  prove,  however,  that  improving  the  quality  of 

schools combined with  investments  in  the community–  thus creating more  stable and 

family‐supported school careers– had a positive effect on outcomes.  

Enhancing the stability of schools careers by investigating "school careers as they 

point to the ways in which young adults navigate the system" (Baysu and De Valk, 2012), 

seems to be a promising approach, as the research which  is reviewed here below may 

illustrate:  all  types of  intermediate  school departure  (obviously with  the exception of 

obtaining a diploma), seem to have a negative influence on the width of the gap, and to 

reduce the odds of achieving a successful secondary school completion.  

Hanushek,  Kain  and  Rivkin  (2001)  found  an  overall  negative  effect  of  school 

transfer for virtually all reasons for school changing: positive motives for switching, such 

as moving to a more affluent neighbourhood, or upward mobility to a higher track or a 

better  school, and negative  reasons,  such as downward educational mobility or being 

expelled from school, all turned out to be harmful to student's prospects. Their research 

underlines, however, that the different mobility effects need to be disentangled in order 

to identify the separate effects of specific types of school changing. 

Swanson  and  Schneider  (1999)  identified  three  categories of pupil mobility: 1) 

movers: moving to another residential area, but staying in the same school; 2) changers: 

going to another school, but continuing to live in the same area; and 3) leavers: changing 

both the area of residence and school. Their study shows that all mobility increases the 

                                                       13 www.hcz.org 

Page 72: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

46 

odds of dropout, yet has different effects in the long or short run. The prospects of some 

students may  improve  in  the  long  run,  possibly  because  a  proportion  of  this  group 

moves on to a better quality school or to a higher  level of education;  in the short run, 

however,  the  interruption of  the  school career burdens all  students, and  the  risks are 

higher  for  low  SES  students.  Furthermore,  the  negative  impact  of  student  mobility 

increases with grade and age; after the 10th grade (15‐16 years), a substantial negative 

effect was  found  on  student's  cognitive  growth.  Better  results  in  the  long  run  for  a 

certain proportion of  students  in  the U.S. may,  furthermore, partially be  the  result of 

"Tiebout effects" (1956): the tendency to move residence until the best neighbourhood 

is found, given the family income; however, schools in the U.S. depend partly on locally 

different  tax  revenues,  implying  that more  affluent  neighbourhoods may  have  higher 

quality schools because of better funding. In contrast, in the Netherlands all schools are 

funded equally by the government. 

Rumberger and Larson (1998) were able to prove that transfer and dropout form 

a  continuum  in educational disengagement;  similar  factors predict both,  and dropout 

increases with the number of school changes. The authors identify a personalized school 

climate (where teachers know students well), an academic  inspiring curriculum, and an 

engaging social climate, as  factors that enhance a beneficial connection between pupil 

and school. Research by Murnane (2009) adds yet another interesting argument to this  

debate on the negative effects of school mobility: substantial levels of pupil mobility at 

school level also burdens those pupils who do not switch themselves, because teachers 

are required to "devote scarce instruction time to assessing the skills and knowledge of 

the  new  entrants,  and  socializing  them  to  classroom  norms  of  behaviour."  Especially 

children from poorer families are burdened by more often attending high‐turnover, low‐ 

quality schools  (Hanushek et al., 2001), and tend to move recurrently within the same 

urban  school  cluster  (Kerbow  et  al.  2003);  such  children  are  referred  to  as  "city 

migrants" by Nakagawa, Stafford, Fisher and Matthews (2002). 

Considering the harmful effects of school transfer, the question seems relevant 

whether the "holding capacity" of a school should be considered as a critical measure for 

school  effectiveness,  instead  of  only  assessing  average  test  scores  (Rumberger  and 

Thomas, 2000). This view is further supported by Payne (2010), who identified six factors 

that might  reduce  the  persistence  of  underperformance  in  urban  schools:  instruction 

Page 73: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

47 

time, intellectually ambitious instruction, a strong and focused community of teachers, a 

challenging academic climate  in combination with strong social support  for pupils,  the 

diagnostic ability of teachers and a whole school approach. Notably, continuity of social 

support for pupils, and guidance by means of formative assessments may benefit from 

stability of school careers and long‐term relations with teachers.  

At the individual level, not surprisingly, high absenteeism, misbehaviour, and low 

social  engagement with  school  and  classmates  predict  school  switching  and  dropout. 

The widespread practice, described by Rumberger and Larson (1998), of administratively 

transferring  troublesome  students  to  other  schools, may  worsen  the  prospects  that 

these underperforming students. In the first place, school changing increases the risk of 

these students will dropout, while at the same time a cluster of schools may nonetheless 

end  up with  the  same  number  of  troublesome  pupils;  they were  just moved  around 

within the cluster from one school to the next. 

Kerbow  et  al.  (2003)  summarize  the  factors  that  actually  cause  the  negative 

consequences of non‐promotional  changing of  schools as  follows: 1)  frequent movers 

miss continued exposure to key concepts of  learning and higher‐order skills; 2) moving 

reduces overall the pace of learning; 3) ability grouping may be biased because teachers 

have less time for assessments; 4) the chance to be placed in high ability groups may be 

reduced  and  5)  cooperation with  parents may  be  jeopardized. Migrant  parents who 

were  interviewed  in  the  context  of  the  OECD  review  on  migrant  education  in  the 

Netherlands (Shewbridge, Kim, Wurzburg and Hostens, 2010) expressed the opinion that 

the complex  tracked system  for secondary education  in  the Netherlands puts an even 

greater  emphasis  on  stable  contacts  between  parents  and  school  and  sustained 

cooperation.  

Stability of school careers and cooperation with parents 

School effectiveness– especially  for children  from  low SES parents–  is positively 

correlated  with  tightly  knitted  bonds  between  school,  parents  and  community: 

"connectedness  matters"  could  be  the  dictum  of  Maroulis  and  Gomez  (2008). 

Furthermore,  Margolis  (2010)  writes:  "Teachers  need  to  be  anthropological 

ethnographers  of  students  (…).  They  work  to  understand  students  through  learning 

centred  conversations  that are  sometimes  focused directly on  the  immediate  content 

Page 74: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

48 

being  explored,  but  other  times may  be  centred  around  personal,  social,  or  cultural 

issues that impact individual and group learning over time". 

The point of view that teachers and school need to stay connected to the places 

where pupils live is further supported by Miller Marsh and Turner‐ Vorbeck (2010), who 

argue that the dissonance between the culture at home and that at school– a substantial 

barrier for low SES children– can best be mitigated by strong neighbourhood schools and 

continuous  school  careers.  Finally,  switching  schools,  notably  for migrant  pupils, may 

threaten the sense of security that pupils have experienced by being among classmates 

at their original school, and even increase exposure to hostility at another school (Leigh 

McGarrigle, 2010).  

Focus on school switching 

In the Netherlands, the reduction of early dropout is one of the main themes on 

the  political  agenda,  on  both  the  government  and  the  municipal  level.  Obviously, 

dropout may have severe consequences for young people. Moreover, over the  lifetime 

of  the  individual,  the  costs  for  society may  be  substantial  as  a  result  of  reduced  tax 

revenues, and the costs of social security due to unemployment– that is often related to 

dropout. On the other hand, the returns on investment of continued education, as Levin 

extensively  reports  throughout  his work  (e.g.  Levin,  2011,  Belfield  and  Levin,  2007), 

exceed  all  the  costs of  supportive programmes  to  guide disadvantaged  young people 

towards a diploma. Although over a lifetime the returns on investment in education are 

substantial, in the short run, avoiding dropout and often repeatedly guiding youths back 

to  school  requires  expensive  interventions  due  to  the  time‐consuming  nature  of  the 

work. Non‐promotional  school  changing  appears  to  be  driven  by  comparable  factors 

such  as  early  school  leaving,  and  similarly  predicts  higher  odds  of  later  dropout. 

However,  the reduction of switching could be  far  less expensive  than  the reduction of 

dropout: enhancing the "holding capacity" of schools means, in practice, different ways 

of  organizing  already  existing  student  guidance  and  counselling  practices,  and  to 

agreements  between  schools  in  order  to  avoid  creating  groups  of  "city  migrants" 

(Nakagawa et al., 2002), who  frequently migrate between schools and may eventually 

end up as dropouts after all. 

Page 75: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

49 

In  summary,  the  literature  predicts  disruptive  effects  of  school  switching, 

through the loss of familiar classmates, teachers, and curriculum. Discontinuity of school 

careers  challenges  the  intellectual  growth  of  pupils,  as  measured  by  formative 

assessments. Pupils enrolled  in high  turnover  schools, who do not  switch  themselves, 

lose  time  to  learn  because  teachers  have  to  spend  time  on  taking  on  board  every 

newcomer.  Finally,  switching  reduces  chances  to  upward  mobility  and  school 

completion. These findings have inspired the current study. 

3.3 Data and Methods14 

3.3.1 Single schools and clusters 

Inspired by Kerbow et al.  (2003), we measured  school  switching at  two  levels, 

and used the following definitions: 1) School switching: switching school, irrespective of 

switching within a cluster or across cluster borders; and 2) cluster switching: switching to 

another school in another cluster. The level of a single school and a cooperating school 

cluster  could be distinguished by means of  the BRIN15  code,  the  school  identification 

number.  Every  single  school  has  4‐digit  BRIN  identification  code;  when  schools 

cooperate  in  a  cluster  two  digits  are  added  to  the  individual  school  code.  Hence 

switching  between  single  schools  could  be  compared  to  switching  across  borders  of 

school clusters. In practice, there may be occasional exceptions to the rule of four plus 

two additional digits in the case of clusters, typically caused by the different histories of 

mergers between schools. Therefore the BRIN codes of all schools and school clusters in 

Amsterdam have been visually checked,  in order to be able to trust the frequencies of 

switching we found. While the literature describes different types of more or less formal 

school  clustering,  as  a  rule,  in  the  Netherlands,  clustered  schools  share  a  common 

board, and have a Principal per individual school.  

3.3.2 Migrant students compared to native Dutch peers 

We measured switching at two different points in time during the school career: 

1)  switching  counted  after  exiting  secondary  education  (with  a  diploma  or  early 

dropout);  and  2)  switching measured  at  the  entrance  of  Year  3  in  secondary  school. 

                                                       14  I am profoundly grateful  to Bregje Zwaan, Adam Booij and  Ilja Cornelisz, who supported me with  their  technical 

genius in carrying out Stata analyses. 15 BRIN: Dutch acronym, Basis Registratie Instellingen. 

Page 76: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

50 

Furthermore, we compared switching in the three zip areas with the rest of Amsterdam, 

separate  for  different  ethnicities  and  with  a  focus  on  pupils  of  Moroccan  descent. 

Finally, we measured how switching  is associated with upward and downward mobility 

between  tracks  in  secondary  school.  The  characteristics  of  the  data  sets  will  be 

described  below,  as  well  as  the  choices  that  have  been  made  in  carrying  out 

measurements; Appendix 3‐1 presents a comparative overview of the data sets.  

3.3.3 Three data sets 

Data for the three zip‐codes (DMO data) 

This data set covers  the  following  information: on  the  reference date of 31  July 2009, 

3168 16‐22 year‐olds live in zip area 1061, 1062 and 1063 in the district of Amsterdam‐

West. We  choose  this  age  range,  instead  of  a  cohort,  for  several  reasons. We  are 

interested  in  the  total  frequency  of  switching when  pupils  exit  secondary  education. 

Depending on the track  in secondary education16, pupils exit school typically at age 16, 

17 or 18. We also wanted to include, however, pupils who repeat classes or even leave 

school and come back again. Therefore we decided to  include youths up to 22 years of 

age. Furthermore, a single year cohort would imply rather small numbers of individuals 

when we consider only these three zip areas. Finally, we were  interested  in a "picture" 

of the youths living in these three zip areas, at a single reference date. We acknowledge, 

however, that cohort data offer valuable other information. Therefore, we repeated our 

analyses with a second data set, covering cohort data, as described below.  

The DMO  data  set  covers  both  school‐going  youths  and  those who  have  already  left 

secondary  education,  either  with  or  without  a  diploma.  In  measurements  of  the 

frequency of school transfers, all individuals who were still enrolled in secondary school 

                                                       16 Official translations of the Dutch educational system in the ISCED classification (International Standard Classification of Education by UNESCO, update1997): 

VMBO:  pre‐vocational  secondary  education,  4  years,  ISCED  2,  qualifying  for  senior  secondary  vocational education; 

HAVO:  senior  general  education,  5  years,  grade  1‐3  ISCED  2,  grade  4‐5  ISCED  3,  qualifying  for  higher education; 

VWO:  pre‐university  education,  6  years,  grade  1‐3  ISCED  2,  grade  4‐6  ISCED  3,  qualifying  for  higher education; 

MBO: senior secondary vocational education, level 1 ISCED 2, level 2‐4 ISCED 3, level 4 qualifying for higher education;  

HBO: universities for applied sciences, ISCED 5B; 

WO: research universities, ISCED 5A. 

The completion of ISCED level 3 is the internationally agreed initial (or basic) qualification for the labour market. School leavers without an ISCED 3 qualification are regarded as early school leavers or dropouts. 

Page 77: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

51 

at the reference date were dropped, since this might have biased our measures: these 

pupils were still enrolled in the 4‐, 5‐ and 6‐year programmes of the different secondary 

tracks, and some of them may still switch before they exit secondary school.  

The data set covers the age, gender and ethnicity of every individual; all individuals were 

registered  by  an  anonymous  code.  The  name,  address  and  BRIN  code  (identification 

code) of schools were also registered. Every enrolment and school exit (with or without 

a diploma) was  registered  at  the exact date,  at  the  individual  level. When  individuals 

drop out of school, the Social Development Service registers them  in a special "priority 

file",  to guide pupils back  to school as quickly as possible. Thus we could  follow every 

pupil  into  re‐enrolment or  final dropout.  The  total number of 3168  individuals  in  the 

data set was  taken  into account  for measuring  the ethnic composition of  the zip‐code 

area,  as well  as  the  proportion  enrolled  in  the  highest  academic  tracks  of  secondary 

education, as a proxy for the current achievement gap between native Dutch pupils and 

pupils with a migrant background.  

Several  additional  choices  were  made  to  prepare  the  data  for  analysis:  tracks  in 

secondary education have been  clustered  in  two main  trajectories,  referred  to as  the 

"vocational track" and the "academic track": 1) VMBO, the vocational track (4 years, pre‐

vocational education, four sub‐tracks and some 40 different vocational programmes that 

correspond with  occupations  on  the  job market),  on  the  one  hand;  and  2) HAVO  (5 

years, senior general education) plus VWO (6 years, pre‐university education) combined 

as academic  tracks, on  the other.  Individuals enrolled  in  senior  vocational  institutions 

were dropped;  these  institutions do not belong  to  secondary education  (the  scope of 

this study) and are an advanced vocational level after secondary school. Pupils typically 

leave  elementary  education  at  age  12;  however,  a  limited  number  of  individuals 

between  16‐22  years  of  age  were  still  registered  as  being  enrolled  in  elementary 

education.  These  students  probably  arrived  in  the Netherlands  only  recently,  or may 

experience  severe  learning  disabilities;  these  individuals  were  dropped  from  the 

analysis.  Individuals with a missing BRIN  identification code were also dropped. When 

individuals re‐entered exactly the same school, this was not counted as switching. This 

reduced the number of individuals in all counts of school switching from 3168 to 2653. A 

further  564  individuals  were  dropped  because  they  were  still  enrolled  in  secondary 

schools. Thus 2089 individuals remained for analysis. 

Page 78: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

52 

This tailor‐made data set was kindly made available by the Social Development Service 

of the municipality of Amsterdam (Dutch acronym: DMO).  

Amsterdam school retention rates (DMO data, school year 2009/2010) 

Additionally, as a reference for the frequency of school switching among pupils living in 

the  three  zips,  the  frequency  of  switching measured  at  school  level  (i.e.  the  school's 

retention rate) for all schools  in Amsterdam was also measured. To this end the Social 

Development Service granted the use of data on the retention rate of secondary schools 

in Amsterdam  (51 schools). The school's  retention  rate,  for  the purpose of  this paper, 

has been defined as the percentage of pupils who are still enrolled in Year 3, in the same 

school where they enrolled initially in Year 1.  

Pupils' position in secondary school Year 3 (BRON data), comparing the three zips with the rest of Amsterdam. 

Having started  in 2001, DUO, the government agency that collects all educational data 

linked to the  individual social security number,  is  in the process of constructing a new 

educational  database  called  BRON.  This  database will  eventually  cover  the  complete 

educational history of all  individuals. From 2008 on, all pupils  in elementary education 

are  recorded  in  the  BRON  files,  starting with  pupils  in  the  final  grade  of  elementary 

school in that year; at the time of the current research, this cohort had reached the third 

year of secondary school. The BRON data make  it possible to compare switching  in the 

district  Amsterdam‐West with  the  frequency  of  switching  in  the  rest  of  Amsterdam. 

Importantly,  with  the  BRON  data,  a  first  attempt  could  be  made  to  measure  how 

switching  is  associated  with  upward  and  downward  mobility  between  different 

secondary  tracks,  a  significant  performance  indicator.  The  BRON  data  set  used  for 

measurements in this study contains 4815 pupils who were enrolled in the last grade of 

elementary education  in 2008  (a  summary of data characteristics  is given  in Appendix 

3‐1). Among these, 310 pupils lived in Amsterdam zip areas 1061, 1062 and 1063. For a 

part of  the  total cohort  (2593  individuals, 414 of Moroccan descent)  the data set  lists 

every  school  transfer,  plus  every  transfer  to  another  secondary  track  (upward  or 

downward) at  the  same  school or another  school. Considering  the  three  zips, data on 

track mobility were  available  for  190  individuals  (98  of Moroccan  descent).  The  data 

allowed  for measuring  the  association  between  school  switching  and  track  switching. 

Page 79: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

53 

Importantly, whereas the number of schools attended by one  individual were counted 

to answer Research question 1,  in the case of upward and downward mobility through 

secondary  tracks and  its association with  switching, we used a dummy‐coded variable 

(switching =1, non‐switching = 0).  

Additionally, we used the BRON data to estimate the average retention rate of schools 

attended by pupils living in the three zips. 

3.4 Results 

3.4.1 The frequency of switching in the three zip‐code areas 

To  set  the  scene, Table 3‐1 provides  information on  the ethnic  composition of 

zip‐code 1061, 1062 and 1063 in the district Amsterdam West, based on all 16‐22 year‐

olds  in  the  data  file,  3168  individuals.  As  can  be  seen  from  Table  3‐1,  youths  of 

Moroccan descent form the largest group with a share of 40%. In total 83% of all 16‐22 

year‐olds in the area have an immigrant background. 

Table 3‐1: Composition by ethnicity of all 16‐22 year‐olds living in zip‐code areas 1061, 1062 and 1063 in the district Amsterdam‐West, at the reference date of 31 July 2009  

Ethnicity  Amount  % 

Native Dutch  543  17.1 

Moroccan  1267  40.0 

Turkish  647  20.4 

Surinamese  229  7.2 

Antillean/Aruban  28  0.9 

Other  454  14.3 

Total  3168  100 

Source: DMO 

As a further illustration, the total number of 3168 individuals was also taken into 

account  to measure  enrolments  in  the  academic  tracks  of  secondary  education  as  a 

proportion of each ethnic group; these indicative measurements have been included as 

an approximation of the width of the achievement gap in the three zip areas (Table 3‐2); 

the available data allow for the measurement of the percentage of individuals who were 

registered at least once as being enrolled in the academic secondary track. The share of 

them who obtained a diploma at this  level  is not considered (it was not  in the data file 

for all individuals).  

Page 80: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

54 

Importantly, the percentage of native Dutch  individuals  in academic tracks may 

be overestimated here: native Dutch youngsters, who grew up  in other cities,  tend  to 

come  to  live  in  the area when  they start  their university studies, attracted by cheaper 

housing opportunities. From the data we cannot derive whether these individuals were 

born  in  the area or elsewhere; but since  the name of all schools  is  registered, we can 

identify that their secondary school was outside the city border. Unfortunately, this did 

not allow us  to drop  these  individuals, since we could not be sure whether  they  lived 

outside the city border or travelled a  further distance to school. This may explain why 

the share of them that has been registered at the academic secondary  level,  is actually 

well above the national average (45.1% in 2009). Enrolments of 15.9% among Moroccan 

pupils in the investigated zip area, and 15.2% of Turkish pupils are, however, well below 

the national average (nationally 20.5% and 20.6%, respectively,  in 2009). Note that we 

do not encounter this problem in measurement based on the BRON data (from Table 3‐6 

on); individuals in this data set have not yet reached the age of access to university.  

Table 3‐2: The share of 16‐22 year‐olds by ethnicity, enrolled in academic secondary tracks. Individuals were registered at least once in the data file as being enrolled at this level 

Ethnicity  Total  In academic tracks  % 

Native Dutch  543  317  58.4 

Moroccan  1267  202  15.9 

Turkish  647  98  15.2 

Surinamese  229  35  15.3 

Antillean/Aruban  28  9  32.1 

Other  454  121  26.7 

Total  3168  782  24.7 

 Source: DMO 

Table 3‐3 presents, separately  for  the six distinct ethnic groups,  the number of 

secondary schools that were attended by pupils counted after they had exited secondary 

education.  The  default  situation would  be  that  a  pupil  attends  one  single  secondary 

school, except when parents move to another area. As can be seen from Table 3‐3, there 

are  differences,  notably  in  the  percentages  enrolled  in  only  one  school,  between 

different  groups:  67.0%  of  pupils  with  a  Moroccan  background  attended  only  one 

secondary school; this  is also the case for 72.5% of Turkish students. The total average 

Page 81: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

55 

over all groups of enrolments  in only one school was 69.4%. Considering native Dutch 

students, 77.4% did not switch.  

Table 3‐3: The number of secondary schools attended by 16‐22 year‐olds, by ethnicity 

No. of schools  

Ethnicity 

1  %  2 % 3 % 4 %  Total  % 

Native Dutch  151  77.4  36  18.5  7  3.6  1  0.5  195  100 

Moroccan  647  67.0  262  27.1  53  5.5  4  0.4  966  100 

Turkish  346  72.5  102  21.4  27  5.7  2  0.4  477  100 

Surinamese  111  64.5  51  29.7  8  4.7  2  1.2  172  100 

Antillean/Aruban  13  76.5  3  17.7  0  0.0  1  5.9  17  100 

Other  181  69.1  62  23.7  15  5.7  4  1.5  262  100 

Total  1449  69.4  516  24.7  110  5.3  14  0.7  2089  100 

Source: DMO. 

Table 3‐4: The number of secondary schools attended by 16‐22 year‐olds, by ethnicity, measuring school switching that includes cluster switching 

No. of schools  Ethnicity 

1  %  2 % 3 % 4 %  Total  % 

Native Dutch  165  84.6  26  13.3  4  2.1  0  0.0  195  100 

Moroccan  765  79.2  175  18.1  25  2.6  1  0.1  966  100 

Turkish  408  85.5  62  13.0  7  1.5  0  0.0  477  100 

Surinamese  133  77.3  35  20.4  3  1.7  1  0.6  172  100 

Antillean/Aruban  35  88.2  2  11.8  0  0.0  0  0.0  17  100 

Other  207  79.0  43  16.4  12  4.6  0  0.0  262  100 

Total  1693  81.0  343  16.4  51  2.4  2  0.1  2089  100 

Source: DMO. 

The measurement of the number of schools attended was repeated at the higher 

aggregate  level  of  a  school  cluster,  in  order  to  compare  the  proportion  of  switching 

between  all  separate  schools,  to  switching  across  borders  of  clustered  groups  of 

cooperating  schools  (Table  3‐4). Measuring  at  the higher  aggregate  level of  clustered 

schools  reduces  the  number  of  switches,  indicating  that  pupils  also migrate within  a 

cluster  of  schools.  Interestingly,  considering  school  switching  that  includes  cluster 

switching, the percentage of pupils of Turkish descent who do not switch (85.5%) is close 

to the percentage among native Dutch pupils (84.6%). Youngsters of Moroccan descent, 

however,  switch more  often,  both  in  counts  at  the  separate  school  level  and  at  the 

cluster level. The substantial difference of 12.2 percentage points considering Moroccan 

Page 82: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

56 

pupils who attend only one  school, between counts at  the  level of  single  schools and 

clusters, indicates they move rather frequently within a cooperating school cluster. 

Table 3‐5 presents 16‐22 year‐old Moroccan students on academic track  levels, 

who have exited  secondary  school  (with or without a diploma), or have moved on  to 

another  school. Three  types of outflow  status are  specified: 1) departure  from  school 

with  a  diploma;  2)  dropout  without  a  diploma;  and  3)  switch  to  another  school. 

Considering pupils who exit  their  first  school of enrolment, a majority of 87.4%  leave 

with  a diploma;  10.3%  switch  and  continue  in  another  school. However, when pupils 

exit, for example, their third school, only 39.1% do so with a diploma. When exiting the 

third school, no less than 43.5% drop out, and do not re‐enrol again in a next school.  

Table 3‐5: Categories of school‐exit among pupils of Moroccan descent, who were enrolled in academic secondary tracks (reference date 31 July 2009) 

No. of schools  Cause 

1 %  2  % 3 % 4 % 5 % 9 %  Total 

Drop out  2  2.3  18  24.7  10  43.5  6  50.0  4  66.7  1  100  41 

Other school  9  10.3  4  5.5  4  17.4  3  25.0  1  16.7  0  0.0  21 

Diploma  76  87.4  51  69.9  9  39.1  3  25.0  1  16.7  0  0.0  140 

Total  87  100  73  100  23  100  12  100  6  100  1  100  202 

Source: DMO. 

3.4.2 The retention rate of schools  

To  illustrate  the occurrence of  switching at  school  level, Figure 3‐1 depicts  the 

retention  rate  for all schools  in Amsterdam  (51 schools). The  retention  rate  is defined 

here as the percentage of pupils who enrolled in a secondary school in Year 1, who are 

still enrolled in the same school in Year 3. Separate schools are observed here, clustering 

is not  considered. While  the average  retention  rate  for  schools  in Amsterdam  is 85%, 

schools differ substantially within a range of 55% to 98% retention  in Year 3. Since the 

schools  in  the municipal  data  file  have  anonymous  codes,  unfortunately, we  cannot 

derive directly  from  these data whether pupils  living  in  the  three  zips attend a  school 

with  a  higher  or  lower  retention  rate.  We  could,  however,  estimate  the  average 

retention rate of schools attended by pupils living in the three zip areas, by means of the 

BRON data set. Therefore we considered all Amsterdam schools where pupils  living  in 

the  three  zip  area  have  enrolled  (in  the  Netherlands  88.7%  of  pupils  in  secondary 

Page 83: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

57 

education are enrolled  in another school than the one closest to their home, as will be 

described in Chapter 5 of this thesis), and derived the retention rate of these schools by 

using the total number of  individuals enrolled  in a school, according to our BRON data 

set. We found that the average retention rate of the schools where pupils  living  in the 

three zip areas were enrolled is 71.2%, a very low retention rate considering Figure 3‐1. 

When we distinguish between the retention rates of schools attended by switchers (≥ 1 

switch)  and  non‐switchers  living  in  the  zip  areas  who  were  still  enrolled  in  Year  3: 

switchers  attend  a  school  with  a mean  retention  rate  of  50.1%,  and  non‐switchers 

attend a school with a retention rate of 79.7%. 

Figure 3‐1: The retention rate in ranking order of all secondary schools (considering separate schools, not clusters) in Amsterdam, defined as the percentage of pupils who entered in Year 1, and are still enrolled in Year 3 

 

Source: DMO; school year 2009/2010. 

3.4.3 Switching and upward and downward mobility between track levels  

Observations in the BRON data have not been collected for long enough to allow 

counting  all  switches  during  secondary  school. We will  use  these  data  to  distinguish 

switchers  (at  least one school switch) and non‐switchers;  for the cohort  in the highest 

grade of elementary school in 2008, we can only measure switching behaviour up to the 

Year 3 in secondary education 

50%

55%

60%

65%

70%

75%

80%

85%

90%

95%

100%

0 10 20 30 40 50 60

Series1

Page 84: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

58 

Table 3‐6 presents school switching in Amsterdam during the first three years of 

secondary  school,  as  a dichotomous  variable. As  a  comparison,  Table  3‐7  depicts  the 

same measurement for zip areas 1061, 1062 and 1063. In the three zips areas, switching 

is  less frequent among native Dutch pupils and more frequent among  immigrant pupils 

(but, except  for Moroccan pupils, the number of observations  in the three zip areas  is 

very small). All over,  the  frequency of switching  is higher  in  the  three zip areas  in  the 

district Amsterdam‐West, than overall in Amsterdam.  

Table 3‐6: School switching among pupils in Amsterdam enrolled in the Year 3 of secondary education in 2011 

Switch Ethnicity 

None  %  ≥1 % Total %

Dutch  1422  81.7  319  18.3  1741  100 

Surinamese/Antillean  581  76.7  177  23.4  758  100 

Turkish  337  73.3  123  26.7  460  100 

Moroccan  577  70.4  243  29.6  820  100 

Other  774  74.7  262  25.3  1036  100 

Total  3691  76.7  1124  23.3  4815  100 

Source: BRON data. 

Table 3‐7: School switching among pupils living in zip‐codes 1061, 1062 and 1063, enrolled in the Year 3 of secondary education in 2011  

Switch Ethnicity 

None  %  ≥1 % Total %

Dutch  20  90.9  2  9.1  22  100 

Surinamese/Antillean  13  72.2  5  27.8  18  100 

Turkish  58  72.5  22  27.5  80  100 

Moroccan  106  68.4  49  31.6  155  100 

Other  24  68.6  11  31.4  35  100 

Total  221  71.3  89  28.7  310  100 

Source: BRON data. 

In Tables 3‐8 and 3‐9, the above counts have been repeated for switching across 

school cluster borders. 

Differences between ethnicities are much  smaller now, except between native 

Dutch  and  immigrant  students.  The  frequency  of  cluster  switching  is  roughly  2/3  of 

school switching,  indicating that 1/3 of school switching occurs within clusters, and 2/3 

crosses  cluster  boundaries:  the  majority  of  school  switching  occurs  across  cluster 

borders.  But,  as  is  the  case  for  school  switching,  cluster  switching  among  immigrant 

Page 85: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

59 

students  is more  frequent  in  the  three  zips  than  in  the  rest  of Amsterdam, whereas 

among native Dutch students  it  is  lower  in  the  three zips  (but numbers are very small 

here). 

Table 3‐8: Cluster switching among pupils in Amsterdam enrolled in  Year 3 of secondary education in 2011 

Switch Ethnicity 

None  % ≥1 % Total %

Dutch  1520  87.3  221  12.7  1741  100 

Surinamese/Antillean  654  86.3  104  13.7  758  100 

Turkish  386  83.9  74  16.1  460  100 

Moroccan  674  82.2  146  17.8  820  100 

Other  880  84.9  156  15.1  1036  100 

Total  4114  85.4  701  14.6  4815  100 

Source: BRON data. 

Table 3‐9: Cluster switching among pupils living in zip‐codes 1061, 1062 and 1063 enrolled in Yea 3r of secondary education in 2011 

Switch Ethnicity 

None  % ≥1 % Total %

Dutch  21  95.5  1  4.6  22  100 

Surinamese/Antillean  14  77.8  4  22.2  18  100 

Turkish  63  78.8  17  21.3  80  100 

Moroccan  123  79.4  32  20.7  155  100 

Other  28  80.0  7  20.0  35  100 

Total  249  80.3  61  19.7  310  100 

Source: BRON data. 

In the tables below, we take a closer look at the nature of pupil mobility: pupils 

can move up  to a higher secondary  track, move down  to a  lower  track, or stay at  the 

same level; we considered all steps of upward and downward mobility, from vocational 

(VMBO)  to  the  senior  general  track  (HAVO), and  from  the  latter  to  the pre‐university 

level  (VWO). Moving  between  tracks  can  occur  both  for  pupils  staying  in  the  same 

school, and for those who enrol in another school.  

We  use  another  subset  for measurements  now,  and  consider  only  individuals 

who  changed  track:  2593  individuals  in  Amsterdam,  where  414  are  of  Moroccan 

descent;  190  individuals  in  the  three  zips,  where  98  are  of  Moroccan  descent. 

Interestingly, the vast majority of pupils in Amsterdam stay in the same track, also after 

switching (Table 3‐10). Upward mobility occurs more often when pupils stay in the same 

Page 86: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

60 

school  (22.7%)  than among pupils who  transfer  to another school  (9%). The  reverse  is 

the  case  for downward mobility: 11% downward mobility when  changing  track  in  the 

same  school,  against  17.3%  among  those who  switch  to  another  school.  Table  3‐11 

presents  similar measurements,  considering  the  three  zip  areas. Correspondingly,  the 

vast majority of pupils continue at  the same  level,  including after switching. However, 

both after switching and when staying  in the same school, upward mobility  is  lower  in 

comparison  with  the  rest  of  Amsterdam.  Furthermore,  switching  is  more  strongly 

associated with downward mobility in the three zips. 

Table 3‐10: Comparison of track level at entry and in Year 3 of secondary school, separate for school switchers and non‐switchers, among pupils in Amsterdam 

Switch Level 

None  %  ≥1  % Total %

Level(s) down  223  11.0  98  17.3  321  12.4 

Same level  1344  66.3  417  73.7  1761  67.9 

Level (s) up  460  22.7  51  9.0  511  19.7 

Total  2027  100  566  100  2593  100 

Source: BRON data. 

Table 3‐11: Comparison of the track level at entry‐ and in Year 3 of secondary school, separate for school switchers and non‐switchers, among pupils in zip‐code areas 1061, 1062 and 1063 

Switch Level 

None  %  ≥1  % Total %

Level(s) down  14  9.8  9  19.2  23  12.1 

Same level  108  75.5  36  76.6  144  75.8 

Level(s) up  21  14.7  2  4.3  23  12.1 

Total  143  100  47  100  190  100 

Source: BRON data. 

Finally, Tables 3‐12 and 3‐13 compare the mobility between tracks among pupils 

specifically with a Moroccan background,  in Amsterdam  (Table 3‐12) and  in  the  three 

zips  (Table  3‐13).  19.5% of pupils of Moroccan descent move up  to  a higher  track  in 

Amsterdam when  they  stay  in  the  same  school;  this  is only  the  case  for  7.6%  of  the 

switchers, who also have a higher rate of downward mobility  (13.2%).  In the three zip 

areas  in  the  district  Amsterdam‐West, Moroccan  youths  staying  in  the  same  school 

realize an upward mobility of 17.1% when they stay in the same school (caution because 

of small numbers). 

Page 87: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

61 

Table 3‐12: Comparison of track level at entry‐ and in Year 3 of secondary school, separately for school switchers and non‐switchers, among pupils of Moroccan descent in Amsterdam. 

Switch Level 

None  %  ≥1 % Total %

Level(s) down  22  7.1  14 13.2 36 8.7

Same level  226  73.4  84 79.3 310 74.9

Level (s) up  60  19.5  8 7.6 68 16.4

Total  308  100  106 100 414 100

Source: BRON data. 

Table 3‐13: Comparison of track level at entry‐ and in Year 3 of secondary school, separately for school switchers and non‐switchers, among pupils of Moroccan descent in zip‐code areas 1061, 1062 and 1063 

Switch Level 

None  %  ≥1 % Total %

Level(s) down  5  7.2  3  10.7  8  8.2 

Same level  53  75.7  23  82.1  76  77.6 

Level (s) up  12  17.1  2  7.1  14  14.3 

Total  70  100  28  100  98  100 

Source: BRON data.  

In summary, more pupils switch in the three zip‐code areas 1061, 1062 and 1063, 

than in the rest of Amsterdam. Furthermore, comparing all ethnic groups, the frequency 

of  switching  is  substantially  higher  among  migrant  students.  Notably  for  Moroccan 

students living in the three zip areas, the incidence of switching is comparatively high at 

31.6% (Table 3‐7). Differentiating switching further into upward and downward mobility 

through  secondary  tracks demonstrates  that upward mobility  is higher  for  all  groups, 

when they stay in the same school. 17.1% of pupils of Moroccan descent in the three zip 

areas, move up to a higher track when staying in the same school (Table 3‐13). 

3.5 Conclusions and policy implications 

The main data set analysed here  is the DMO data file, kindly made available by 

the Social Development Service of the municipality of Amsterdam. This data set covers 

the detailed school history of youth living in the three studied zip areas. This local data 

set  has  been  an  important  source  of  information  about  key  policy  issues  for  the 

municipality of Amsterdam, notably the reduction of early dropout, and is, furthermore, 

far more detailed  than national data  sets. However,  its architecture makes  the BRON 

data much more accessible for research. Ideally, the architecture of the national data set 

Page 88: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

62 

should  be  compatible  with  the municipal  data  set,  in  order  to  support  local  policy 

makers in translating more general national educational analyses to local practice which 

fit  every  school's  specific  context. While  currently  the  two  data  sets we  used  have  a 

largely different architecture, we hope that our comparison of the data sets may support 

further  initiatives to make the data compatible, a promising endeavour considering the 

fascinating opportunities the BRON data will offer  in the near  future, when we will be 

able  to  follow  the  2008  cohort  from  elementary  school  until  tertiary  education,  and, 

finally,  on  the  labour market.  By  then,  for  example, we will  be  able  to  estimate  the 

causal effects of school switching on school success.  

3.5.1 The three zip areas 

Overall, a larger share of Dutch pupils in comparison with migrant pupils, do not 

switch.  The  share  of  pupils  with  a Moroccan  background  who  do  not  switch  is  the 

lowest, at 67%  (Table 3‐3). Whereas 22% of native Dutch pupils switch school, 33% of 

Moroccan pupils  switch. Among both  groups,  the  share of  school  switchers who  also 

switch cluster is about 2/3: 21% of Moroccan pupils switch cluster, whereas only 15% of 

their native Dutch peers switch cluster (Table 3‐3 and Table 3‐4). These findings suggest 

that  Moroccan  migrant  students  may  be  extra‐burdened  by  the  assumedly  serious 

disruptive  effects  of  switching,  as  the  literature  predicts.  Considering  that  the 

achievement gap  is most wide  in the case of pupils of Moroccan descent, especially  in 

their  case,  agreements  between  schools  to  reduce  switching,  or,  when  switching  is 

unavoidable,  to  securely guide  the  transfer, may offer promising opportunities, within 

the  span  of  control  of  schools,  at  relatively  low  costs.  Clustered  schools  have  the 

advantage  that  they  can  make  agreements  between  their  own  schools  to  reduce 

switching.  Although  the  currently  available  data  do  not  yet  allow  for  demonstrating 

causal evidence,  the  association between obtaining  a diploma  at  academic  secondary 

levels and staying in the same school during the whole school career, convincingly points 

in  the  same  direction  as  the  causal  evidence  presented  in  the  reviewed  literature  in 

Section  3.2  (e.g.  Hanushek  et  al.,  2009).  Considering  these  Moroccan  youths  who 

remained  in  one  school,  87.4%  obtained  a  diploma  on  the  academic  secondary  level 

(Table 3‐5). Those who exited a second school already showed a dropout percentage of 

24.7%. 

Page 89: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

63 

3.5.2 A hidden problem? 

Remarkably, the vast majority of switchers continue in the same track in another 

school. This raises the question why they did switch. This finding may refer to a specific 

problem;  pupil movements  between  schools may  partly  be  the  result  of  disciplinary 

school policies, specifically expulsion policies. Both the municipality of Amsterdam and 

the Inspectorate of Education are well aware of increased risks of later school dropout, 

when a pupil  is expelled  from school. Therefore  the school has a duty  to  report every 

pupil  that  has  been  expelled. However,  sending  a  pupil  to  another  school  under  the 

same board within a cluster may de jure not imply that the pupil has been expelled, and 

the school may not  report  these  transfers  to  the  Inspectorate.  In  this way  involuntary 

departure may  be  disguised  as  voluntary  departure‐  and  behavioural  problems  that 

induced  change  to  another  school  may  persist.  To  estimate  the  occurrence  of  not 

reporting school departure within clusters would be a question for further research.  

The  substantial  differences  in  retention  rates  of  all  secondary  schools  in 

Amsterdam  (Figure  3‐1), which  range  from  roughly  55%  to  95%  retention  (measured 

between  Years  1  and  3),  illustrate  that  indeed  the  frequencies  of  switching  do  differ 

substantially, which leaves ample room for improvement. 

3.5.3 Youth  in  the  three  zips  might  benefit  from  agreements  between  schools  to reduce switching 

The analyses based on the national BRON data confirm that pupils with a migrant 

background switch more often than native Dutch students, and that switching is higher 

still among migrant pupils living in the three studied zip areas.  

The BRON data made it possible, furthermore, to carry out a first exploration of 

possible  relations  between  switching  and  educational  outcomes.  Considering  the 

tracked system of secondary education in the Netherlands, switching especially appears 

to reduce upward mobility to a higher secondary track: In Amsterdam 22.7% of all pupils 

move up to a higher track; this is, however, only the case for 9% among switchers, while 

in  their  case  downward  mobility  is  17.3%  (Table  3‐10).  These  differences  between 

switchers and non‐switchers are far more articulate  in the three zip areas (Table 3‐11): 

among switchers only 4.3% move up to a higher track. Furthermore, looking specifically 

at students of Moroccan descent (total of Amsterdam) who did not switch until Year 3 of 

secondary school, a remarkable 19.5% has improved their track level; even in the three 

Page 90: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

64 

zip areas, this is still 17.1%. Interestingly, Van de Werfhorst and Hofstede (2007) in their 

study on  relative  risk aversion  (in  this  case  the avoidance of downward mobility)  find 

that it is rather concerns with mobility (i.e. the risk involved in moving to a higher track) 

than with social class that are associated with ambitions to progress to higher  levels of 

schooling.   However, switching substantially reduces the upward mobility of Moroccan 

youths  to 7.6%  in Amsterdam  and 7.1%  in  the  three  zip‐code  areas  (Tables 3‐12  and 

3‐13). Upward mobility potentially may have a substantial positive effect on the closing 

of the achievement gap, since migrant pupils more often may need some extra time to 

move  up  to  their  level  of  abilities.  Staying  in  the  same  school,  the  current  research 

suggests, improves the chances of upward mobility. The study by Van de Werfhorst and 

Hofstede (ibid.) seems to support this finding, and, moreover, staying in the same school 

when moving to a higher track, may reduce the experienced risk of mobility, notably for 

migrant pupils. 

3.5.4 Schools offering all tracks may be favourable for migrant students 

As Crul and Schneider describe  in their comparative study on school success of 

pupils with a Turkish background in Germany and in the Netherlands (2009), the tracked 

Dutch  system offers opportunities  to  follow  a  longer  route  through  successive higher 

levels  of  secondary  education,  eventually  allowing  access  to  higher  education. 

Stimulating  upward  mobility  within  the  same  familiar  school,  which  offers  a  broad 

choice  of  tracks, may  be  in  favour  of  low  SES  students,  since  this may minimize  the 

disruptive  effect  of  school  transfer  for  students.  This  assumption  would  counter, 

however,  the  current  trend  in  the  Netherlands  to  organize  vocational  and  academic 

secondary tracks in separate schools, and also even diversify further between academic 

tracks.  

3.5.5 Reducing switching may decrease costs 

Finally, active policies to reduce switching might also be worthwhile considering 

in terms of cost‐effectiveness. As has been mentioned earlier in this chapter, policies to 

reduce dropout  tend  to be  very  costly, because of  the  labour  intensive,  complex and 

tailor made nature of such strategies. Dropout may be the final result of problems that 

worsened  over  a  long  period  of  time.  The  literature  referred  to  in  Section  3.2  offers 

convincing arguments for regarding school changing as a predictor for later dropout, and 

Page 91: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

3 School switching, a concern for migrant pupils 

65 

the findings in the current paper suggest the same tendency: dropout increases with the 

number of schools attended. Presumably, the costs of the reduction of switching may be 

considerably  lower than the repair of dropout, since such policies can be agreed upon 

between  schools and mainly  require  transparency  in  the  reasons  for student mobility; 

clustered  schools  in  particular  have  the  opportunity  to make  agreements  to  reduce 

switching between their own schools. 

In  Summary, when  pupils  of Moroccan  descent  stay  in  the  same  school while 

enrolled in the academic track in the three zips, 87.4% obtains a diploma; already when 

they have been enrolled in two schools (one switch), the percentage of diplomas drops 

to  69.9%,  and  dropout  is  24.7%  (5.5%  transfer  to  yet  another  school).  These  last 

percentages  are  especially worrying, because here we  consider Moroccan pupils who 

qualified  for  the  academic  secondary  level:  this  dropout  directly  widens  the 

achievement gap, although we cannot know from our data which share eventually takes 

up their studies again after the age of 22. Furthermore, a first exploration of the effect 

of  switching,  in  terms  of  upward  or  downward  mobility  between  secondary  tracks 

(measured  in  Year  3)  resulted  in  substantial  differences:  upward  mobility  among 

Moroccan pupils who do not switch, is 19.5% in Amsterdam and 17.1% in the three zips, 

almost 1 in 5 pupils: these track improvements may add in important ways to the closing 

of the achievement gap. However, among switchers, upward mobility decreases to 7.6% 

and 7.1%  (three  zips)  respectively. As a consequence, agreements between  schools  in 

Amsterdam to reduce switching may be expected to have positive results on the closing 

of the gap. We hope that insights into the occurrence and effects of switching described 

in our  study may motivate  schools  to  reduce  switching,  and  to be  transparent  about 

expelling pupils as a disciplinary measure, although this may be a more complex issue. 

 Interesting  opportunities  for  further  research will  arise,  once  the  BRON  data 

cover  the exit‐exam  results  in  secondary education of  the cohort  that was enrolled  in 

the final year of elementary school in 2008. Hopefully this paper may add to convincing 

the Ministry of Education, Culture and Science of the  importance of also making these 

high‐quality data available for scientific research in the future.  

   

Page 92: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

66 

3.6 Appendix 

Appendix 3‐1: Comparison of the two main data sets. 

  DMO data set Amsterdam municipal data 

BRON data set National educational data 

# Individuals  3168  4815

Characteristics   All 16‐22 year‐olds living in zip‐code areas 1061, 1062 and 1063 in district Amsterdam‐West. Exact date of all school transfers, exact BRIN code of all schools, plus the BRIN code of clusters of schools. Ethnicity of all pupils, plus track level in secondary school and exit status when leaving secondary education. 

The cohort, living in Amsterdam, which was enrolled in the last grade in elementary school in 2008. Postal code of the residence of all pupils; BRIN code of all schools on the school and cluster level. Ethnicity, home address and track level of all pupils. 

Reference date  July 2009  The cohort enrolled in the last grade of elementary school in 2008, living in Amsterdam. The data were updated to 2011, when pupils typically were enrolled in Year 3 of secondary school. 

Selection of sample 

Kept: all pupils who had left secondary school at the reference date, with or without a diploma, whose complete history of school attendance plus school BRIN code was registered: 2089 individuals. The total number of individuals (3168) was used to indicate neighbourhood composition and the educational achievement gap between native Dutch and second generation immigrant pupils.  

Kept: all pupils whose complete school transfer history was listed in the data set; 2593 individuals in Amsterdam, 190 of them living in the three zip areas 1061, 1062, 1063.  

Time of measurement 

After exit from secondary school In Year 3 of secondary school 

Switching variable 

Count of the number of schools in which a pupil has been enrolled. 

Dummy‐coded variable (switching =1, non‐switching=0). No distinction between once or repeated switching. 

The  third data set  (DMO) covers  the percentage, at school  level, of pupils who 

are  still  enrolled  in  the  same  school  in  Year  3,  as  the  school  they  entered  in  Year  1; 

reference date: school year 2009/2010. 

Page 93: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

67 

4 Patterns in secondary school selection in 

the context of unlimited choice 

I should also add further that this liberty (i.e. the liberty of 

judgment) is absolutely essential to the advancement of the arts 

and sciences; for they can be cultivated only by those with a free 

and unfettered judgment. 

Baruch de Spinoza, 1670  

4.1 Introduction 

Considering  the  stubborn  problem  of  unequal  access  to  higher  levels  of 

education  for  different  socio‐economic  groups,  offering  broader  options  for  school 

choice  is a policy measure  that generally  raises high expectations.  Interestingly,  in  the 

Netherlands, free school choice has been practically unlimited for almost a century, by 

the  constitutional  law  of  1917.  This  law  also  gives  parents  the  right  to  found  a  new 

school  according  to,  for  example,  a  specific  religious  or  pedagogical  concept,  on  the 

condition  that  a  sufficient  number  of  pupils  will  attend  the  new  school,  and  no 

comparable  school  is available within  reasonable distance.  In practice, because of  the 

density of schools  in the Netherlands, the constitutional right to found a new school  is 

nowadays  rarely  exercised.  Free  school  choice  is, moreover,  guaranteed  because  all 

                                                        This chapter is based on: van Welie, E.A.A.M., Hartog, J. and Cornelisz, I. (2013). Forthcoming.  

Page 94: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

68 

schools are funded equally by the state, both schools with a board based on private law 

or  those with a board based on public  law. Additionally,  in elementary and secondary 

schools no tuition fee is required, except for a contribution for certain activities outside 

the curriculum, like school festivities. This contribution is, however, not obligatory; when 

parents cannot afford to pay, the child is still entitled to participate. Finally, schools are 

independent of local taxes (since all funding comes from the Ministry of Education), and 

all  children  in  elementary  education  growing  up  in  low  SES  families  are  additionally 

funded  through  a  voucher  system,  the money  following  the  pupil.  Secondary  schools 

receive extra  funding when  the  student population  crosses a  threshold percentage of 

pupils  who  live  in  low  SES  neighbourhoods  (the  threshold  percentage  is  lower  for 

vocational tracks and higher for academic tracks). This results  in schools with relatively 

more  low  SES pupils actually  receiving more  funding  than  schools with more  children 

from affluent parents. In summary, the long history of free school choice, equal funding 

by  the  state,  no  tuition  fees,  and  extra  funding  for  low  SES  pupils,  offers  interesting 

opportunities to study possible differences  in school choice between distinct groups  in 

society. 

In the current chapter we compare patterns of school choice among native Dutch 

youth  and  youth with  a migrant  family  history.  In most  cases,  these migrant  pupils' 

grandfathers came to the Netherlands as unskilled workers in the 1969s and 1970s. The 

average level of education of the children and grandchildren of these labour immigrants 

has  shown  a  steady  increase  over  the  past  decades.  However,  considering  the 

importance of attainment  levels  that  fit pupils' capacities– both  for  the  individual and 

for society at large– we still need a deeper understanding of the actual choices migrant 

pupils make,  in  order  to  develop  policies  that may  accelerate  the  process  of  finding 

equal access to high  levels of schooling. According to the OECD17 (2011), a knowledge‐

based  economy may  lose  part  of  the  population's  potential  to  expand  scientific  and 

cultural knowledge, and may receive lower tax revenues over the lifetime of individuals, 

when substantial numbers of citizens remain under‐schooled.  

In this chapter we focus on secondary schools, and diversify between secondary 

vocational tracks (VMBO) on the one hand, and two secondary academic  levels (HAVO 

                                                       17 OECD: Organization for Economic Cooperation and Development. 

Page 95: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

69 

and VWO), on  the other18.  In  the Dutch educational system VMBO  is  the  lowest  level, 

and  VWO  the  highest  level.  The  achievement  gap  in  Dutch  secondary  education  is 

typically described as the lower enrolment levels in academic tracks among youth with a 

migrant  background,  as  compared  to  their  native Dutch  peers.  Both  academic  tracks 

offer access to higher education;  in the text we refer to these two  levels combined as 

the academic level; in some tables we diversify between HAVO and VWO.  

In Europe and the U.S. alike, the debate on raising educational opportunities for 

underrepresented  groups  in  the  higher  strata  of  education  seems  to  concentrate  on 

extending school choice, assuming  that  this will enhance  the desegregation of schools 

along  the  lines  of  race,  ethnicity,  or  social  status;  the  underlying  assumption  for 

desegregation  policies  in  general,  is  that  low  SES  children  may  benefit  from  peers 

belonging to other social strata. 

Interestingly,  geographical  distance  measurements  offer  opportunities  to 

disentangle  the  effects  of  school  choice,  school  composition,  neighbourhood 

composition, school density, and urbanicity19. Using  the  right  to select a school  freely, 

which  may  result  in  choosing  another  school  than  the  one  closest  to  the  pupils' 

residence, unavoidably involves the extra effort of a longer distance to travel to school. 

In the present chapter, we use measurements of the distance from home to the chosen 

secondary  school,  relative  to  the nearest available  school, controlling  for  indicators at 

the  individual, neighbourhood and school  level,  in order to be able to analyse whether 

and how patterns of school choice differ between ethnic groups.  

Analysing  the effect of secondary school choice  is complicated by  the  fact  that 

measuring school success– as the combined effect of individual, neighbourhood, school 

quality,  and  educational  system  variables–  requires  data  on  prior  achievement  in 

elementary  education  to  start  with.  Interestingly,  in  2001,  the  Dutch  Ministry  of 

                                                       18 Tracks in the Dutch system for secondary education, plus the ISCED translation (International Standard Classification  of Education by UNESCO, update1997): 

VMBO: pre‐vocational secondary education, 4 years, ISCED 2, qualifying for senior secondary vocational education. 

HAVO: senior general education, 5 years, grade 1‐3 ISCED 2, grade 4‐5 ISCED 3, qualifying for higher education (professional universities).  

VWO: pre‐university education, 6 years, grade 1‐3 ISCED 2, grade 4‐6 ISCED 3, qualifying for higher education (research universities). 

19 Urbanicity: a measure for the intensity of human activity in a given area, based on the number of addresses per km² (definition Statistics Netherlands). 

Page 96: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

70 

Education,  Culture  and  Science  started  to  create  a  longitudinal  educational  data  set 

based on  individual  social  security numbers,  covering  all  students  and  their  complete 

educational history.  In 2008, all pupils  in  the  final year of elementary education were 

recorded  for  the  first  time  in  this  new  database  (students  in  other  strata  in  the 

educational  system  had  preceded  them).  In  the  coming  years, when  the  total  school 

history  of  all  individuals will  be  recorded,  performance  in  elementary  education  can 

actually be  linked  to  the exit‐exam outcomes of secondary education at  the  individual 

level20. At  the  time of our  research,  the  cohort  that was enrolled  in  the  final  year of 

elementary education  in 2008 has  reached  the Year 3  in  secondary  school. Therefore, 

we  could  relate  prior  achievement  in  elementary  education  to  achievements  at  the 

entrance of the third secondary year. 

Among  policy  makers  and  researchers  alike,  positive  effects  of  equal 

opportunities  in  education  are  expected  from:  1)  free  school  choice;  2)  equal 

(government)  funding of schools; 3)  independence of  local  taxes; and 4) extra  funding 

for  low  SES  pupils.  These  four  conditions  have  all  been  met  system‐wide  in  Dutch 

schools. Although the outcomes of the Dutch system of education rank among the top‐

ten  in  international comparisons  (PISA, 2009),  the pattern of unequal access  to higher 

levels of schooling among low SES youth per se, resembles patterns in countries without 

the four above mentioned conditions.  

While we use distance to the preferred non‐nearest school as a measure for the 

selectivity of choice, we acknowledge that choosing the proximity allocation may also be 

based on  selective choice; however,  in  this case we do not know whether choice was 

based  on  qualitative  considerations,  or whether  parents  and  pupils  simply  chose  the 

nearest  school because  it was close  to  their home. Both high SES and  low SES pupils, 

notably  youth  with  a migrant  background, may  have  reasons  to  choose  the  school 

closest to the residence‐ or prefer another school, albeit these reasons may be different 

between both groups. Obviously, choosing  the nearest  school  takes  the  least effort  in 

terms  of  travelling  time  and  costs.  Since  there  is  no  school  bus  system  in  the 

                                                       20 We are indebted to Cees Vermeulen at DUO (the Dutch government organisation that collects all educational data), who provided us with an extended data set of great quality. We acknowledge his unique professionalism in designing and loading this new database. Moreover, we valued his advice in finding our way in the database. We are also deeply grateful for the support of Erik Smits and Rob Kerstens (Director‐ General of DUO). 

Page 97: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

71 

Netherlands,  avoiding  the  dangers of  traffic may  also  be  a  consideration  for  parents. 

There may  be  other  reasons  though:  native Dutch  pupils  living  in  affluent  areas may 

prefer  a  school  close  to  home,  because  this  school  presumably mirrors  the  high  SES 

composition of  the residential area. Migrant students, on  the other hand, may choose 

the nearest school because  this school may be more specialized  in migrant education, 

for example with an extra emphasis on language proficiency. Migrant students may also 

have reasons to choose a school where they form the majority. In summary, preferring 

the nearest school may, or may not be, a deliberate choice; however, on the basis of our 

available data we cannot diversify between different motives behind the choice for the 

nearest school. We assume, however,  that making  the extra effort  to travel  further  to 

school is the result of active choice.  

We explore the following research questions:  

1. Does  distance  to  school,  as  a measure  of  selectivity  of  school  choice,  differ 

between ethnic groups, and does  the group of pupils who  choose  the nearest 

school differ from those who opt for a school further away?  

2. How is school choice associated with ethnic segregation?  

3. Do pupils benefit from selective school choice? 

 

The outline of the chapter is as follows: previous research is reviewed in Section 

4.2; the data sources we used are described in Section 4.3; the results are presented in 

Section  4.4;  conclusions,  discussion  and  possible  policy  implications  for  policies  are 

presented  in Section 4.5. The results show that, of all pupils  in secondary schools  living 

in the four largest cities, 88.7 % are not enrolled in the school nearest to their residence. 

Pupils  of Dutch  origin,  living  in  poorer  neighbourhoods,  have  a  stronger  tendency  to 

choose a  school at a  further distance  than migrant  students  in  such neighbourhoods. 

Differences between the group of pupils who choose the nearest school and those who 

choose  the  non‐nearest  school  are  marginal.  Prior  achievement  in  elementary 

education, secondary school average exam scores, and school‐level upward mobility are 

only marginally associated with distance travelled to school. The school's percentage of 

students  with  a  non‐Western  immigrant  background  and  the  school  population's 

average  SES,  however,  do  significantly  drive  school  choice.  Choice  patterns  seem  to 

reveal ethnic segregation by choice among migrant pupils. 

Page 98: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

72 

4.2 Previous research  

Our data allow for measurements of socio‐economic pupil, neighbourhood, and 

school characteristics;  for qualitative characteristics on  the  individual and school  level; 

and for distance measurements. In the literature reviewed in this section, we have made 

an attempt to address this wide array of factors related to the patterns and geographic 

distribution  of  school  choice.  Segregation,  neighbourhood  schools,  differential  effects 

related  to  distinct  pupil  characteristics,  postponed  tracking,  distance  from  home  to 

school  and  extended  school  choice  are  the  key  issues  we  explored  in  academic 

literature: 1) We studied: 1) research that points out the benefits of mixed desegregated 

schools,  identifying  peer  effects  and  parental  support  by  better‐educated  parents  as 

main mechanisms behind higher expected outcomes; but 2)  in contrast,  research  that 

finds positive effects of segregated neighbourhood schools in the case of migrant pupils, 

that  indicates the  importance of close cooperation between parents and school; 3) yet 

other  research  findings  that  demonstrate  differential  effects  for  different  groups.  For 

example,  students  in  vocational  trajectories who might  benefit  from  neighbourhood 

schools,  and,  on  the  other  hand,  low  SES,  talented  students, who might  have  better 

opportunities in a school with a higher average SES; 4) scientific research that considers 

the effect of postponement of the age of tracking  in secondary schools, as well as the 

availability of  comprehensive  schools  that offer all  tracks  (thus enhancing  chances  for 

upward mobility);  5)  research  that  considers  the  distance  to  school  specifically  and 

different patterns of school choice between ethnic groups. And, finally, 6) the expected 

effects of extending school choice.  

School composition 

The  expected  mechanism  behind  desegregation  policies  (typically  involving 

greater  distances  to  school)  is  often  based  on:  1)  the  assumed  positive  peer  effects 

between pupils from various backgrounds; 2) assumed positive effects of more support 

in  school  activities by higher  SES parents;  and  3)  a probable  teacher  selection effect, 

because  high  quality  teachers  might  prefer  high  SES  schools.  Desegregation  in  this 

context  involves  the  reduction  of  ethnic  segregation  between  schools.  If  residential 

areas  are  ethnically  segregated,  this  implies  aiming  for  a  school  composition  that 

deviates from the ethnic composition of the neighbourhood where the school is located. 

Page 99: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

73 

Research  outcomes  on  the  effects  of  school  composition  appear,  however,  to  be 

ambiguous. While, according to some authors, the overall effect may be  limited, other 

authors  warn  that  assumedly  underestimated  compositional  effects  or  differential 

effects  should be  treated with  caution  (e.g. Palardy, 2008, Vigdor  and  Ludwig, 2010). 

Thrupp, Lauder and Robinson  (2003),  in  their comparison of school compositional and 

peer effects in the United States, the United Kingdom, Belgium and New Zealand, argue 

that the  little consensus over these effects might be due to  inadequate theorizing and 

research  design.  In  order  to  be  able  to  attribute  differences  in  school  quality  and 

outcomes to school or pupil characteristics, it is a prerequisite, so they argue, to be able 

to distinguish between the effects of school  leadership and  instructional quality on the 

one  hand,  and  pupil  composition  and  peer  effects  on  the  other  hand.  The  authors 

emphasize that peer and school effects can, in fact,  only be measured adequately, when 

prior achievement at the individual pupil level is known. This means in the Dutch context 

that  the pupils' CITO21 score at  the end of elementary education, plus  the elementary 

teacher's advice for the appropriate track level in secondary education, should be linked 

to  later actual achievements  in secondary education,  in order  to assess  the secondary 

school's added value. 

Ho Sui‐Chu and Willms (1996) find that children perform better in schools with a 

high mean SES of parents, and  illustrate that this positive effect  is  indeed mediated by 

parental  involvement.  The  educational  systems  in  the  United  States  and  in  the 

Netherlands  differ,  however,  in  important  characteristics  related  to  these  research 

outcomes: although in the United States many schools with low SES pupils receive extra 

funding  schools  also  depend  on  local  tax  revenues;  poor  neighbourhoods may more 

often have poor schools with a high  turnover of  teachers, which  further burdens poor 

children. In the Netherlands all children from low SES parents receive extra funding, and 

all  schools  receive  equal  standard  funding  from  the  government.  Schools  in  the 

Netherlands,  therefore,  hardly  depend  on  financial  support  by  parents.  However, 

parental assistance  in school processes and projects may vary also  in the Netherlands, 

depending on average levels of schooling of parents. 

                                                       21 The final test of elementary education; scores range between 500‐ and 550; the national average is 535. 

Page 100: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

74 

Opdenakker  and  van Damme  (2001)  suggest  that  composition  effects may  be 

larger  than many  researchers assume, but  the mechanism behind  these effects might 

not  so  much  be  a  peer  effect,  but  also  the  result  from  the  influence  of  school 

composition  on  school  processes,  notably  an  orderly  learning  environment  and 

structural and effective cooperation between  teachers. Therefore,  the authors suggest 

that school effectiveness measurements might be biased  in favour of high SES schools, 

where  high  standard  outcomes might wrongly  be  attributed  to  high  standard  school 

processes,  while  in  fact  these  are  the  emergent  effects  of  school  composition. 

Interestingly, Opdenakker and van Damme (ibid.) find that low SES, high ability students 

might be twice as sensitive to school composition as low SES, low‐ability students.  

Neighbourhood schools 

Ryan Wells (2010) raises the issue that policies aimed at the enhancement of the 

levels of schooling of low SES youth, may in fact work out differently for migrant pupils, 

than  for  low  SES  native‐born  pupils.  Ryan  Wells  (ibid.),  in  accordance  with  often‐ 

reported research findings in the Unites States, reports that migrant youths on average 

express higher  expectations of  their  future  attainment  levels  than native‐born pupils. 

Whether this is also the case in the Dutch context, we do not know. Interestingly, Wells 

(ibid.) also reports, seemingly in contrast with these higher expectations among migrant 

youth,  that  immigrant  status  is  associated with  (self‐reported)  lower  expectations  in 

high  SES  schools‐  all  else  being  equal.  He  assumes  that  the  rationale  behind  these 

findings  is  that  migrant  pupils  may  rather  benefit  from  the  presence  of  successful 

migrant  students  as  a  reference  and  role model,  in  schools with  high  percentages  of 

migrant students, and may feel isolated in high SES schools. 

Especially  in  the  tracked and  fairly  complex  system  for  secondary education  in 

the Netherlands, access to information is an important prerequisite for parents in order 

to  be  able  to  choose  the  best  school  for  their  children.  Parents  need  to  be 

knowledgeable  specifically  about  the  consequences  of  the  choice  for  a  particular 

secondary  track,  considering  their  children's  future  options  for  tertiary  education. 

According to Cabrera and La Nasa (2001): "SES gaps are reduced, if not eliminated, once 

a number of school‐based and  family‐oriented  factors are taken  into account…  [these] 

practices  are  as  important,  if  not  more,  than  is  family's  SES  in  becoming  college 

Page 101: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

75 

qualified…" Following this line of reasoning, primary schools in migrant neighbourhoods 

may  play  an  important  role  in  informing  parents  about  the  options when  choosing  a 

secondary  school.  In  practice,  schools with  large  numbers  of migrant  pupils may  put 

family‐based policies that involve cooperation with parents much higher on the agenda 

than schools with more affluent, well‐educated parents. 

Finally, Bauder (2002) severely criticizes policies to send migrant pupils to schools 

outside  their  residential area,  for  the  sole purpose of desegregating  schools:  this may 

wrongly  stigmatize  students  attending  neighbourhood  schools  in  the  view  of  future 

employers,  because  by  implication  the  message  appears  to  be  that  neighbourhood 

schools in migrant communities are of insufficient quality.  

Mixed and differential effects 

Cullen,  Jacob  and  Levitt  (2000)  investigated  the  effect  of  expanded  parental 

choice within  the  Chicago  Public  School  System.  This  educational  reform  resulted  in 

about half of all the students opting out to attend another school. The authors describe 

how this intervention, in the first place, dramatically increased sorting. Disproportionally 

more motivated  students opted out, and  this  indeed  raised  the odds of graduating  in 

their  school of  choice.  They  found no evidence of  a negative effect on  students who 

remained  in  their  assigned  neighbourhood  school.  The  overall  effect  on  the 

desegregation  of  schools,  however,  turned  out  to  be  limited.  In  sum,  the  authors 

demonstrate  that  expanded  choice  affects  students  and  their  parents  differently, 

resulting  in more motivated  students  to opt  for  a new  school. With  the exception of 

what are called  the "Career Academies"22,  the  increased chance  to graduate  for  those 

who opt  for  another  school may be  correlated with motivation,  rather  than with  the 

quality of the chosen school. Finally, in the words of the authors, the "greatest puzzle" is 

why so many students opt for another school, while the academic benefits turn out to 

be  limited. This  is an  interesting question  in the Dutch context as well, where parental 

choice is free, but especially migrant students tend to choose a school on average closer 

to their residence address than native‐born students. 

                                                       22 Career Academies promote a college‐preparatory curriculum and career‐focused education in different fields. Students get the opportunity to visit local business, and shadow business professionals in various career areas. 

Page 102: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

76 

Rumberger and Palardy  (2005)  raise  the question of  (reversed)  causality when 

finding that schools with mostly lower‐income students tend to be organized differently 

from  schools with  affluent  students:  Should  school  reform  introduce  effective  school 

characteristics  in  schools  with  many  low‐income  students?  Or  are  effective  school 

characteristics  the  emergent  result  of  having  affluent  students?  In  the  first  option, 

school characteristics and processes may make the difference; in the second option, the 

composition of students affects outcomes. Furthermore, Rumberger and Palardy (ibid.) 

explore  whether  school  policies  and  organizational  characteristics  may  have  a 

differential effect on black and white students. Their study demonstrates  that  teacher 

expectations  (albeit  raising  teacher  expectations  structurally  may  be  a  complicated 

endeavour) and an academic  school climate  foster  the  success of  low SES  students;  if 

schools with many low SES students were to focus on these quality aspects, the authors 

do not expect that desegregation would offer any extra educational advantages. 

Finally, Konstantopoulos and Borman (2011) gained new interesting insights from 

their replicated analyses of Coleman's famous data (1966), with current, more advanced, 

statistical instruments. Coleman (ibid.) found that the pupils' background was a stronger 

predictor  for  student  achievement  than  school  quality.  While  these  findings  were 

reaffirmed by Konstantopoulos and Borman (ibid.), because of the current availability of 

multi‐level  models  for  statistical  analyses,  they  could  also  demonstrate,  however, 

significant  between‐school  variance:  "Our  results  also  indicated  that  schools  play 

meaningful  roles  in  distributing  equality  or  inequality  of  educational  outcomes  to 

females, minorities, and the disadvantaged." Similar to Coleman, they found that within‐

school variance  is  in  fact  larger  than  in‐between  school variance; but notwithstanding 

this, they could also demonstrate that "40 % of the total variability in achievement was 

attributable to differences among schools, and ...that schools have nontrivial effects [ ...] 

on the achievement gap." 

Opportunities for upward mobility  

Migrant students may need more time to discover their ambitions and capacities, 

for example, because  they may  lack  role models  in  their own  family. Especially when 

Dutch  is  not  the  language  spoken  at  home,  they may  also  need  some  extra  time  to 

acquire  the  level of Dutch  language proficiency  that  is needed  for  academic  tracks  in 

Page 103: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

77 

secondary education. This would imply that broad secondary schools that offer all tracks 

(vocational  and  academic),  and  which  are  specialized  in  assessing  the  capacities  of 

migrant  students  repeatedly  during  their  school  career,  would  be  an  advantage  for 

migrant students. Recent studies seem to support this hypothesis.  

Pekkarinen, Uusitalo and Kerr  (2009) made use of  the unique opportunity of a 

system‐wide school reform in Finland to investigate the effect of postponed tracking on 

the correlation between the labour market position of parents and the level of schooling 

of their children. Between 1972 and 1977, the  former Finnish two‐tracked system was 

replaced by a comprehensive secondary school system that shifted the age of choosing 

between  a professional or  an  academic  track,  from  age  11  to  16.  This  later  selection 

resulted in a substantial decrease of 23 % in intergenerational income elasticity.  

Comparable  research outcomes  (Hanushek and Woessman, 2006, Brunello and 

Checchi, 2007, Bauer and Riphahn, 2006) demonstrate that early selection and tracking 

negatively  affects  educational  outcomes  of  low  SES  students.  Considering  the 

intergenerational elasticity of high‐ and middle‐income students in the Swiss educational 

system,  Bauer  and  Riphahn  (2006)  find  that  "early  tracking  increases  the  absolute 

benefit  of  having  highly  vs.  mid‐way  educated  parents  and  magnifies  the  relative 

advantage of highly educated parents." Finally, also Bjorklund and Salvanes (2010) also 

find  that  postponement  of  tracking  may  potentially  reduce  intergenerational 

correlations with parental schooling. 

Interestingly,  Van  Elk  et  al.  (2011)  make  use  of  the  opportunity  that  in  the 

Netherlands  early  tracking  at  age  12,  and  postponed  tracking  at  age  13/14,  exist  in 

parallel. As described above, schools, furthermore, may offer all secondary tracks or may 

be specialized in either vocational or academic tracks. Considering schools that offer all 

tracks (lower vocational and academic), and that start with one or two comprehensive 

years, they find that pupils starting in the lower secondary vocational track, have a 26 % 

chance to complete higher education later on. Note that without upward mobility to an 

academic track, the secondary  lower professional track does not directly qualify pupils 

for  access  to  higher  education.  In  schools  that  only  offer  the  secondary  professional 

tracks (and no academic tracks), however, this chance is 21 %. The findings of van Elk et 

al.  (ibid.)  suggest  that  upward mobility  is more  feasible, when  all  tracks  are  offered 

within the same school. A 5 percentage point increase in the chance to complete higher 

Page 104: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

78 

education, after attending broad secondary schools  in the case of students starting on 

the pre‐vocational  level, seems of extra  importance for  low SES migrant students, who 

may need some more time to discover their ambitions and capacities.  

The distance between home and school 

Interestingly,  Nihad  Bunar  (2010)  explored  the  question  why  the  children  of 

immigrants in Sweden would prefer the (segregated) neighbourhood school even when 

they know the school is low‐performing and located in a high‐poverty area, while– as is 

the  case  in  the Netherlands–  school  choice  is  free.  Bunar  (ibid.)  interviewed migrant 

pupils  in  two  urban  schools  in Malmö  and  Stockholm,  and  found  that  neither  school 

quality, nor a lack of information on school quality, nor the costs of travelling to a school 

at  greater  distance,  can  fully  explain  the  choice  to  stay  in  an  urban  school  in  the 

proximity of the residential area. Bunar (ibid.) writes: "The answer is to be found in the 

process  of  negotiation  taking  place  within  the  realms  of  families  and  peer  groups 

oscillating  around  the  importance  of  relationships  that  provide  safety,  the  feeling  of 

belonging  and  cultural  recognition,  on  the  one  hand.  On  the  other  hand,  there  are 

detrimental  effects  of  categorization  and  stigmatization  attached  to  immigrants, 

neighbourhoods and schools."  

Harris,  Johnston and Burgess  (2007) report  in  their geodemographic analysis of 

ethnicity and school choice in Birmingham (England) that the likelihood of attending the 

nearest state‐funded secondary school indeed varies with the ethnic composition of the 

neighbourhood.  They  find  that white  pupils  are  always more  likely  not  to  attend  the 

nearest school, and that this likelihood is further increased by greater exposure to other 

ethnic groups  in  the  residential area. Harris et al.  (ibid.)  find evidence  that pupils may 

prefer to attend a school that is more representative for a pupils' own ethnic group; this 

may  result  in  a  stronger  segregation  at  the  school  level,  as  compared  to  the 

neighbourhood  level.  Allen  (2007)  reports  similar  findings  that  demonstrate  that 

choosing  another  school  than  the  nearest  school  tends  to  increase  social  and  ability 

sorting. Allen (ibid.) developed a new model, based on the availability of the pupils' zip‐

codes, to compare actual school choice data to the proximity counterfactual. Allen (ibid.) 

provides us with a telling  illustration of how choosing another school than the nearest 

one  implies  a  larger  average  distance  to  school  as  a  consequence:  "The  proximity 

Page 105: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

79 

allocation indicates that the typical journey currently made by a pupil is 60% longer than 

the minimum necessary. In fact, over 5 million kilometres of additional travel are made 

by 11 ‐ 16 year‐olds every school day…" 

Finally, Andersson, Malmberg and Östh (2012) describe how the liberalization of 

school choice in Sweden (20 years ago) indeed increased the average distance travelled 

to  school,  but  that  foreign‐born  students  travel  shorter  distances.  They  find, 

furthermore, that Swedish‐born students choose more distant schools, and tend to do 

so more often when residential areas have larger proportions of foreign‐born students, 

or larger proportions of socio‐economically disadvantaged families.  

Extended school choice 

Overall,  the  better match  of  pupils  and  schools  is  the  rationale  behind,  the 

expected educational  gains of extended  school  choice  (Gibbons, Machin,  Silva, 2008). 

Secondly, extended school choice is often thought of as an important strategy to reduce 

segregation  along  lines  of  SES,  ethnicity  or  race  (e.g.  Cullen  et  al.,  2000).  Ladd  et  al. 

(2009) find, however, that schools are to a  large extent segregated  in the  larger Dutch 

cities, notwithstanding the fact that school choice has been free and universal for almost 

a  century.  They  suggest  that  free  choice may  lead  to  Dutch  parents  to  avoid multi‐

cultural schools, thus increasing segregation. 

In  summary,  the  literature  reviewed above predicts  that  school  choice may be 

related  to  ethnicity  and  segregation,  to  school‐  and  neighbourhood  socio‐economic 

composition  and  to  school  quality;  furthermore,  those  effects  may  be  different 

depending  on  pupil  characteristics.  This  has  inspired  us  to  choose  distance 

measurements  as  the  core  of  our  research:  first,  distance measurements  allow  for  a 

comparison  of  school  choice  patterns  between  native Dutch  pupils  and  youth with  a 

migrant background,  relative  to  the  socio‐economic  characteristics of  their  residential 

area.  Second,  measuring  distance  to  the  more  distant  schools  as  a  proxy  for  the 

selectivity of choice, allows for analyses of school characteristics that may drive school 

choice,  like  average  exam  scores  and  the  percentage  of  upward mobility  to  a  higher 

track,  but  also  school  ethnic  composition  and  school  average  SES.  Third,  distance 

measurements may  shed  a  light  on  the  possible  benefits  of  deliberately  choosing  a 

school at further distance. 

Page 106: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

80 

4.3 Data and Methods  

4.3.1 BRON data and additional data sources 

Starting  in  2001  (based  on  a  new  law),  DUO23,  the  government  agency  that 

collects  all  educational  data  linked  to  the  individual  social  security  number,  is  in  the 

process of constructing a new educational database called BRON24. This database covers 

all  relevant  socio‐economic  background  characteristics  of  pupils,  including  their 

complete educational history. From 2008 on, all pupils  in elementary education will be 

recorded in the BRON data, starting with pupils in the final grade of elementary school in 

that year; at  the  time of our  research  this  first cohort had  reached  the  third grade of 

secondary education. In the Dutch tracked secondary system, many schools are (partly) 

comprehensive for the first two years. In the third year, however, the majority of pupils 

have been placed  in a specific track. This offered us the opportunity to relate the track 

level  in  the  third  year  at  secondary  school,  to  prior  achievement  in  elementary 

education. We  have  added  a  list with  the  definitions  of  all  variables we  used  in  this 

chapter in Section 4.6.1. 

4.3.2 Ethnic diverse populations in the four major cities  

In  the  current  chapter we  focus on  a  comparison of  school  choice patterns of 

pupils with  a  Dutch  background,  and  those with  a migrant  family  history.  Since  the 

majority  of  migrant  students  live  in  the  four  major  cities  (Amsterdam,  Rotterdam, 

Utrecht, and The Hague), we based our analyses on youths living in these cities. Another 

consideration for our choice to concentrate on these four cities was our assumption that 

in smaller cities, and especially in the countryside, school choices may structurally differ 

from  those  taken  in  urban  environments.  For  example,  considering  the much  lower 

density of schools in the countryside, school choice could be driven to a large extent by 

the sheer presence of a school. 

Our total data set contains in total 170.465 individuals living in the Netherlands, 

who  were  enrolled  in  the  last  grade  of  elementary  school  in  2008.  The  data  cover 

information  from 2008 up  to and  including 2011. From  this  total data set we selected 

pupils  living  in  the  four major  Dutch  cities  throughout  the  years  2008‐2011,  17.192 

                                                       23 Dutch acronym: Dienst Uitvoering Onderwijs. 24 Dutch acronym: Basis Register Onderwijs; official educational database. 

Page 107: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

81 

individuals. Next we dropped 997 individuals with a distance from home to school ≥ 10 

km  (elementary  school) or  ≥ 20 km  (secondary  school). These eliminations  resulted  in 

16.195  individuals. We  dropped,  furthermore,  124  individuals  because  of  incomplete 

distance data.  

This resulted  in N = 16.071  individuals we kept  for analyses. This  includes 2624 

pupils living in one of the four cities, who were enrolled in a school outside the city (but 

within  20  km).  The  four  major  cities  differ  in  size,  urbanicity  and  unobserved 

characteristics. Therefore, we added city dummies in our estimates to account for such 

aggregate effects. When we  introduced  the characteristics of  the nearest school as an 

explaining variable,  this reduced  the number of  individuals  to 11.023, since we do not 

have relative exam scores for the nearest school in the case of all individuals. 

In our analysis of the difference between pupils who choose the nearest school 

to  their  residence,  and  those who  do  not,  our  counts  demonstrate  that  1824  pupils 

chose the nearest school, while 14.247 chose a school at further distance (total 16.071). 

For measuring  individual upward mobility to a higher track  in secondary school, 

we selected all pupils who were enrolled in Year 1 of the lower vocational track (VMBO), 

and enrolled in Year 3 at the time of our research: 4343 pupils. 

Unless  clearly  indicated otherwise, we distinguish  five ethnic  groups: 1) native 

Dutch  pupils;  pupils  with  a  2)  Surinamese/Antillean‐,  3)  Turkish‐,  or  4)  Moroccan 

background; and 5) the combined group of all other non‐Dutch backgrounds, referred to 

in  the  tables  and  figures  as  "other  background".  The  last  group  is  very  diverse  and 

includes  many  nationalities  and  a  wide  variance  in  educational  attainment  and  in 

motives for immigration (e.g. the children of high skilled labour immigrants, but also the 

children of  low educated  refugees). The data set  includes 6368 pupils of Dutch origin, 

2450 of Moroccan descent, 1995 of Turkish descent, and 2269 came from Suriname or 

the Dutch Antilles. 2928  individuals had another type of  immigrant background; for 43 

individuals we did not know their ethnicity. The vast majority of pupils with a Turkish or 

Moroccan  background  were  born  in  the  Netherlands;  they  belong  to  the  second 

generation  immigrants.  For  the  sake  of  sufficient  statistical  power,  in  some  tables  a 

Page 108: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

82 

distinction  was  made  between  the  combined  group  of  all  immigrants  with  a  non‐

Western migrant background25 and others.  

4.3.3 Distance measurements 

For our distance measurements we used  the BRON  information on  the pupils' 

residence (4‐digit postal code) and the school address (6‐digit postal code). A student's 

geographic  coordinates  refer  to  the  centroid  of  the  neighbourhood  of  residence; 

neighbourhoods are defined by their corresponding 4‐digit postcode. In compliance with 

privacy  laws, the data do not allow the student to be  located at the 6‐digit postal code 

and  the  individual  household  level.  The  average  number  of  individuals  per  4  digit 

postcode area  is 100.7  (SD 52.3)26. Correspondingly,  in our data set clusters of around 

100  students  on  average,  residing  in  the  same  4‐digit  zip‐code  area,  share  the  same 

residential location. For each school, geographic coordinates are available at the exact 6‐

digit zip‐code school location level.  

Using these geographic coordinates for students and schools, Euclidian distances 

were calculated  in order  to derive  the distance‐to‐school measures. For each  student, 

the  school‐distance measure  is  thus  defined  as  the  Euclidian  distance  between  the 

centroid of the student's 4‐digit zip‐code neighbourhood, and the 6‐digit zip‐code school 

location he/she attends, measured in kilometres. We constructed a data set that covers 

the distance to the nearest relevant school (i.e. the nearest school that indeed offers the 

track  the  pupil  is  enrolled  in)  and  the  distance  to  the  actually  attended  school  (if 

different), for every individual.  

We used  three  types of distance measurements: 1)  the absolute distance  from 

home to school; 2) the difference in distance between the nearest and the actual school; 

3) a dummy for not attending the nearest school.  

                                                       25 The following definitions by Statistics Netherlands (CBS, the national statistics office, www.cbs.nl) have been used: 

Western immigrant: someone originating from a country in Europe (exclusive of Turkey), North America, Oceania, Indonesia or Japan. 

Non‐Western immigrant: someone origination from Africa (in the Netherlands the majority group of immigrants from Morocco), South America, Asia (exclusive of Indonesia and Japan) or Turkey. 

26 Percentiles of the numbers of individuals in our data set living in one 4‐digit postal code area: 41 (10%); 61 (25%); 90 (50%); 125 (75%); 174 (90%). 

Page 109: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

83 

For some analyses, we also make use of density measures, capturing the number 

of  schools within a  reasonable distance of  the pupils'  residence. We used  the data by 

Statistics Netherlands (CBS) on the number of schools within a 5 km radius.  

4.3.4 SES indices 

We  included  information  on  the  student's  residential  area  characteristics.  In 

particular,  we  make  use  of  data  collected  by  the  Netherlands  Institute  for  Social 

Research  (SCP).  These  data  provide  us  with  a  socio‐economic  index  for  every 

neighbourhood, known as  the  "status  scores".27  Information  for  these  status  scores  is 

collected  through household  (telephone)  surveys, one household per 6‐digit  zip‐code. 

Next,  these  data  are  aggregated  at  the  4‐digit  zip‐code  level.  The  data  include 

neighbourhood mean  income, employment and  level of  schooling  (all self‐reported by 

inhabitants). In our figures we present poverty indices in the reverse order, compared to 

SCP:  in all our  figures on neighbourhood SES and distance  to school,  the standardized 

poverty  score  ranges  from  ‐4  (poor)  to +4  (affluent). Lastly, we merged our data with 

additional  data  on  neighbourhood  characteristics  that  are  updated  annually  (e.g. 

demographics and urbanicity), provided by Statistics Netherlands (CBS).28 

4.3.5 Indicators of pupils' prior achievement and secondary school quality 

Prior  achievement  in  elementary  education  is  of  crucial  importance  for  the 

assessment of results in secondary schools. At the end of elementary education, children 

have a final test (the CITO test). The score on this test, plus the recommendation of the 

pupils' elementary teacher, typically determine the level of enrolment in a specific track 

in Dutch secondary education. The BRON database we used contains both the score on 

the  final elementary CITO  test, and  the  teacher's  recommendation, but unfortunately, 

we  only  had  the  results  of  the  CITO  test  in  the  case  of  3208  individuals  and  the 

elementary teacher's advice  in the case of 6081 pupils. We tested these subsets of the 

data on all  control  variables we use; we  found  that  the  regression  coefficients of  the 

controls remained stable. Therefore, we trusted that the subsets did not differ from the 

total set of 16.071 individuals in major ways. Furthermore, we compared the subsample 

with  only  the  CITO  score,  and  the  group with  only  the  teacher's  advice.  The  pupils 

                                                       27 SCP data set "Statusscores Postcodegebieden 2006". 28 CBS data set "Buurtkaart met Cijfers 2008" (update 2). 

Page 110: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

84 

without  CITO  scores  had  a  slightly  lower  mean  teacher's  advice  score,  but  similar 

individual  and  neighbourhood  characteristics.  We  acknowledge  the  limitation  of 

introducing  a  smaller  subsample  with  information  on  prior  achievement  of  pupils; 

however, we  considered  this  additional  information  relevant  in  combination with  the 

data on school quality we use in our analyses, in our search for quality‐driven elements 

(at the  individual and school  level) of school preference. The Inspectorate of Education 

(2011)  found  no  evidence  that  children  with  a migrant  background  receive  a  lower 

advice from their elementary teacher than native Dutch children. 

We used data provided by the Inspectorate of Education for our measurements 

of  school mean  exam  scores  and  upward mobility  to  higher  tracks,  as  indicators  of 

secondary school quality (see Section 4.6.1 for definitions of the variables).  

In  the  tables we  present  below, we  diversified  between  three  different main 

tracks  in  Dutch  secondary  education:  VMBO  (lower  secondary  vocational  education), 

HAVO  (senior  general  education),  and  VWO  (pre‐university  education),  plus  the 

combined advice HAVO/VWO. We used VMBO (the lowest level) as a reference, and the 

other  levels as dummy variables. We  included pupils who are eligible for extra funding 

because  of  special  educational  needs;  these  pupils  can,  in  principle,  enrol  in  every 

school.  In  the Netherlands only pupils with  specific  special needs  (e.g. blind  children) 

attend special schools; these pupils are not listed in our data set. 

4.3.6 Limitation 

We  had  to  accept  some  limitations,  because,  as  we  described  earlier,  the 

development of  the BRON data set  is currently  in progress. As a consequence,  for  the 

time  being,  we  could  not  link  results  on  (future)  final  secondary  exams  to  prior 

achievement  in elementary education. We attempted to bypass this barrier somewhat, 

by  using  mean  exam  results  at  the  school  level  (collected  by  the  Inspectorate  of 

Education)  to  estimate  whether  distance  to  the  preferred  school,  controlling  for 

individual  achievement  in  elementary  school,  is  associated  with  school  quality,  as 

expressed in the average exam score. 

4.4 Results  

In this Section we start with a general overview of summary statistics of the total 

data set of 16071 pupils  living  in the four Dutch major cities, who were enrolled  in the 

Page 111: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

85 

last grade of elementary education  in 2008; most of  them were enrolled  in Year 3 of 

secondary school (2011) at the time of our research (Table 4‐1). Next we present a first 

overall analysis of  the association between distance  to school and control variables at 

the  individual,  neighbourhood,  and  school  level; we  added  the  characteristics  of  the 

nearest school as an explaining variable for school choice (Table 4‐2). We continue with 

a comparison of the group of pupils who choose the nearest school, and those who opt 

out for a school at greater distance; we present the summary statistics in Table 4‐3, and 

regression  analyses  in  Table  4‐4.  In  Table  4‐5  we  consider  the  distance  difference 

between  the  nearest  and  the  actual  school,  as  a  proxy  for  the  selectivity  of  choice, 

assuming that pupils who take the extra effort to travel further made a more deliberate 

choice. In Table 4‐6 we present a comparison between the nearest school and the actual 

school, with a  focus on patterns of ethnic  segregation. We  investigate  these patterns 

specifically  by  considering  the  SES  difference  and  the  difference  in  percentages  of 

migrant pupils, when comparing the nearest (not chosen) and the actual school (Tables 

4‐7 and 4‐8). Finally we look at upward mobility to a higher track, as a specific example 

of a possible benefit of selective school choice in Table 4‐9. 

4.4.1 A general overview of socio‐economic measurements and distance to school 

As a visualization of our data, we have added  in Appendix 4‐1  the maps of  the 

four  cities, which  present  a  view  of  the  ethnic  composition  of  neighbourhoods,  the 

average  distance  travelled  to  school,  and  the  distribution  of  secondary  schools.  The 

maps give a  first  impression of  the association between  the ethnic composition of an 

area and the distance travelled to school.  

In Table 4‐1, we  set  the  scene and present  the  summary  statistics  for  the  four 

ethnicities  (including  native  Dutch)  that  we  distinguish  in  this  chapter,  plus  the 

combined group "other immigrants". Counts were carried out on the pupil‐, school‐ and 

neighbourhood level. In addition to socio‐economic variables, we present school quality 

indicators (mean exam score and percentage of pupils with upward mobility to a higher 

track) and  indicators of pupils' prior achievement  in elementary school (the CITO score 

and elementary teacher's advice for the track level in secondary school). 

Page 112: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

86 

Table 4‐1: Summary statistics for pupils living in the four major Dutch cities, at the individual, neighbourhood and school level 

   Dutch Sur./Ant. Turkish Moroccan Other Imm.  Total   mean mean mean mean mean  mean 

VARIABLE  (SD) (SD) (SD) (SD) (SD)  (SD) individual level       distance to primary school (km) 1.16 1.29 0.93 0.91 1.23  1.12    (1.21) (1.44) (1.17) (1.15) (1.4)  (1.27)   distance to secondary school (km)  3.49 3.30 2.92 2.67 3.17  3.21    (2.4) (2.48) (2.02) (1.89) (2.20)  (2.28)   school outside municipality border  0.21 0.17 0.13 0.10 0.12  0.16    (0.41) (0.38) (0.34) (0.30) (0.33)  (0.37)   CITO‐score  537.44 529.32 528.77 528.29 533.51  533.16    (9.93) (10.59) (9.68) (10.64) (11.03)  (11.03)   VMBO advice  0.42 0.72 0.75 0.76 0.54  0.58    (0.49) (0.45) (0.43) (0.43) (0.50)  (0.49)   HAVO advice  0.15 0.11 0.12 0.12 0.15  0.14    (0.36) (0.31) (0.32) (0.32) (0.36)  (0.34)   HAVO/VWO advice  0.18 0.10 0.07 0.08 0.14  0.13    (0.39) (0.30) (0.26) (0.26) (0.35)  (0.34)   VWO advice  0.24 0.07 0.06 0.05 0.17  0.15    (0.43) (0.26) (0.23) (0.22) (0.37)  (0.36)   individual upward mobility in Year 3  0.18 0.09 0.07 0.09 0.17  0.14    (0.39) (0.29) (0.26) (0.29) (0.38)  (0.35)   male  0.50 0.50 0.51 0.47 0.51  0.50    (0.50) (0.50) (0.50) (0.50) (0.50)  (0.50)   non‐western background  0.00 1.00 1.00 1.00 0.60  0.53    (.) (.) (.) (.) (0.49)  (0.50)   2nd generation non‐western background  0.00 0.85 0.91 0.93 0.81  0.52    (.) (0.36) (0.28) (0.26) (0.39)  (0.50)   weighted student funding  0.03 0.33 0.52 0.57 0.32  0.27    (0.10) (0.43) (0.44) (0.43) (0.42)  (0.40)   one‐parent household  0.06 0.26 0.02 0.02 0.12  0.09    (0.24) (0.44) (0.15) (0.12) (0.32)  (0.28)   nearest school chooser  0.12 0.14 0.10 0.10 0.11  0.11    (0.32) (0.34) (0.30) (0.30) (0.31)  (0.32)   distance difference chosen and nearest school  2.40 2.39 2.09 1.85 2.29  2.26    (2.26) (2.44) (1.96) (1.79) (2.12)  (2.17) neighbourhood level       neighbourhood SES index  0.84 ‐0.47 ‐0.91 ‐0.90 ‐0.12  0.00    (1.33) (1.33) (1.16) (1.14) (1.44)  (1.50)   urbanicity  3.42 3.60 3.81 3.81 3.67  3.60    (0.76) (0.57) (0.40) (0.39) (0.53)  (0.63)   distance to nearest relevant school (km)  1.08 0.91 0.83 0.82 0.88  0.95    (0.84) (0.64) (0.63) (0.64) (0.65)  (0.74)   # of relevant schools within 5 km  15.44 17.36 21.19 20.85 18.47  17.81    (8.00) (9.52) (9.07) (9.26) (9.35)  (9.10) school level       # of tracks offered at secondary school  2.76 2.93 2.99 3.07 2.86  2.88    (1.18) (1.30) (1.41) (1.28) (1.30)  (1.27)   relative exam scores secondary school  0.09 ‐0.04 ‐0.10 ‐0.13 0.02  0.01    (0.26) (0.25) (0.27) (0.26) (0.29)  (0.28)   % upward mobility at school level  15.50 19.10 21.05 22.23 17.28  17.93    (9.26) (8.73) (9.78) (10.27) (9.51)  (9.79)   N  6386 2269 1995 2450 2928  16071 

 

The mean average distance to secondary school is lowest for Moroccan youth at 

2.67 km. Native Dutch pupils more often attend the nearest school (12%) than migrant 

groups, with the exception of youth with a Surinamese/Antillean background (14%). Like 

native Dutch students, Surinamese and Antillean students travel a larger mean distance 

to  school,  but  at  the  same  time more  often  choose  the  nearest  school.  Presumably 

Surinamese  and Antillean pupils prefer a  school with  a  large  share of pupils with  the 

same background;  consider  for example Amsterdam: a  large  share of Surinamese and 

Antillean pupils  live  in the Bijlmer district  in the south‐eastern part of Amsterdam, and 

Page 113: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

87 

tend to go to local schools with large percentages of pupils belonging to this group; this 

might explain why they more often attend the nearest school than is the case for other 

groups. At the same time, Surinamese and Antillean pupils who do not live in the Bijlmer 

district, may still prefer to attend a school with many Surinamese and Antillean students 

in the Bijlmer district; this could explain the higher mean distance to school.  

Native Dutch pupils prefer more often prefer  the nearest  school;  the odds are 

that these students more often  live  in more affluent areas than  is the case for migrant 

students, and tend to attend a  local school with a population that mirrors the affluent 

neighbourhood. For example 57% of Moroccan pupils are eligible for weighted student 

funding in elementary education, against 3% among native Dutch pupils. 

The mean neighbourhood SES  index differs  substantially between  the different 

ethnic groups, by more than 1 Standard Deviation, and between native Dutch pupils and 

all  other  groups.  The  share  of  Dutch  pupils  enrolled  in  a  school  outside  the  city 

boundaries, at 21% differs  considerably  from,  for example,  students with a Moroccan 

background  (10%).  This  alludes  to  our  further  findings  further below, which  illustrate 

that native Dutch pupils tend to prefer the nearest school when they live in an affluent 

neighbourhood, but  leave  lower SES  residential areas  to go  to school elsewhere more 

often than migrant pupils living in the same area. 

Considering  pupils'  prior  achievement,  the mean  CITO  score  of  native  Dutch 

pupils  is  close  to  1  Standard  Deviation  higher  than,  for  example,  it  is  for  pupils  of 

Moroccan descent. Substantially more often, migrant children are advised to proceed in 

secondary vocational education (the ranking order of secondary tracks is VWO (highest), 

HAVO,  VMBO);  24%  of  native  Dutch  students  are  advised  to  proceed  to  the  highest 

secondary track (pre‐university track) from their elementary school teacher, against 5% 

and 6% of pupils with, respectively, a Moroccan or Turkish background.  

Successful progress  in  secondary  school, as measured by upward mobility  to a 

higher  track, differs  substantially between native Dutch and migrant  students: among 

immigrant groups, 7‐9% move up  to a higher secondary  track at  the start of Year 3 of 

Page 114: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

88 

secondary school, whereas this is the case for 18% of the Dutch students29. Besides the 

percentage of upward mobility, we considered the school's relative exit‐exam score to 

be  an  important  indicator  of  school  quality;  we  defined  this  indicator  as  the 

performance at each track level, measured by the mean exit‐exam score, compared with 

the mean exit‐exam score of all other schools  in  the  four major cities which offer  this 

same track. The school mean exam score is higher in the case of native Dutch pupils. 

4.4.2 Distance to school  

In Table 4‐2 we present OLS regressions  for  the  individual distance  travelled  to 

school. Except  for Surinamese/Antillean students,  immigrant students travel  less  far to 

school  than  native  Dutch  students.  This  conclusion  still  holds  with  the  inclusion  of 

additional  controls.  The magnitude  of  the  difference  declines  if  we  add  school  and 

neighbourhood characteristics, but  is restored  if we add school advice and CITO score. 

Specification  1  (individual  student  characteristics)  demonstrates  that  poorer  students 

(who  were  eligible  for  weighted  student  funding  in  elementary  education),  among 

whom  there are many Turkish and Moroccan migrant students, on average  travel  less 

far to school.  

In  Specification  2 we  introduce  neighbourhood  characteristics:  an  increase  in 

neighbourhood SES decreases distance to school. This may seem  in contrast with what 

we find at the individual level: poorer students (as measured by the mark‐up on funding 

per pupil  in elementary education– weighted student funding) travel  less far to school. 

However, while poorer migrant students tend to attend their neighbourhood school, the 

same  also  appears  to  be  the  case  for  affluent  native  Dutch  youth  in  higher  SES 

neighbourhoods.  

                                                       29 Data files of the Inspectorate of Education show, as a reference, that native Dutch pupils are enrolled in schools with a mean 15.5 % upward mobility (measured after the completion of final exams), while Turkish pupils are enrolled in schools with a mean 21.1 % upward mobility, and Moroccan pupils in schools with a mean 22.2 % upward mobility. This indicates that migrant students more often attend a school with a larger upward mobility, but do so less themselves (as table 1 shows) than native‐born students. 

Page 115: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

89 

Table 4‐2: Regression results (OLS) for distance to secondary school in km, on individual‐, school‐ and neighbourhood variables (standard errors in parentheses). 

   (1) (2) (3) (4) (5)    distance to distance to distance to distance to  distance to VARIABLES  sec. school sec. school  sec. school sec. school   sec. school  individual level   male  0.10*** 0.08** 0.05 0.16**  0.20**    (0.038) (0.038) (0.047) (0.063)  (0.081)   weighted student funding ‐0.30*** ‐0.18*** ‐0.19** ‐0.28***  ‐0.09    (0.076) (0.056) (0.074) (0.106)  (0.155)   Surinamese/Antillean  ‐0.11 ‐0.13 ‐0.12 ‐0.21  ‐0.18    (0.124) (0.089) (0.097) (0.137)  (0.216)   Turkish  ‐0.41*** ‐0.18* ‐0.23 ‐0.35**  ‐0.48*    (0.120) (0.102) (0.138) (0.158)  (0.258)   Moroccan  ‐0.65*** ‐0.52*** ‐0.54*** ‐0.65***  ‐0.65***    (0.111) (0.082) (0.107) (0.148)  (0.243)   Other Immigrant Background  ‐0.24*** ‐0.18*** ‐0.21*** ‐0.30***  ‐0.34**    (0.091) (0.064) (0.076) (0.097)  (0.161)   one‐parent household  0.08 0.12* 0.12 0.07  ‐0.02    (0.088) (0.069) (0.083) (0.102)  (0.133)   HAVO advice track dummy  ‐0.84***    (0.109)   HAVO/VWO advice track dummy  ‐0.89***    (0.122)   VWO advice track dummy ‐0.61***    (0.168)   CITO test score  ‐0.02***    (0.007)   Non‐western x CITO test score  0.00    (0.000) neighbourhood level   neighbourhood SES index ‐0.21*** ‐0.14** ‐0.04  0.05    (0.052) (0.068) (0.078)  (0.125)   Surinamese/Antillean x SES  0.14** 0.13* 0.12  0.14    (0.059) (0.072) (0.096)  (0.124)   Turkish x SES  0.19*** 0.19** 0.16*  0.17    (0.064) (0.076) (0.090)  (0.113)   Moroccan x SES  0.19*** 0.21** 0.15  0.04    (0.069) (0.085) (0.097)  (0.169)   Other Immigrant Background x SES  0.08** 0.09* 0.06  ‐0.02    (0.041) (0.046) (0.058)  (0.094)   urbanicity  ‐0.33*** ‐0.42*** ‐0.19  ‐0.28    (0.113) (0.157) (0.155)  (0.187)   # of relevant schools within 5 km  ‐0.03*** ‐0.04*** ‐0.05***  ‐0.05***    (0.008) (0.009) (0.009)  (0.011)   distance to nearest relevant school (km)  0.59*** 0.47*** 0.45***  0.51***    (0.117) (0.139) (0.143)  (0.163)   Utrecht municipality dummy  ‐0.13 0.12 0.10  0.07    (0.193) (0.250) (0.255)  (0.303)   The Hague municipality dummy  ‐0.55*** ‐0.67*** ‐0.49**  ‐0.88***    (0.139) (0.189) (0.193)  (0.257)   Rotterdam municipality dummy  ‐0.51*** ‐0.50*** ‐0.32*  ‐0.53***    (0.137) (0.174) (0.177)  (0.201) school level   relative exam scores nearest school  ‐0.32 ‐0.30  ‐0.17    (0.266) (0.262)  (0.322)   # of tracks offered at nearest school  ‐0.02 ‐0.01  ‐0.10    (0.051) (0.054)  (0.080)   % upward mobility at nearest school  ‐0.49 0.09  ‐0.38    (0.606) (0.570)  (0.704)   average SES index at nearest school  ‐0.25*** ‐0.27***  ‐0.42***    (0.093) (0.104)  (0.130)   Constant  3.44*** 4.89*** 5.48*** 5.08***  18.22***    (0.117) (0.600) (0.828) (0.805)  (3.695)   Observations  16,071 16,060 11,023 6,081  3,208 R‐squared  0.02 0.14 0.14 0.16  0.18 Adj. R‐squared  0.0200 0.139 0.139 0.158  0.171 

Notes: Robust clustered standard errors (at neighbourhood level) in parentheses *** p<0.01, ** p<0.05, * p<0.1 

 

Page 116: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

90 

In Specification 3 we introduce the characteristics of the nearest relevant school 

as an explaining variable. The two school quality indicators we consider in this chapter– 

mean exam score and percentage upward mobility– do not seem to drive school choice. 

A higher average SES of  the nearest school, however, decreases distance  to school,  in 

accordance with what we described above: native Dutch pupils  in richer areas tend to 

attend their neighbourhood school.  

In  Specification  4  we  add  the  elementary  teacher's  advice  for  the  pupils' 

appropriate  track  in  secondary  school. Pupils who are advised  to  follow  the academic 

secondary  tracks  (HAVO  and VWO),  relative  to  the  lowest professional  track  (VMBO), 

travel  less  far  to school. This again seems  related  to what we described above: native 

Dutch pupils more often have affluent parents and  live  in higher SES residential areas; 

the  children of more  affluent, better‐educated parents more often enrol  in  academic 

tracks (e.g. 58% among native Dutch, and 25% among pupils of Moroccan descent, Table 

4‐1). We have seen that affluent native Dutch pupils tend to attend their neighbourhood 

school, hence the lower mean distance to school among pupils in academic tracks.  

More migrant  than  native Dutch  students  start  in  secondary  vocational  tracks 

(VMBO);  on  average,  enrolment  in  a  vocational  track may  require  somewhat  longer 

travel distances:  in contrast to academic tracks, at the vocational  level a wide range of 

professional  programmes  is  offered  (that  correspond  to  professions  on  the  labour 

market), but not all  these options are offered  in every  school. Therefore, enrolling on 

the preferred vocational programme may imply further travelling to school. 

Finally,  in Specification 5, we add  the CITO  test score; unfortunately, similar  to 

the  elementary  teacher's  advice,  these  analyses  are  based  on  a  substantially  smaller 

subsample.  We  decided,  however,  to  include  these  regressions,  because  we  found 

hardly  any  instability  to  controls. We  trusted,  therefore,  that  the  subsample may not 

differ  in major ways  from  the  total  data  set.  The  effect  of  the  pupils'  CITO  score  on 

distance to school is marginal but statistically significant.  

As we presented in Table 4‐1, only 10‐14% (depending on the ethnicity) of pupils 

chooses  the nearest  school  to  their  residence. For  this  reason, we  looked  further  into 

possible differences between the two groups, those who choose the nearest school, and 

those who do not. 

Page 117: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

91 

Figure  4‐1  demonstrates  these  two  effects.  The  interaction  of  ethnicity  with 

neighbourhood  SES,  furthermore,  diminishes  the  effect  of  ethnicity  on  distance  to 

school: when  neighbourhood  SES  increases,  the  odds  are  that  Turkish  and Moroccan 

students travel further to school. We see an opposite effect between native Dutch and 

migrant students here: while the first group tends to choose a school closer to home in a 

high SES area, our analyses estimate the opposite effect for migrant students. 

A  greater  distance  to  the  nearest  relevant  school:  a  greater  distance  to  the 

nearest school may point at a  lower population density and, accordingly,  lower school 

density. In that case, a preferred school may tend to be further away than  in a densely 

populated area.  

Figure 4‐1: Linearly fitted lines (OLS), one for each of the 5 ethnic groups on the correlation between absolute distance to secondary school and relative neighbourhood SES 

 

4.4.3 Choosing the nearest school or not 

Only  1824  (11.3%)  pupils  chose  the  nearest  school  to  their  residence,  while 

14.247 (88.7%) pupils chose another school, illustrating that the right to choose freely is 

largely  exercised  (Table  4‐3).  Note  that  preferring  the  nearest  school may  also  be  a 

deliberate choice, however,  in this case we cannot know; therefore, the percentage of 

pupils who actively choose a school could in reality even be higher. In these descriptive 

Page 118: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

92 

measurements  we  used  the  higher  aggregate  level  of  the  combined  group  of  non‐

Western  immigrants,  in order to avoid  losing too much statistical power  if we were to 

specify 1824 pupils (nearest school choosers) into five ethnicities.  

Table 4‐3: Summary statistics of the comparison between the group of students who choose the nearest relevant school (i.e. a school that indeed offers a desired track) to their home, and those who do not (standard deviations in parentheses) 

Non‐Nearest Nearest   mean meanVARIABLE  (SD) (SD)individual level   distance to primary school (km) 1.13 1.08   (1.28) (1.19)  distance to secondary school (km)  3.49 1.04   (2.26) (.85)  CITO test score  532.96 534.53   (11.06) (10.77)  VMBO advice  0.59 0.47   (.49) (.5)  HAVO advice  0.13 0.18   (.34) (.39)  HAVO/VWO advice  0.13 0.15   (.34) (.36)  VWO advice  0.15 0.19   (.35) (.4)  individual upward mobility in Year 3  0.14 0.15   (.34) (.36)  male  0.50 0.51   (.5) (.5)  non‐western background  0.53 0.53   (.5) (.5)  2nd generation non‐western background  0.52 0.52   (.5) (.5)  weighted student funding  0.27 0.25   (.4) (.39)  one‐parent household  0.09 0.09   (.28) (.29)  nearest school chooser  0.00 1.00   (.) (.)  distance difference chosen and nearest school  2.55 0.00   (2.14) (.)neighbourhood level   neighbourhood SES index  ‐0.03 0.17   (1.49) (1.57)  urbanicity  3.62 3.42   (.61) (.76)  # of relevant schools within 5 km  18.17 15.01   (9.04) (9.17)  distance to nearest relevant school (km)  0.94 1.04   (.72) (.85)school level   # of tracks offered at secondary school  2.86 3.04   (1.28) (1.2)  relative exam scores secondary school  0.01 0.00   (.28) (.27)  % upward mobility at school level  17.89 18.32   (9.87) (9.15)  N  14247 1824

Measurements  overall  demonstrate  only  marginal  differences  between  the 

groups  of  nearest  and  non‐nearest  school  choosers.  The main  differences we  found 

concern  the  elementary  teacher's  advice  to  follow  the  vocational  track  (VMBO)  and 

neighbourhood SES  (richer native Dutch pupils who choose the nearest school). As we 

described  above,  attending  the  preferred  VMBO  programme  may  require  further 

Page 119: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

93 

travelling, since not all schools offer all vocational programmes. We  found virtually no 

difference  between  the  two  groups  considering  immigrant  family  background  or 

eligibility for weighted student funding in elementary school. 

Table 4‐4: Regression results (OLS) for choosing the nearest school (i.e. comparing pupils who choose the nearest school with those who choose another school), with choosing the nearest school = 1 

(1) (2) (3) (4) (5)VARIABLES  nearest nearest nearest nearest nearest   male  0.00 0.00 0.01 0.01 0.02    (0.006) (0.006) (0.007) (0.009) (0.012) weighted student funding ‐0.01 0.01 0.01 ‐0.00 ‐0.01    (0.009) (0.009) (0.011) (0.015) (0.023) Surinamese/Antillean  0.03 0.03* ‐0.03 ‐0.03 ‐0.02    (0.017) (0.014) (0.022) (0.024) (0.036) Turkish  ‐0.01 0.01 ‐0.03 ‐0.03 0.01    (0.014) (0.014) (0.022) (0.025) (0.040) Moroccan  ‐0.01 0.01 ‐0.03 ‐0.01 0.01    (0.015) (0.013) (0.024) (0.028) (0.037) Other Immigrant Background  ‐0.00 0.01 ‐0.02 ‐0.01 ‐0.01    (0.011) (0.009) (0.014) (0.018) (0.027) one‐parent household  ‐0.01 ‐0.01 ‐0.02 0.00 ‐0.00    (0.009) (0.008) (0.011) (0.016) (0.027) neighbourhood SES index 0.00 ‐0.01 ‐0.01 ‐0.01    (0.007) (0.010) (0.012) (0.016) Non‐western x SES  ‐0.01 0.01 0.00 0.00    (0.008) (0.010) (0.012) (0.015) urbanicity  ‐0.05*** ‐0.04** ‐0.06** ‐0.05*    (0.017) (0.022) (0.025) (0.030) # of relevant schools within 5 km  ‐0.00*** ‐0.00*** ‐0.00*** ‐0.00***    (0.001) (0.001) (0.001) (0.001) distance to nearest relevant school (km)  ‐0.02 ‐0.00 ‐0.00 ‐0.00    (0.017) (0.026) (0.028) (0.033) Utrecht municipality dummy  ‐0.04* ‐0.06* ‐0.09** ‐0.09**    (0.024) (0.034) (0.037) (0.041) The Hague municipality dummy  ‐0.01 ‐0.01 ‐0.02 ‐0.03    (0.018) (0.025) (0.027) (0.032) Rotterdam municipality dummy  ‐0.03 ‐0.02 ‐0.03 ‐0.01    (0.019) (0.023) (0.024) (0.034) relative exam scores nearest school  0.06 0.06 0.05    (0.036) (0.042) (0.056) # of tracks offered at nearest school  ‐0.01 ‐0.01 ‐0.00    (0.006) (0.007) (0.011) % upward mobility at nearest school  0.00 ‐0.04 ‐0.06    (0.076) (0.091) (0.121) average SES index at nearest school  0.02 0.02 0.04    (0.017) (0.018) (0.025) % non‐western at nearest school  0.00 ‐0.02 ‐0.06    (0.059) (0.053) (0.084) % non‐western at nearest school x non‐western 0.09*** 0.09*** 0.17***    (0.026) (0.030) (0.064) HAVO advice track dummy 0.05**   (0.023)HAVO/VWO advice track dummy  0.01   (0.027)VWO advice track dummy 0.02   (0.028)CITO test score  0.00    (0.001) Non‐western x CITO test score  ‐0.00    (0.000) Constant  0.11*** 0.37*** 0.39*** 0.47*** ‐0.24    (0.011) (0.072) (0.096) (0.108) (0.576)   Observations  16,071 16,017 10,993 6,065 3,208 R‐squared  0.00 0.02 0.03 0.04 0.05 Adj. R‐squared  0.000901 0.0189 0.0287 0.0400 0.0436 

Notes: Robust clustered standard errors (at neighbourhood level) in parentheses *** p<0.01, ** p<0.05, * p<0.1 

Page 120: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

94 

Table 4‐4  lists  the  regressions  (OLS) on  these differences between nearest and 

non‐nearest  school  choosers,  and  we  used  again  the  characteristics  of  the  nearest 

school  as  explanatory  variables  for  the  choice  for  the  proximity  school  or  a  school 

further from the residence. Similar to the summary statistics in Table 4‐3, regressions in 

Table  4‐4  demonstrate  hardly  any  dissimilarity  between  pupils  choosing  the  nearest 

school,  and  those  who  do  not.  The  key  result  is  a  significant  effect  of  interaction 

between ethnicity and ethnic school composition: non‐Western pupils are more likely to 

choose the nearest school if this school holds more non‐Western pupils. The coefficients 

(not statistically significant) we  found considering  the student's ethnic background are 

not sensitive  to additional controls. Understandably, urbanicity– a measure  for human 

activity  in a neighbourhood based on the number of addresses– drives the choice for a 

more distant school: there are more schools to choose from in densely populated areas, 

and, as a consequence, the distance difference between the nearest and the preferred 

school tends to be smaller; the same can be said for the number of schools within 5 km. 

Recapitulating the above, we have seen that socio‐economic indicators influence 

distance  to  school,  but  that  the  comparison  between  pupils who  attend  the  nearest 

school and  those who prefer another  school does not demonstrate major differences 

between these two groups. Next, we introduce measurements of the distance difference 

between the nearest and the actual school as a proxy for the selectivity of choice. 

4.4.4 Selectivity of choice 

Unlike  in  Table  4‐2, where we measured  the  absolute  distance  from  home  to 

school,  in Table 4‐5 we consider  the distance difference between  the nearest and  the 

actual school, as a measure for the selectivity of choice, assuming that selecting a school 

at  greater  distance  implies  a  more  deliberate  choice,  and  has  a  higher  cost.  In 

accordance with the measurements we presented above, an increase in neighbourhood 

SES decreases this distance difference significantly, indicating that pupils living in higher 

SES  areas,  more  often  choose  the  nearest  school  (i.e.  more  pupils  with  distance 

difference = 0). As we also found earlier (Table 4‐2), the  interaction term non‐Western 

migrant background x SES neighbourhood reduces the effect of ethnicity considerably.  

Page 121: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

95 

Table 4‐5: Regression results (OLS) for the distance difference (i.e. between the actual and the nearest school), with distance difference = 0 for pupils who choose the nearest school 

(1) (2) (3) (4) (5) (6)VARIABLES  Dist. Diff  Dist. Diff.  Dist. Diff.  Dist. Diff.  Dist. Diff.  Dist. Diff. 

  male  0.08** 0.08** 0.05 0.08*  0.16**  0.20**   (0.038) (0.038) (0.047) (0.047)  (0.063)  (0.080)weighted student funding ‐0.22*** ‐0.20*** ‐0.20*** ‐0.22***  ‐0.30***  ‐0.11   (0.064) (0.055) (0.073) (0.068)  (0.105)  (0.149)Surinamese/Antillean  0.01 ‐0.10 ‐0.02 ‐0.06  ‐0.10  ‐0.24   (0.126) (0.086) (0.122) (0.124)  (0.164)  (0.205)Turkish  ‐0.20** ‐0.18** ‐0.17 ‐0.26*  ‐0.26  ‐0.59***   (0.090) (0.085) (0.147) (0.158)  (0.175)  (0.218)Moroccan  ‐0.43*** ‐0.52*** ‐0.50*** ‐0.65***  ‐0.55***  ‐0.66***   (0.087) (0.070) (0.138) (0.156)  (0.158)  (0.207)Other Immigrant Background  ‐0.07 ‐0.13** ‐0.12 ‐0.19**  ‐0.20*  ‐0.33**   (0.079) (0.062) (0.087) (0.091)  (0.119)  (0.150)one‐parent household  0.20** 0.13* 0.14* 0.11  0.09  ‐0.03   (0.088) (0.069) (0.082) (0.105)  (0.103)  (0.133)neighbourhood SES index ‐0.19*** ‐0.11 ‐0.16**  ‐0.01  0.07   (0.050) (0.067) (0.076)  (0.077)  (0.119)Non‐western x SES  0.13*** 0.12** 0.17***  0.07  0.06   (0.044) (0.057) (0.058)  (0.066)  (0.096)urbanicity  ‐0.32*** ‐0.43*** ‐0.72***  ‐0.20  ‐0.32*   (0.112) (0.155) (0.197)  (0.151)  (0.181)# of relevant schools within 5 km  ‐0.03*** ‐0.04*** ‐0.05***  ‐0.05***  ‐0.05***   (0.008) (0.009) (0.011)  (0.009)  (0.011)distance to nearest relevant school (km)  ‐0.40*** ‐0.52*** ‐0.66***  ‐0.55***  ‐0.51***   (0.117) (0.141) (0.162)  (0.145)  (0.162)relative exam scores nearest school  ‐0.30 ‐0.11  ‐0.25  ‐0.21   (0.260) (0.309)  (0.252)  (0.309)# of tracks offered at nearest school  ‐0.01 ‐0.04  ‐0.00  ‐0.08   (0.052) (0.060)  (0.055)  (0.077)% upward mobility at nearest school  ‐0.49 ‐0.57  0.09  ‐0.51   (0.602) (0.711)  (0.556)  (0.686)average SES index at nearest school  ‐0.32*** ‐0.33**  ‐0.33**  ‐0.55***   (0.114) (0.135)  (0.131)  (0.167)% non‐western at nearest school  ‐0.20 ‐0.19  ‐0.07  ‐0.05   (0.302) (0.332)  (0.336)  (0.457)% non‐western at nearest school x non‐western ‐0.10 0.18  ‐0.17  ‐0.75*   (0.193) (0.207)  (0.195)  (0.407)Utrecht municipality dummy  ‐0.12 0.14 ‐0.03  0.12  0.04   (0.194) (0.247) (0.274)  (0.256)  (0.299)The Hague municipality dummy  ‐0.55*** ‐0.67*** ‐0.85***  ‐0.49***  ‐0.90***   (0.138) (0.183) (0.210)  (0.187)  (0.254)Rotterdam municipality dummy  ‐0.51*** ‐0.53*** ‐0.72***  ‐0.33**  ‐0.65***   (0.137) (0.162) (0.185)  (0.168)  (0.197)HAVO advice track dummy ‐0.85***    (0.106) HAVO/VWO advice track dummy  ‐0.90***    (0.122) VWO advice track dummy ‐0.63***    (0.170) CITO test score  ‐0.03***   (0.007)Non‐western x CITO test score  0.00*   (0.001)Constant  2.36*** 4.84*** 5.58*** 7.62***  5.10***  19.03***   (0.082) (0.600) (0.829) (0.955)  (0.794)  (3.714)  Observations  16,071 16,017 10,993 9,504  6,065  3,208R‐squared  0.01 0.05 0.07 0.13  0.09  0.10Adj. R‐squared  0.0100 0.0528 0.0708 0.127  0.0913  0.0949

Notes: Robust clustered standard errors (at neighbourhood level) in parentheses *** p<0.01, ** p<0.05, * p<0.1 

Page 122: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

96 

Table 4‐6: Comparisons of nearest (non‐chosen) school and the actual school, considering non‐nearest school choosers only (14247 individuals, 88.7 % of the data set) 

Non‐nearest school choosing Dutch Sur./Ant. Turkish Moroccan Other Imm. Total   mean  mean mean mean mean  meanVARIABLE  (SD)  (SD) (SD) (SD) (SD)  (SD) NEAREST (NON‐CHOSEN) SCHOOL % non‐western at school  54.35  66.81 66.49 63.9 60.74  60.25   (29.18)  (27.96) (29.27) (30.) (29.95)  (29.75)school level average SES index ‐0.08  ‐0.69 ‐0.73 ‐0.52 ‐0.46  ‐0.38   (1.)  (.86) (.85) (.86) (.95)  (.97)relative exam school average  0  ‐0.1 ‐0.16 ‐0.16 ‐0.05  ‐0.07   (.28)  (.3) (.31) (.32) (.3)  (.3) school level average CITO intake 529.23  526.14 524.58 526.5 527.39  527.55   (9.17)  (8.96) (9.31) (9.83) (9.69)  (9.49)% upward mobility at school level  0.18  0.2 0.23 0.21 0.2  0.2    (.1)  (.1) (.11) (.11) (.11)  (.11)#  of  tracks  offered  at  secondary  2.29  2.35 1.95 2.07 2.32  2.23   (1.54)  (1.6) (1.75) (1.69) (1.56)  (1.61)ACTUAL (CHOSEN) SCHOOL % non‐western at school  32.56  63.24 74.09 75.53 52.82  52.4   (20.64)  (25.14) (23.23) (23.2) (28.12)  (29.46)school level average SES index 0.56  ‐0.33 ‐0.6 ‐0.58 ‐0.04  0    (.86)  (.91) (.82) (.77) (.96)  (.99)relative exam school average  0.09  ‐0.03 ‐0.08 ‐0.11 0.03  0.02   (.26)  (.26) (.28) (.27) (.3)  (.29)school level average CITO intake 535.81  529.37 528.63 528.71 532.95  532.4   (8.89)  (9.26) (8.95) (8.85) (9.71)  (9.62)% upward mobility at school level  0.15  0.19 0.21 0.22 0.17  0.18   (.09)  (.09) (.1) (.1) (.1)  (.1) #  of  tracks  offered  at  secondary  2.75  2.9 2.96 3.06 2.84  2.86   (1.18)  (1.33) (1.43) (1.29) (1.32)  (1.28)ACTUAL ‐ NEAREST    mean  mean mean mean mean  meanVARIABLE  (Pr{|T| >  (Pr{|T| >  (Pr{|T| >  (Pr{|T| >  (Pr{|T| >  (Pr{|T| > % non‐western at school  ‐21.79  ‐3.57 7.6 11.63 ‐7.92  ‐7.85   (0.000)  (0.000) (0.000) (0.000) (0.000)  (0.000)school level average SES index 0.64  0.36 0.13 ‐0.06 0.42  0.38   (0.000)  (0.000) (0.000) (0.002) (0.000)  (0.000)relative exam school average  0.09  0.07 0.08 0.05 0.08  0.09   (0.000)  (0.000) (0.000) (0.002) (0.000)  (0.000)school level average CITO intake 6.58  3.23 4.05 2.21 5.56  4.85   (0.000)  (0.000) (0.000) (0.000) (0.000)  (0.000)% upward mobility at school level  ‐0.03  ‐0.01 ‐0.02 0.01 ‐0.03  ‐0.02   (0.000)  (0.004) (0.000) (0.026) (0.000)  (0.000)#  of  tracks  offered  at  secondary  0.46  0.55 1.01 0.99 0.52  0.63   (0.000)  (0.000) (0.000) (0.000) (0.000)  (0.000)  N  5649  1959 1796 2198 2605  14247

Furthermore,  a  higher  SES  index  of  the  nearest  school  reduces  the  average 

distance  difference,  reflecting  that more  pupils  attend  the  nearest  school when  this 

school has a higher SES population. Compared with the vocational track as a reference 

(VMBO),  the distance difference  is  reduced  in  the case of pupils enrolled  in academic 

tracks  (HAVO  and  VWO).30  While  overall  we  found  no  major  differences  between 

analyses of the absolute distance to school and of the distance difference between the 

actual and the nearest school, in the latter case the magnitude seems more sensitive to 

controls.  As we  found  repeatedly  in  the  different measurements we  present  in  this 

research, the quality of the nearest school as expressed by the relative exam score, does 

                                                       30 Since the difference in distance must be ≥ 0, we also carried out a Tobit regression. However, we opted for the OLS regression because we found no major discrepancies with the Tobit regression, probably because only around 10 % of individuals choose the nearest school. OLS tables may be easier to interpret, since Tobit regressions require an extra calculation of marginal effects. 

Page 123: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

97 

not demonstrate a significant effect on the distance difference. We added Appendix 4‐2, 

to  illustrate  that  the  exam  score  at  the  proximity  school  is  indeed  only  marginally 

associated with the distance difference, in the case of all ethnic groups.  

4.4.5 School choice and segregation 

 In  this  paragraph  we  take  a  closer  look  at  links  between  school  choice  and 

segregation. We have already seen that notably Turkish and Moroccan migrant students 

travel less far to school. In Table 4‐6 we consider the subset of pupils who do not choose 

the nearest  school  (14247  individuals,  88.7% of  the  total data  set),  and  compare  the 

characteristics  of  their  chosen  school  with  the  nearest  school  they  did  not  choose. 

Considerable  dissimilarities  appear  between  choice  patterns  of  native  Dutch  pupils, 

compared with pupils who have a migrant background. Native Dutch students, relative 

to  the not‐chosen nearest  school, choose a  school with  lower percentages of migrant 

pupils, a higher school average SES , a higher relative exam score, and a higher average 

CITO  score  intake.  In  contrast,  consider,  for  example  students  with  a  Moroccan 

background:  they  choose  a  school  with  a  higher  percentage  of  pupils  with  a  non‐

Western migrant background, about the same school SES index as the nearest school, a 

slightly better relative exam score and more often a broader secondary school, offering 

more  tracks. We  carried out  a paired  t‐test  and  found  that  the  characteristics of  the 

chosen more distant school  in all cases differed significantly  from those of the nearest 

school.  

The  different  choice  patterns  between  native  Dutch  and  migrant  pupils 

apparently  lead  to more  segregation,  and may  reveal  segregation  by  choice  among 

migrant parents. Only  in the case of native Dutch students we did find a school quality 

induced movement. 

We  carried  out  regression  analyses  for  these  different  choice  patterns,  and 

analysed  the difference between  the nearest and  the actual school  for all choosers of 

more  distant  schools,  concerning  four  different  indicators  at  school  level.  Note  that 

these  four  tables are based on different numbers of observations;  in  some  cases,  the 

information on all school quality indicators was not listed in our data set. We carried out 

tests  to  check  for  unobserved  selection mechanisms,  and  found  that  coefficients  on 

Page 124: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

98 

socio‐economic  variables  remained  stable  between  the  different  subsamples.  We 

analysed the following school characteristics: 

The school's mean exam score (Appendix 4‐3); 

The  school's  percentage  of  upward  mobility  ("pupward")  to  a  higher  track 

(Appendix 4‐4); 

The average SES of the school population (Table 4‐7); 

The school's percentage of pupils with a non‐western background (Table 4‐8). 

 

We found the most significant coefficients for the difference in school SES (Table 

4‐7)  and  school  percentage  of  pupils with  a  non‐Western migrant  background,  as  a 

further Specification of school SES (Table 4‐8). Tables on relative exam score and upward 

mobility have been included in the Appendices. 

Looking at the difference in the average school SES between the nearest and the 

actual school  (Table 4‐7), the  intercept  in Specification  (1)  indicates that the reference 

group  of  native  Dutch  gains  from  not  choosing  the  nearest  school.  This  outcome 

reiterates our finding that native Dutch pupils tend to opt for a school with less migrant 

pupils when their nearest school has many migrant pupils. Immigrant students gain less, 

and, in fact, Moroccan students do not gain SES status at all. Neighbourhood SES has no 

effect  on  the  gain,  but  the  percentage  of  non‐Western  pupils  at  the  nearest  school 

clearly  increases  the  gain,  by  some  0.15  SES  score  (about  one‐tenth  of  a  standard 

deviation)  for  a  10  percentage‐point  increase.  Weighted  student  funding  in  the 

elementary school also reduces the SES gain between the nearest and the actual school 

significantly.  Finally,  a  recommendation by  the primary  school  teacher  for one of  the 

academic secondary  tracks also  increases  the SES difference between  the nearest and 

the actual school significantly: the mean SES among pupils  in academic tracks tends to 

be higher than among pupils in vocational tracks. In The Hague and Rotterdam, the gain 

is substantially higher than in the other two cities. 

   

Page 125: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

99 

Table 4‐7: Regression results (OLS) for SES difference (i.e. average SES of chosen school minus average SES of nearest school), considering non‐nearest school choosers only (14197 individuals) 

(1) (2) (3) (4) (5)VARIABLES  SES diff SES diff SES diff SES diff SES diff   Distance difference to nearest school  0.03*** 0.06*** 0.05*** 0.05*** 0.05***    (0.010) (0.013) (0.010) (0.010) (0.013) male  0.01 0.01 0.01 0.01 0.02    (0.013) (0.013) (0.013) (0.016) (0.022) weighted student funding ‐0.16*** ‐0.15*** ‐0.12*** ‐0.16*** ‐0.17***    (0.048) (0.033) (0.029) (0.040) (0.051) Surinamese/Antillean  ‐0.20*** ‐0.26*** ‐0.18** ‐0.05 ‐0.06    (0.045) (0.041) (0.073) (0.073) (0.065) Turkish  ‐0.41*** ‐0.52*** ‐0.41*** ‐0.25*** ‐0.26***    (0.082) (0.057) (0.083) (0.084) (0.082) Moroccan  ‐0.59*** ‐0.58*** ‐0.46*** ‐0.31*** ‐0.28***    (0.091) (0.057) (0.088) (0.091) (0.090) Other Immigrant Background  ‐0.16*** ‐0.20*** ‐0.13** ‐0.06 ‐0.04    (0.050) (0.034) (0.050) (0.054) (0.058) one‐parent household  ‐0.12** ‐0.12*** ‐0.15*** ‐0.07* ‐0.09*    (0.045) (0.030) (0.028) (0.038) (0.048) neighbourhood SES index ‐0.02 0.08* 0.05 0.05    (0.054) (0.047) (0.053) (0.042) Non‐western x SES  0.00 0.01 0.02 0.04    (0.034) (0.031) (0.034) (0.035) urbanicity  0.36*** 0.33*** 0.22*** 0.17*    (0.110) (0.075) (0.075) (0.087) # of relevant schools within 5 km  ‐0.00 ‐0.00 0.00 ‐0.00    (0.005) (0.004) (0.005) (0.005) distance to nearest relevant school (km)  0.29** 0.17*** 0.15*** 0.13**    (0.112) (0.057) (0.056) (0.053) # of tracks offered at nearest school  ‐0.02 ‐0.02 ‐0.03    (0.038) (0.039) (0.043) % non‐western at nearest school  1.55*** 1.43*** 1.60***    (0.273) (0.296) (0.296) % non‐western at nearest school x non‐western ‐0.22* ‐0.22 ‐0.18    (0.124) (0.137) (0.226) Utrecht municipality dummy  0.03 0.11 0.10 0.08    (0.166) (0.116) (0.133) (0.127) The Hague municipality dummy  0.66*** 0.56*** 0.70*** 0.73***    (0.166) (0.112) (0.124) (0.125) Rotterdam municipality dummy  0.30** 0.38*** 0.41*** 0.49***    (0.134) (0.134) (0.142) (0.154) HAVO advice track dummy 0.25***   (0.057)HAVO/VWO advice track dummy  0.36***   (0.066)VWO advice track dummy 0.51***   (0.079)CITO test score  0.02***    (0.003) Non‐western x CITO test score  ‐0.00    (0.000) Constant  0.56*** ‐1.25*** ‐1.86*** ‐1.70*** ‐11.12***    (0.090) (0.474) (0.409) (0.392) (1.525)   Observations  14,197 14,147 14,147 7,912 4,145 R‐squared  0.08 0.18 0.35 0.37 0.43 Adj. R‐Squared  0.0747 0.181 0.344 0.371 0.427 

Notes: Robust clustered standard errors (at neighbourhood level) in parentheses *** p<0.01, ** p<0.05, * p<0.1 

   

Page 126: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

100 

Table 4‐8: Regression results (OLS) for the difference in the percentage of pupils with a non‐Western migrant background (i.e. percentage at the actual school minus the percentage at the nearest school), considering non‐nearest school choosers only (14197 individuals) 

(1) (2) (3) (4) (5)VARIABLES  % NW diff % NW diff % NW diff % NW diff % NW diff   Distance difference to nearest school  ‐1.70*** ‐1.99*** ‐1.30*** ‐1.77*** ‐1.03**    (0.372) (0.374) (0.289) (0.328) (0.417) male  0.05 0.10 0.39 0.73 0.14    (0.554) (0.529) (0.512) (0.670) (0.914) weighted student funding  8.90*** 8.62*** 9.26*** 10.06*** 7.11**    (2.119) (1.882) (1.567) (1.836) (2.998) Surinamese/Antillean  14.19*** 15.06*** 17.64*** 14.83*** 7.45***    (1.837) (1.603) (1.763) (1.835) (2.521) Turkish  24.26*** 26.39*** 26.87*** 22.34*** 17.81***    (3.309) (2.914) (2.698) (2.059) (3.307) Moroccan  27.48*** 28.51*** 27.18*** 22.23*** 16.84***    (3.513) (2.672) (2.524) (2.139) (3.325) Other Immigrant Background  10.44*** 10.66*** 11.15*** 8.94*** 3.89*    (1.537) (1.238) (1.288) (1.321) (2.103) one‐parent household  5.38*** 5.20*** 6.67*** 4.74*** 4.61**    (1.489) (1.281) (1.103) (1.355) (2.128) neighbourhood SES index  1.27 ‐4.49*** ‐3.13 ‐3.10*    (1.713) (1.590) (1.954) (1.871) Non‐western x SES  0.88 0.94 1.52 1.06    (1.450) (1.513) (1.546) (1.656) urbanicity  ‐3.29 1.74 4.81 5.68    (4.322) (3.123) (3.523) (3.760) # of relevant schools within 5 km  0.08 0.16 ‐0.01 0.16    (0.264) (0.213) (0.269) (0.235) distance to nearest relevant school (km)  ‐8.44* ‐5.05 ‐5.60 ‐3.50    (5.072) (3.818) (3.518) (3.267) # of tracks offered at nearest school  ‐3.48** ‐3.89*** ‐4.08***    (1.349) (1.451) (1.519) SES index at nearest school  21.34*** 22.66*** 24.76***    (2.282) (2.652) (2.631) SES index at nearest school x non‐western  1.14 ‐1.18 ‐0.79    (2.062) (1.865) (2.042) Utrecht municipality dummy  ‐2.21 ‐15.52** ‐17.86*** ‐19.07***    (8.542) (6.666) (6.415) (5.864) The Hague municipality dummy ‐10.35* ‐8.35* ‐9.30* ‐9.52**    (6.164) (4.341) (4.885) (4.524) Rotterdam municipality dummy 0.88 11.92** 11.33** 16.87***    (6.120) (5.076) (5.534) (5.781) HAVO advice track dummy  ‐10.91***   (2.381)HAVO/VWO advice track dummy  ‐13.43***   (2.948)VWO advice track dummy  ‐16.69***   (2.729)CITO test score  ‐0.65***    (0.117) Non‐western x CITO test score 0.01***    (0.005) Constant  ‐17.57*** 3.92 ‐7.36 ‐4.42 327.73***    (2.775) (17.828) (14.916) (16.965) (63.260)   Observations  14,197 14,147 14,147 7,912 4,145 R‐squared  0.14 0.18 0.38 0.41 0.47 Adj. R‐Squared  0.137 0.174 0.375 0.408 0.465 

Notes: Robust clustered standard errors (at neighbourhood level) in parentheses *** p<0.01, ** p<0.05, * p<0.1 

Page 127: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

101 

In Table 4‐8 we chose the difference between the percentages of pupils with a 

non‐western  immigrant  background  of  the  actual  minus  the  nearest  school  as  the 

outcome  variable.  As  indicated  by  the  intercept  in  Specification  (1),  native  Dutch 

reference  students, on  average,  travel  to  a  school with  lower percentages of migrant 

pupils. Turkish and Moroccan students travel to schools with higher percentages of non‐

Western immigrants, whereas the difference for the other immigrant groups is smaller. 

In  the  case  of Moroccan  pupils,  the  percentage  non‐western  pupils  increases  at  the 

actual school by 9.9 percentage points  (‐17.57 + 27.48), minus 1 percentage point per 

km  increase  in  distance  difference  (Specification  5).  Since  distances  to  school  are 

relatively  small  in  the  Netherlands  (Table  4‐1),  we  may  assume  that  the  different 

patterns found  in school choice between native Dutch and pupils of Moroccan descent 

are indeed the effect of different ethnic preferences. We found comparable patterns of 

choice between Turkish and Moroccan  students; also Surinamese/Antillean pupils and 

those  with  another migrant  background,  on  average,  prefer  a  school  with  a  higher 

percentage  of migrants,  but  to  a  lesser  degree  than  pupils  of  Turkish  and Moroccan 

descent.  Introducing the CITO test score (at the cost of substantially  lower numbers of 

observations) is also associated with a lower percentage of migrant pupils at the chosen 

school. These  findings  confirm our earlier outcomes: migrant pupils and  their parents 

appear to choose further segregation. 

4.4.6 Benefits of school choice?  

Upward  mobility  to  a  higher  track  may  be  a  particularly  important  extra 

opportunity  for migrant  pupils.  They  are  relatively more  often  the  first  one  in  their 

family with the option to enrol in academic secondary tracks, and may need extra time 

to  arrive  at  this  level  and  later  on  in  higher  education.  At  the  start  of  secondary 

education,  far  higher  percentages  of  migrant  pupils  enter  the  lower  secondary 

vocational track: e.g. 76 % of Moroccan youths, against 42 % of native Dutch youngsters 

(Table 4‐1). In fact, this  is the width of the educational achievement gap, which  indeed 

could be reduced by offering migrant students extra opportunities to improve their track 

level while  in secondary school. The Dutch educational system has two trajectories  for 

upward mobility:  1)  after  finishing one  track with  a  diploma,  pupils  can  re‐enrol  in  a 

higher track and obtain a second diploma at this higher level. 2) Pupils can also move up 

Page 128: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

102 

between grades, and finish with a diploma on a higher level than their level of entrance 

in Year 1. In Table 4‐9 we consider this second trajectory for upward mobility between 

grades. We  included only pupils who  started on  the  lower  vocational  track  in Year 1, 

since notably  this  lowest  level  leaves open  the possibility of opportunities  for upward 

transfer  to  a  higher  track. We  counted  only  individuals who  chose  the more  distant 

school,  in  order  to  be  able  to  assume  active  choice.  4343  individuals  in  our  data  set 

started  in  the vocational  track when  they entered secondary school. This number was 

further reduced to 2730  individuals when we  introduced characteristics of the nearest 

school (which are not in all cases listed in the data set). Other socio‐economic variables, 

however, remained stable in comparison with the total number of individuals in our data 

set.  

The coefficients  in Table 4‐9  show a  small, but  significant, negative correlation 

between a pupils' upward mobility and the distance difference to the nearest school: the 

predicted probability for a student to have moved up to a higher track in Year 3 reduces 

by 0.01 for every kilometre  increase  in the distance difference between the actual and 

the nearest school. Notably migrant pupils seem to  lose opportunities to move up to a 

higher  track, when  they  travel  further  to school. Furthermore,  the neighbourhood SES 

index is positively associated with upward mobility, indicating that pupils living in a more 

affluent  residential  area  also  have  a  higher  chance  to move  up  while  in  secondary 

education. The positive neighbourhood effect is reduced for immigrant pupils. 

Page 129: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

103 

Table 4‐9: OLS regressions of individual upward mobility at the entrance of Year 3, considering only pupils who started at the lower vocational secondary level in Year 1 (4343 individuals); non‐nearest school choosers only 

Non‐nearest choosing VMBO advice students (1) (2) (3)VARIABLES  level‐up in Year 3 level‐up in Year 3 level‐up in Year 3   Distance difference to nearest school  ‐0.01*** ‐0.01*** ‐0.01***    (0.002) (0.003) (0.003) male  ‐0.03*** ‐0.03*** ‐0.04***    (0.010) (0.010) (0.013) weighted student funding ‐0.06*** ‐0.06*** ‐0.06***    (0.014) (0.014) (0.018) Surinamese/Antillean  ‐0.06*** ‐0.05*** ‐0.05**    (0.017) (0.017) (0.020) Turkish  ‐0.06*** ‐0.04** ‐0.05*    (0.019) (0.018) (0.024) Moroccan  ‐0.04** ‐0.03 ‐0.03    (0.019) (0.018) (0.024) Other Immigrant Background  0.01 0.02 0.02    (0.018) (0.018) (0.023) one‐parent household  ‐0.03* ‐0.02 ‐0.01    (0.015) (0.014) (0.016) neighbourhood SES index 0.05*** 0.05***    (0.010) (0.013) Non‐western x SES  ‐0.02** ‐0.03***    (0.010) (0.012) urbanicity  ‐0.02 ‐0.01    (0.017) (0.023) # of relevant schools within 5 km  0.00* 0.00**    (0.001) (0.001) distance to nearest relevant school (km)  ‐0.03** ‐0.04**    (0.013) (0.017) relative exam scores nearest school  0.02    (0.033) # of tracks offered at nearest school  0.00    (0.006) % upward mobility at nearest school  ‐0.05    (0.086) average SES index at nearest school  0.01    (0.020) % non‐western at nearest school  ‐0.00    (0.034) Utrecht municipality dummy  ‐0.03 0.00    (0.025) (0.035) The Hague municipality dummy  ‐0.04** ‐0.00    (0.019) (0.022) Rotterdam municipality dummy  ‐0.04** ‐0.01    (0.019) (0.022) Constant  0.22*** 0.29*** 0.25**    (0.020) (0.075) (0.098)   Observations  4,343 4,325 2,730 R‐squared  0.03 0.04 0.05 Adj. R‐Squared  0.0253 0.0401 0.0462 

Notes: Robust clustered standard errors (at neighbourhood level) in parentheses *** p<0.01, ** p<0.05, * p<0.1 

   

Page 130: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

104 

4.5 Conclusions and discussion 

In  the  current  research we explored 1) how  the distance  travelled  to  school  is 

related  to  ethnicity  and  other  socio‐economic  variables  on  the  individual‐, 

neighbourhood‐ and school level; 2) the differences between the group that chooses the 

nearest school, and the group that prefers another school; 3) the effects of free school 

choice  on  ethnic  segregation;  and  4)  upward mobility  to  a  higher  track  as  a  possible 

benefit of school choice.  

4.5.1 Who travels further to school? 

All over, migrant pupils  travel  less distance  to  school  than native Dutch pupils, 

except for native Dutch pupils in affluent residential areas. Average distance to school is 

obviously  influenced  by  population  density  and  the  corresponding  larger  number  of 

schools  in densely populated areas. The number of schools to choose  from  in the  four 

Dutch major  cities  is,  for  example,  15  schools within  a  radius  of  5  km  in  the  case  of 

native Dutch students,  to 21 schools  in  the case of students of Turkish descent  (Table 

4‐1).  Since more  students with  a migrant  background  at  present  live  in  less  affluent, 

more densely‐populated neighbourhoods,  this may partly explain why  they  travel, on 

average, less far to school. However, urbanicity, according to our findings, cannot solely 

explain the lower mean distance travelled to school by migrant pupils.  

Students  of  Dutch  origin  travel  the  largest mean  absolute  distance,  3.49  km. 

However, the absolute distance to school is inversely correlated with residential SES: in 

the  poorest  residential  areas,  native  Dutch  youths  travel  further  than  Turkish  and 

Moroccan  youths, while  in  contrast,  the  average  distance  to  school  is  lower  among 

native  Dutch  students  in  affluent  areas.  This may  also  explain  our  findings  that  the 

elementary  teacher's  recommendation  for  the academic  secondary  track  is associated 

with  a  lower  distance  to  school:  substantially  higher  percentages  of  native  Dutch 

students  obtain  this  advice,  and  they more  often  live  in  an  affluent  area where  the 

preferred school may be nearby.  

4.5.2 Differential sorting, mobility increases segregation 

The literature predicts strong effects on ability sorting of extended school choice 

(e.g. Cullen et al., 2000). In the context of unlimited choice in the Netherlands, however, 

we  found,  somewhat  to our  surprise,  virtually no  systematic differences between  the 

Page 131: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

105 

group of  choosers  for  the nearest  school and  those who  chose a  school  further  from 

their home address. We did find, however, differential sorting when we compared the 

characteristics of  the nearest school and  the selected school,  that was not  induced by 

pupils' abilities, but related to the pupils' ethnicity: native Dutch pupils seem to weigh 

other aspects when selectively choosing a school at a further distance than  is the case 

among migrant  pupils. When migrant  students  prefer  a  school  further  outside  their 

residential area, they tend to choose a school with even higher percentages of migrant 

pupils than the neighbourhood school. Native Dutch pupils, on the other hand, tend to 

make a quality‐driven choice for a school with a higher mean exam score and a higher 

mean CITO intake. They prefer, furthermore, a school with a higher SES population and a 

lower  percentage  of migrant  students  than  the  nearest  school.  Remarkably, migrant 

students,  in  contrast, more  often  choose  schools with  higher  percentages  of migrant 

students,  than  the often already high percentages at  the nearest  school, while at  the 

same time, they show only a marginal preference for higher school quality.  In fact, we 

may have revealed here the segregation of migrant pupils by choice. Our findings bring 

to mind research by Bunar in Sweden (2010), who learned from interviews with migrant 

students  in  Stockholm  and  Malmö  how  important  the  sense  of  belonging  and 

recognition within the pupils' own ethnic group seems to be. 

4.5.3 Upward mobility, an important opportunity for migrant students  

Finally, although the effect we found is fairly small, we still consider as important 

the outcome  that  the slightly higher odds of migrant students  to move up  to a higher 

track  when  they  attend  a  school  closer  to  their  home.  In  the  domain  of  academic 

research  into  the  achievement  gap  between  the  children  of  low  SES  and  high  SES 

parents, the question whether disadvantaged pupils should be bussed to better schools, 

or better schools should be brought to poorer neighbourhoods, seems centre stage. Our 

findings rather support the second strategy. 

4.5.4 Policy implications and further research 

We plan to follow up on the current research, and investigate more in‐depth why 

migrant parents and pupils  seem  to prefer a  school with high percentages of migrant 

pupils. For the moment we reflect on two possible explanations: Over the years, schools 

with  high  percentages  of migrant  students  (up  to  90%),  have  largely  invested  in,  for 

Page 132: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

106 

example,  the  development  of  tailor‐made  strategies  to  support  first‐time  academic 

learners,  effective  cooperation  with  parents,  and  the  encouragement  of  upward 

mobility. There may be, therefore, reasons  for migrant parent to choose a school with 

large numbers of migrant children; 2) we cannot exclude, however, that migrant pupils 

tend to wish to avoid stigmatization in a school with predominantly native Dutch pupils, 

and prefer a school where they are the absolute majority.  In the Dutch context, as we 

described earlier, it is important to keep in mind that no financial concerns limit school 

choice  in  any  way;  neither  do  schools  with  lower  SES  pupils,  in  general,  have  less 

qualified  teachers,  or  struggle  with  a  high  turnover  of  teachers.  This  may  enhance 

opportunities  to  identify  other  motives  behind  school  choice,  for  example,  choices 

related to ethnicity, as mentioned above.  

Furthermore, we  reflected on our  finding  that  the average  school exam  scores 

only marginally drive school choice, even though this  information  is freely available on 

the internet (Inspectorate of Education). Since the average school exam score is strongly 

correlated to the average  level of schooling of parents,  it  is complicated for parents to 

identify  the  school's added value which best  fits  their  child. Supported by our  finding 

that  the  school's upward mobility does not drive  school choice whereas  this could be 

expectedly  an  important  element  in  the  closing of  the  achievement  gap– we  suggest 

that the Inspectorate of Education publishes on the Internet both exam results and the 

upward mobility of all schools, specified for the separate ethnic groups. More in general, 

informing migrant  parents  is,  according  to,  for  example  Douglas  Archbald  (2004),  a 

requirement for free school choice as a Liberation Model.  

We  contemplated  on  the  overall marginal  associations we  found with  school 

quality  indicators.  It  is  imaginable  that  almost  a  century  of  free  school  choice,  the 

absence of tuition fees, and the publication of the Inspectorate's assessment reports of 

all  schools  on  the  Internet  have  all  together  resulted  in  a  transparent,  fairly 

homogeneous school market. Out of about 650 secondary schools nationally, around 30 

schools  are  under  intensified  supervision  of  the  Inspectorate  because  of  severe 

underperformance. Currently only two of these 30 schools are located in the major four 

cities  (www.onderwijsinspectie.nl).  Moreover,  underperforming  schools  typically 

improve within  one  or  two  years,  under  intensified  supervision  of  the  Inspectorate. 

Taking this thought further, in the Netherlands, perhaps less so than in other countries, 

Page 133: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

107 

choosing one school or another may not result  in serious divergences  in opportunities 

for pupils.  

Finally, another aspect of school choice that may  inspire further research  is the 

fact that pupils do have an important say themselves in the choice of a secondary school 

in the Netherlands. They may, for example, want to go to the same secondary school as 

their  classmates  in  elementary  education,  they  may  feel  inspired  by  the  extra 

programmes  a  secondary  school  offers,  or  they  may  feel  attracted  to  a  new  well‐

equipped  school building.  Free  school  choice  and  the  absence of  tuition  fees make  it 

possible for pupils to follow such preferences. How pupil preferences influence patterns 

of choice and school results would be an interesting further research question. 

We look forward to be able to use the data on final exams in the coming years, as 

the BRON database is further developing. We acknowledge that assembling data sets for 

the purpose of academic research is time‐consuming, and that the government institute 

that collects these data (DUO) does currently not have the extra personnel to carry out 

this task. This has deepened our gratitude towards Cees Vermeulen, Erik Smits and Rob 

Kerstens, who have already invested their time and support. We do hope that our work 

has illustrated the high quality and the possibilities of the BRON data, and its potential to 

improve  the  quality  of  education  at  large,  through  a  better  understanding  of  these 

complex processes.  

4.6 Appendices 

In  the  general  overview  depicted  in  Appendix  4‐1, we  considered  the  higher 

aggregate  level  of  non‐Western  background  (includes  the  vast majority  of  pupils  of 

Turkish or Moroccan descent) versus Dutch background. The following maps give a first 

impression  of  the  association  between  the  ethnic  composition  of  an  area,  and  the 

distance  travelled  to  school.  For  example,  the  southern  part  of  the  city  centre  in 

Amsterdam has a large share of native Dutch inhabitants (only 0‐20% inhabitants with a 

migrant background), and the average distance travelled to school  in this area  is  lower 

(lightest shade of blue) than in other districts in Amsterdam.  

Some differences between the four cities can be observed. In Rotterdam and The 

Hague, neighbourhoods in the city centre more often have a population predominantly 

consisting of people with a non‐Western background  than  is  the  case  for Amsterdam 

Page 134: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

108 

and Utrecht. This relates to differences in city structure: in Amsterdam and Utrecht, top‐

quality residential areas are located in the heart of the city.  

 Interestingly, in Amsterdam, more so than in the three other cities, schools tend 

to  be  segregated  more  than  neighbourhoods:  20  out  of  84  secondary  schools  in 

Amsterdam  enrolled  between  80  and  100  %  pupils  with  a  non‐Western  migrant 

background,  whereas  no  neighbourhood  in  Amsterdam  has  this  high  percentage  of 

inhabitants with a migrant background; the degree of segregation of these 20 schools, 

exceeds  the  degree  of  segregation  of  the  surrounding  neighbourhood.  While  in 

Rotterdam  the  variety  in  ethnic  composition  of  neighbourhoods  reflects  the  ethnic 

composition  of  schools more  closely,  also  in  this  city  21  out  of  113  schools  enrolled 

between 80‐ 100% of pupils of non‐ Western migrant descent. In comparison, The Hague 

and Utrecht  seem  to  be  less  segregated,  although  The Hague  is  the  only  city with  6 

neighbourhoods  in the city‐centre where between 80‐100% of  inhabitants have a non‐

Western immigrant background.  

We also depicted the schools outside the city boundaries which were attended 

by pupils  living within  the city. There are  remarkable differences  in  the percentage of 

pupils attending a school outside the city boundary among the four cities: 

Amsterdam:  230/5153=4.4% 

Rotterdam:  964/4853=19.8% 

The Hague:  791/4000=19.7% 

Utrecht:  639/2065=30.9% 

 

In  the  current  chapter  we  did  not  investigate  further  the  causes  for  these 

differences. 

4.6.1 Explanation of variables 

Average  SES  index  school:  the  average  SES  of  (the  parents  of)  children  at  a 

particular school. 

CITO‐score:  final test  in elementary education; scores range between 500‐ 550, 

the national average is 535. 

HAVO,  HAVO/VWO  and  VWO  advice  track  dummy:  dummy‐coded  variable 

relative to VMBO (lowest vocational track). 

Page 135: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

109 

Individual upward mobility: pupil who moves up to a higher secondary track.  

Municipality dummy‐coded  fixed effects of Utrecht, Rotterdam and The Hague, 

relative to Amsterdam. 

Nearest  relevant school:  the nearest school  to  the pupils'  residence  that offers 

the pupils' chosen track. 

Non‐Western immigrant: The following definitions by Statistics Netherlands (CBS, 

the National Statistics Office, www.cbs.nl) have been used: Western  immigrant: 

someone  originating  from  a  country  in  Europe  (exclusive  of  Turkey),  North 

America,  Oceania,  Indonesia  or  Japan.  Non‐western  immigrant:  someone 

originating  from  Africa  (including Morocco),  South  America,  Asia  (exclusive  of 

Indonesia and Japan) or Turkey. 

Other  immigrant  background:  a  wide  range  of  nationalities,  including  highly 

educated workers,  refugees and people  seeking asylum. The variance  in SES  is 

large, accordingly. 

Outside municipality border: pupils, who live within city boundaries, but attend a 

school outside the city. 

Relative  exam  score:  the  performance  per  track  level, measured  by  the mean 

exit‐exam score, compared with the mean exit‐exam score of all other schools in 

the four major cities offering the same track level.  

Relative  neighbourhood  SES  index:  Neighbourhood  SES  score  relative  to  the 

average in the four major cities; SES scores have been centred on zero, ‐4 (poor) 

to +4 (affluent). 

Secondary  school advice  score: advice by  the elementary  teacher about which 

track to follow in secondary education. 

Tracks offered at secondary school: the number of tracks (1 ‐ 4 main tracks) may 

differ  per  school.  Comprehensive  schools  offer  tracks  on  the  vocational  and 

academic  levels.  Schools may  also  offer  either  vocational  or  academic  tracks; 

schools may be diversified further and offer only one academic track.  

Upward mobility  at  school  level:  the  percentage  of  pupils  at  school  level who 

move up to a higher secondary track. 

Page 136: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

110 

Urbanicity:  a measure  for human  activity  in  an  area, based on  the number of 

addresses per km². 

Weighted student  funding:  in primary schools children qualify  for extra  funding 

depending on the educational attainment  level of the parents; the pupil weight 

for  funding  is  0  for  better‐educated  parents  (means  only  standard  funding); 

children  of  less  well‐educated  parents may  receive  0.3  to  1.2  extra  funding, 

additional to standard funding.  

Appendix 4‐1: Maps of Amsterdam, Rotterdam, Utrecht and The Hague.  

In  shades  of  blue,  the  average  distance  travelled  to  secondary  school  on  the 

neighbourhood  level;  in  shades of  red,  the percentage of non‐Western  immigrants  at 

neighbourhood  level.  In parentheses, the number of neighbourhoods sharing the same 

percentage of non‐Western immigrants, and the number of neighbourhoods sharing the 

same average distance  travelled  to school,  respectively. The cities are not depicted  to 

scale, but have been formatted to fit in this text. The ranking order of size is Amsterdam 

(largest), Rotterdam, The Hague, Utrecht. 

Page 137: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

111 

 

Page 138: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

112 

 

Page 139: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

113 

 

Page 140: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

114 

 

Page 141: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

115 

Appendix 4‐2: LOWESS regression (Locally Weighted Scatterplot Smoothing; bandwidth = .6) and linear regression (OLS) of the extra distance travelled to the preferred school (as compared with the nearest school) and school's mean final exam score (national standardized exam score) 

 

   

Page 142: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

116 

Appendix 4‐3: Regression results (OLS) for difference in the mean relative exam score at school level (i.e. the actual school exam score minus exam score at the nearest school), considering non‐nearest school choosers only 

(1) (2) (3) (4) (5)VARIABLES  exam diff exam diff exam diff exam diff exam diff   Distance difference to nearest school  0.00 0.01* 0.01 0.00 0.00    (0.005) (0.005) (0.005) (0.005) (0.006) male  ‐0.01 ‐0.01 ‐0.01 ‐0.02* 0.00    (0.009) (0.009) (0.008) (0.011) (0.014) weighted student funding  ‐0.01 ‐0.04* ‐0.03 ‐0.05** ‐0.04    (0.032) (0.022) (0.021) (0.026) (0.040) Surinamese/Antillean  ‐0.05 ‐0.08*** ‐0.01 ‐0.04 ‐0.00    (0.033) (0.027) (0.048) (0.042) (0.050) Turkish  ‐0.04 ‐0.13*** ‐0.04 ‐0.07 ‐0.08    (0.060) (0.043) (0.065) (0.048) (0.064) Moroccan  ‐0.09 ‐0.16*** ‐0.07 ‐0.08* ‐0.06    (0.058) (0.042) (0.067) (0.045) (0.063) Other Immigrant Background  ‐0.05* ‐0.06*** ‐0.02 ‐0.03 ‐0.01    (0.027) (0.024) (0.033) (0.031) (0.040) 2nd generation non‐Western Immigrant  0.03 0.04** 0.05*** 0.03 0.02    (0.019) (0.018) (0.017) (0.020) (0.029) one‐parent household  ‐0.04 ‐0.05** ‐0.06*** ‐0.03 ‐0.05    (0.028) (0.020) (0.021) (0.020) (0.033) neighbourhood SES index  ‐0.06** ‐0.02 ‐0.02 ‐0.01    (0.021) (0.026) (0.025) (0.027) Non‐western x SES  ‐0.00 ‐0.01 ‐0.01 0.01    (0.018) (0.018) (0.020) (0.023) urbanicity  ‐0.05 ‐0.05 ‐0.03 ‐0.05    (0.052) (0.052) (0.053) (0.062) # of relevant schools within 5 km  0.00* 0.01* 0.00 0.00    (0.003) (0.003) (0.003) (0.003) distance to nearest relevant school (km)  0.01 ‐0.02 ‐0.03 ‐0.03    (0.040) (0.036) (0.034) (0.036) # of tracks offered at nearest school  0.02 0.00 0.00    (0.022) (0.020) (0.027) average SES index at nearest school  ‐0.02 ‐0.07 ‐0.06    (0.059) (0.061) (0.071) % non‐western at nearest school 0.40** 0.33** 0.35*    (0.159) (0.164) (0.193) % non‐western at nearest school x non‐western  ‐0.17** ‐0.09 0.00    (0.073) (0.063) (0.115) Utrecht municipality dummy  0.28*** 0.32*** 0.29*** 0.27***    (0.094) (0.083) (0.082) (0.082) The Hague municipality dummy 0.28*** 0.27*** 0.27*** 0.20***    (0.055) (0.056) (0.059) (0.060) Rotterdam municipality dummy 0.11* 0.13* 0.12* 0.10    (0.067) (0.065) (0.077) HAVO advice track dummy  0.01   (0.032)HAVO/VWO advice track dummy  0.02   (0.034)VWO advice track dummy  0.19***   (0.041)CITO test score  0.00**    (0.002) Non‐western x CITO test score ‐0.00    (0.064) (0.000) Constant  0.11*** 0.08 ‐0.18 ‐0.15 ‐2.26**    (0.030) (0.236) (0.285) (0.289) (0.985)   Observations  8,503 8,471 8,471 4,720 2,524 R‐squared  0.01 0.09 0.16 0.18 0.17 Adj. R‐Squared  0.00420 0.0924 0.153 0.173 0.158 

Notes: Robust clustered standard errors (at neighbourhood level) in parentheses *** p<0.01, ** p<0.05, * p<0.1 

   

Page 143: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

4 Patterns in secondary school selection in the context of unlimited choice 

117 

Appendix 4‐4: Regression results (OLS) for difference in school percentage of upward mobility to a higher track (i.e. upward mobility at the actual school minus upward mobility at the nearest school), considering non‐nearest school choosers only 

(1) (2) (3) (4) (5)VARIABLES  pupward diff pupward diff pupward diff pupward diff  pupward diff  Distance difference to nearest school  0.16 0.28 0.36* 0.11  0.27   (0.202) (0.207) (0.184) (0.179)  (0.221)male  0.27 0.20 0.27 0.20  ‐0.11   (0.326) (0.316) (0.296) (0.358)  (0.459)weighted student funding ‐0.78 0.37 0.33 ‐0.60  ‐2.39   (0.886) (0.777) (0.736) (0.917)  (1.603)Surinamese/Antillean  1.34 2.81** 1.95 ‐0.49  ‐0.80   (1.186) (1.083) (1.940) (1.750)  (1.925)Turkish  1.10 3.19** 2.06 0.14  ‐0.63   (1.937) (1.538) (2.269) (1.738)  (2.223)Moroccan  4.07* 6.20*** 4.82** 2.62  1.64   (2.091) (1.571) (2.318) (1.625)  (2.067)Other Immigrant Background  0.27 1.15 0.56 ‐0.12  ‐2.11   (0.917) (0.843) (1.189) (1.178)  (1.321)2nd generation non‐Western Immigrant  0.40 0.32 0.14 0.02  1.29   (0.704) (0.643) (0.644) (0.828)  (1.113)one‐parent household  1.36* 2.11*** 2.44*** 0.33  0.77   (0.735) (0.622) (0.646) (0.628)  (0.870)neighbourhood SES index 1.98*** 0.42 1.18  1.47*   (0.699) (0.834) (0.829)  (0.868)Non‐western x SES  ‐0.57 ‐0.33 ‐0.43  ‐0.85   (0.627) (0.639) (0.843)  (1.022)urbanicity  0.60 0.80 0.61  0.48   (2.023) (2.185) (2.260)  (2.376)# of relevant schools within 5 km  0.00 0.01 ‐0.06  ‐0.09   (0.108) (0.106) (0.118)  (0.128)distance to nearest relevant school (km)  ‐0.89 ‐0.15 ‐0.14  ‐0.64   (1.976) (1.743) (1.622)  (1.613)# of tracks offered at nearest school  ‐1.34 ‐1.77**  ‐1.27   (0.889) (0.823)  (1.060)average SES index at nearest school  2.68 2.21  1.84   (1.962) (2.274)  (2.419)% non‐western at nearest school  ‐6.55 ‐7.40  ‐12.29*   (5.234) (6.366)  (7.236)% non‐western at nearest school x non‐western 2.72 5.44***  8.00*   (2.646) (1.974)  (4.692)Utrecht municipality dummy  ‐2.50 ‐4.16 ‐7.06**  ‐6.55**   (4.066) (3.373) (3.342)  (2.918)The Hague municipality dummy  0.10 0.20 ‐1.24  1.57   (2.178) (2.166) (2.236)  (2.630)Rotterdam municipality dummy  ‐1.11 ‐0.29 ‐0.59  ‐1.35   (2.415) (2.459) (2.570)  (2.787)HAVO advice track dummy ‐3.16**    (1.272) HAVO/VWO advice track dummy  ‐3.54***    (1.334) VWO advice track dummy ‐8.34***    (1.422) CITO test score  ‐0.31***   (0.104)Non‐western x CITO test score  ‐0.00   (0.006)Constant  ‐3.45*** ‐6.16 0.45 7.96  170.51***   (1.309) (9.372) (10.264) (10.023)  (54.278)  Observations  9,025 8,997 8,997 5,046  2,709R‐squared  0.01 0.04 0.10 0.14  0.18Adj. R‐Squared  0.00967 0.0396 0.0957 0.134  0.170

Notes: Robust clustered standard errors (at neighbourhood level) in parentheses *** p<0.01, ** p<0.05, * p<0.1 

Page 144: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit
Page 145: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

119 

5 A tailor made transfer of scientific 

knowledge to school practice 

But despite the fact that natural knowledge is divine, its 

practitioners cannot be called prophets. For other men may 

discern and embrace what they teach with as much certainty and 

entitlement as they do themselves. They do not just accept it on 

faith. 

   Baruch de Spinoza, 1670 

5.1 Introduction  

The  importance  of  scientific  research  for  school  practice  is  generally 

acknowledged  by  teachers  and  school  leaders,  educational  researchers,  and  policy 

makers.  Educational  policies  at  the  government  level,  furthermore,  seem  increasingly 

supported by academic research (e.g. the No Child Left Behind Act, U.S. Congress, 2001). 

So far, however, the effects of research findings on school outcomes seem disappointing 

(Slavin, 2002). The academic  literature  in the domain of the utilization of knowledge  in 

the practice of schools reports very limited transfer of knowledge to schools.  

                                                        This chapter is based on: Van Welie, E.A.A.M., Borghans, L. and De Wolf, I. (2013). Using research outcomes in school practice. Working paper. 

Page 146: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

120 

Ideally, the process to connect school questions with existing research outcomes 

should  include 1)  the articulation of  school questions; 2)  the  search  for a match with 

research  findings;  3)  a  knowledge based  intervention  in  school  processes;  and  4)  the 

analyses of the effects on school outcomes.  

The aim of  this paper was  to  investigate what obstacles hinder  the process of 

knowledge transfer, and how the application of scientific knowledge in schools could be 

enhanced.  So  far  the  limited  use  of  scientific  knowledge  in  schools  appears  to  be  a 

stubborn problem. We  therefore  chose  to  carry out a detailed  study  in  six  secondary 

schools 

We  carried  out  in‐depth  semi‐structured  interviews with  Principals  to  identify 

and articulate school questions. Next, we carried out a literature review in order to link 

these questions  to  state‐of‐the‐art  research publications. Together with  the  six  school 

leaders, we  concluded  this explorative  study with an evaluation, again by means of a 

structured  personal  interview,  of  the  applicability  in  school  practice  of  a  tailor‐made 

selection of research publications.  

In 2011 two reports were published in the Netherlands on the importance of an 

enhanced transfer of knowledge from educational research to an effective utilization in 

schools  (Education  Council  of  the  Netherlands,  2011;  Commissie  Nationaal  Plan 

Toekomst  Onderwijswetenschappen  [Committee  for  the  National  Plan  for  the 

Educational  Sciences], 2011).  Inspired by  these  reports,  the present  study  sets out  to 

investigate what the nature of school questions is, and whether these questions can be 

linked to relevant, high‐quality, existing research.  

We carried out our  research  in secondary schools with a diverse student body, 

including  at  least  30%  of  pupils with  a migrant  background. We  focused  on  themes 

related  to  the  quality  of  school  outcomes  and  equal  opportunities  for  all  pupils. We 

made  this choice  for diverse schools because  the attempt to understand  the stubborn 

and persistent achievement gap between migrant pupils and native‐born youths, and, 

more  in  general,  between  low  SES  (socio‐economic  background)  children  and more 

affluent youth,  is an ambitious policy goal both  in Europe and  in  the U.S. A wealth of 

research outcomes is published yearly on the achievement gap in education, and it may 

be assumed that researchers in the field of education implicitly expect that schools could 

Page 147: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

121 

benefit from the results of their work, when these results actually would be translated 

to school strategies.  

Previous  research  is described  in Section 5.2. We describe our methodology  in 

Section 5.3; the six schools are portrayed in Section 5.4; the findings are listed in Section 

5.5 and the conclusions in Section 5.6. The findings support our assumption that schools 

need  to  be  understood  in  their  specific  context,  and  that  a  timely match  between 

question and answer seems a prerequisite. The outcomes show that school  leaders do 

indeed have a demand for a knowledge base for their strategic themes; the majority of 

queries could effectively be linked to research publications. In five out of six schools, the 

selected  scientific  insights  have  been  used  to  guide  new  school  strategies,  support 

existing programmes, or  inform discussions with  the  school board and  local  (political) 

governors.  Based  on  this  project,  school  leaders  recommend  the  formation  of  an 

intermediate function between academe and secondary schools.  

5.2 Previous research 

Different kinds of knowledge 

According to McIntyre (2006), educational research and school practice are "two 

sharply contrasting kinds of knowledge…that are at the opposite ends of the spectrum": 

researchers abstract ideas from complex realities, while teachers simultaneously work in 

many dimensions:  the  concrete  setting of  their  class, but also  the  combined  levels of 

school,  parents,  local  stakeholders,  the  curriculum,  and  the many  unpredictable  daily 

events  in a  large  school. McIntyre  (ibid.)  suggests  that  the  transfer of knowledge vice 

versa between  these  two different worlds might benefit  from  the work of researchers 

who  generate  general  research‐based  suggestions  for  practice  in  their  publications, 

discuss  possible  implications  of  their  research  with  teachers,  and  cooperate  in 

monitoring the effects on outcomes of knowledge‐based interventions.  

Landry, Amara and Lamari (2001),  instead of comparing the different categories 

of knowledge in academe and in schools, focus on the question whether academe raises 

barriers,  through  its  culture  and  organization,  and  notably  in  terms  of  additional 

transition  costs,  against  the  dissemination  of  the  results  of  scientific  research  to 

classroom practice. To start with, the authors acknowledge that dissemination efforts at 

best play a minor role in the evaluation of the quality of research; this expectedly does 

Page 148: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

122 

not stimulate the transfer of research to classrooms. Based on their research, Landry et 

al. (ibid.) describe a  list of predictions on how the transmission of knowledge might be 

improved.  Several  of  these  predictions  seem  especially  relevant  for  the  present 

research:  1)  educational  research  projects  should  be  customized  to  applicability  in 

practice,  and  correspond with  the needs  of  teachers;  2)  explanatory  variables  should 

reflect the user's context, thereby adding to the credibility of research; 3) the utilization 

of research possibly may  increase when results are pertinent and reach schools at the 

right  time,  coinciding with  current  school  questions;  4) more  easily  readable  reports 

with specific recommendations for practice could  improve the applicability of research, 

and,  perhaps  even  more  important,  "factors  such  as…efforts  to  focus  on  variables 

amenable to intervention by users", could be expected to bridge the worlds of academe 

and practice.  Interestingly, Landry et al. (ibid.)  introduce the concept of a new feature: 

the linkage mechanism (i.e. the translation of research outcomes to practice as a distinct 

expertise). With reference to the power of context an important finding of Landry et al. 

(ibid.) should be mentioned here: taking  into account the user's context, turned out to 

be a far stronger variable than the user's needs. The recommendation by Landry et al. 

(ibid.) to  invest  in the further development of skills that are required to  identify school 

themes,  and  to  explain  the  applicability  of  research  findings,  have  inspired  the 

methodology we describe in the current study.  

In  the  Netherlands,  a  successful  linkage  mechanism,  albeit  in  the  field  of 

agriculture,  started as early as 1876, when  the predecessor of Wageningen University 

and Research Centre was founded. Almost from its beginning, in 1877 the first regionally 

based  testing  stations  were  set  up,  that  brought  workers  in  agriculture  and  cattle 

breeding in contact with scientific researchers. Largely supported by the opportunities of 

the Internet, to this day Wageningen is firmly associated with institutions and individual 

workers in the domains of food production, the environment, and health. Their website 

offers up to date reports and research outcomes considering current themes in the field 

(for  example  http://www.livestockresearch.wur.nl),  and,  through  a  web‐based 

information centre, workers  in  the  field can submit  research questions emerging  from 

practice,  when  no  financial  means  for  initiating  research  are  available  to  them. 

Admittedly,  it  is  a  complicating  factor  that  educational  research  is  carried  out  in 

different scientific disciplines, not organized  in one specific university, as  is the case  in 

Page 149: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

123 

Wageningen. We feel, however, that the example of Wageningen University might offer 

a valuable example for education, as will be elaborated at the end of this paper. 

Furlong (2004) adds to the goal of a more effective validation of knowledge, the 

noteworthy caution not to get involved again in "epistemological divisions and paradigm 

wars",  referring  to  the  constant  emphasis  in  political  and  policy  debates  on  proven 

"what  works"  research,  and  the  strong  priority  for  controlled  randomized  trials  in 

research design. In Furlong's view, researchers should consider the work of Schön (1983) 

on the "reflective practitioner", and invest in connecting researchers with professionals 

in schools while acknowledging the teacher's own subject knowledge and judgment. The 

critical assessment of academic work, furthermore, is regarded to be a necessary specific 

skill that should be embedded in on‐the‐job training of teachers in order to equip them 

better  for  research  utilization.  Furlong  (ibid.)  also  advocates  that  researchers  and 

practitioners should work together directly, in order to make research work in schools. 

Reciprocal views of educational practitioners and academic researchers  

Vanderlinde  and  Braak  (2009)  investigated  in  Flanders  the  opinions  that 

teachers,  school  leaders,  researchers  and  intermediaries  hold  of  one  another's  roles, 

with  regard  to  the  gap  between  research  and  practice.  The  outcomes  of  their  focus 

group  interviews  indicate  that  particularly  teachers  are  sceptical  about  the  value  of 

research  for  their  classroom  practice,  since  they  feel  that  research  does  not  offer 

enough practical applicability. School  leaders, however, report that they do,  in fact, try 

to  use  research  findings,  notably  for  themes  concerning  the  school  organization. 

Researchers expressed that their first and foremost concern was the assessment of their 

work by  the academic community, and  that  they  lack  the  time and  incentive  to make 

their work readable to a wider public, although they recognize that the use of technical 

language may form a hindrance for practitioners. Not surprisingly, intermediaries did not 

experience  the distance between  research and practice  to be a problem, but  rather a 

challenge.  Arguably,  however,  the  opinion  of  the  researchers  in  the  focus  group,  to 

delegate the transmission of knowledge to intermediaries, could be interpreted either as 

an innovative idea or as yet a further seclusion of scholars in academe.  

Page 150: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

124 

Intermediate actors 

Nutley, Walter  and  Davies  (2002)  take  the  epistemological  discussion  on  the 

preferred design of educational research further, in distinguishing the instrumental from 

conceptual use of  research. By  their nature,  results  from  controlled  randomized  trials 

might  more  often  be  instrumental,  since  preferably  specific  discrete  practices  are 

compared  in  these experiments. Bringing  into mind, however,  the complex and highly 

contextualized  characteristics  of  schools,  also  conceptual  insights  from  research  (for 

example  the  role and  function of  the  school amidst  societal  themes  concerning equal 

opportunities) might be appropriate and helpful. Nutley et al. (ibid.) refer specifically to 

the  significance  of  a  dissemination  medium  for  linking  research  to  schools.  They 

emphasize, furthermore, the  importance of the timeliness and accessibility of research 

findings, and – not surprisingly– the flexibility, reliability and credibility of research. 

Research to support school reform and evaluation 

While  the  current  paper  wants  to  elucidate  how  the  transfer  of  knowledge 

between  research  and  individual  schools  could  be  improved,  the  research  base  for 

government policies, however, should also be mentioned here. The assumption  is, that 

the  knowledge  underlying  government  policies  should  preferably  be  accessible  to 

schools,  in  order  to  enhance  the  acceptance  and  implementation  of  general  policies. 

Slavin  (2002), when reflecting on the enormously  increased technical potential of data 

analyses,  speaks  of  a  "scientific  revolution  that  has  the  potential  to  profoundly 

transform policy, practice and research". Yet, he records the fact that large‐scale reform 

programmes in the U.S., for example the No Child Left Behind Act (U.S. Congress, 2001)– 

albeit such programmes may need a  longer period of time to demonstrate outcomes– 

have  so  far  failed  to demonstrate  significantly  improved educational outcomes at  the 

system  level.  In  the Netherlands, heads of  school and  teachers express doubts on  the 

knowledge base of government reform programmes, and doubt sometimes whether the 

feasibility  of  translating  reform  programmes  to  school  practice  per  se,  has  been  the 

subject of research at all. The Education Council of the Netherlands (2011), furthermore, 

states that the tendency seems to be that  large‐scale educational reforms have not so 

demonstrated the expected results. In his inaugural lecture of 2001, Hessel Oosterbeek 

already expressed his concern that teachers' motivation might erode, when they have to 

Page 151: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

125 

comply with new government policies and  then never  learn about  the effects of  such 

policies. Slavin  (ibid.) critically notices  that  too often government programmes are not 

based  on  evidence  of  actual  effectiveness  in  practice.  Slavin  (ibid.),  furthermore, 

remarks critically that research can be found to support virtually any point of view given 

a  specific  theme  in  a  school;  contradictory  research  outcomes  considering  the  same 

theme, may severely hamper the use of academic knowledge.  

Even  apart  from  doubts  on  an  adequate  research  base  for  national  school 

reforms, Morell and Noguera (2011) criticize government policies for neglecting the fact 

that  some  schools,  for  example,  those  with  many  low  SES  or  migrant  pupils, 

demonstrate excellent outcomes, while others do not. National policies tend, however, 

to address all such schools equally as disadvantaged. Worldwide, schools struggle with 

context  specific  obstacles,  when  trying  to  implement  national  reform  programmes. 

Morell and Noguera (ibid.) convincingly explain how long term meaningful partnerships 

between research and practice, leading to joint research agenda setting, research based 

interventions  in  school  practice  and  subsequent monitoring  of  new  strategies, might 

enhance  the  effectiveness  of  school  reform.  Research  insights  may  be  applied  in 

different ways in schools. Johnson (1998) distinguishes between instrumental (basis for 

action),  conceptual  (offering  insights), process  (learning about  current processes), and 

symbolic (e.g. in political discussions) use of evaluations.  

Finally, the Education Council of the Netherlands (2011) recommends three main 

strategies for a more effective transmission of research to practice:  

1. The  formation  of  networks  consisting  of  schools,  universities,  and  centres  for 

education,  by  analogy with medical  schools  in  academe.  In  addition,  at  each 

school several teachers should develop competences for analysing their results in 

a more scientific way, in order to be able to study the effect of interventions. 

2. The  foundation  of  a  new  agency,  consisting  of  representatives  from  schools, 

academe  and  the  government,  which  coordinates  agenda‐setting  for 

fundamental, practice‐based, and policy‐oriented educational research.  

3. Considering  the  ambition  that  all  schools,  in  the  future,  should  invest  more 

systematically in quality improvement, the Council proposes that the Ministry of 

Education should allow schools sufficient free rein, in order to develop methods 

that fit the  individual school best. The government, furthermore, should ensure 

Page 152: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

126 

that school  reforms are  introduced  in a way  that can be monitored effectively, 

notably by the schools themselves. 

 

Inspired  by  the  academic work  described  in  this  Section, we  have  set  out  to 

investigate  in  the  current  study, what obstacles hamper  the use of existing  academic 

knowledge in school practice and how this process could be organized more effectively. 

Based on  the  reviewed  literature and our own experiences, we had  two assumptions. 

First, we assumed that schools are to a considerable extent  influenced by their specific 

context, and that this context should strictly be taken into account in order to make an 

accurate match with relevant research publications; second, we assumed that research 

could work  in schools when there  is a timely match with existing school questions (e.g. 

Landry et al., 2001; Morell and Noguera, 2011). The method we designed and tested was 

based  on  these  two  presuppositions.  Through  in‐depth  interviews we  collected main 

school questions,  and  carried out  a  literature  review  for every question  in  search  for 

state‐of‐the‐art publications  to match  these school  themes. Furthermore, we analysed 

the nature of school questions and the different categories of academic knowledge that 

supported  best  school  practice.  In  this  way,  we  set  out  to  achieve  a  deeper 

understanding  of  concrete  intermediary  practices  that  actually  link  in‐depth  explored 

school questions with adequate research outcomes. Finally, we  took  into account  that 

our method  should  allow  scaling  up  to  larger  numbers  of  schools, within  reasonable 

costs.  

5.3 Methodology 

We based our methodology on the following research questions: 

1. Do school leaders have significant questions that demand for academic research 

outcomes as a guidance?  

2. Can school questions be linked to manageable, high quality outcomes of existing 

research?  

3. What  is  the  nature  of  school  questions,  and  what  categories  of  academic 

knowledge can best be matched with these queries?  

4. Can  selected  research  findings  subsequently  be  applied  in  school  practice  and 

strategies? 

Page 153: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

127 

5. Can specific impediments for the transfer of applicable knowledge to schools be 

identified? 

6. What  are  the  requirements  for  an  intermediate  practice  that  could  bridge 

schools and academe satisfactory?  

 

Below we will describe the interventions we carried out, and motivate our choice 

to  focus  on  the  role  of  the  Principal  in  the  transfer  and  use  of  knowledge. We will 

describe the method of semi‐structured interviews we used, as well as our procedure for 

establishing  the  match  between  school  questions  and  applicable  research  findings. 

Finally, we will describe the format of the evaluative interview. 

5.3.1 Six participating school leaders 

We choose  to  focus on six schools because we wanted: a) a detailed  insight  in 

the present use– or the lack thereof– of scientific knowledge in individual schools; while 

b) we wanted to be able to compare schools and find possible patterns; and c) to have a 

sufficiently  large  scale  in  terms of  the number of pupils  served by  these  schools. We 

carried out our study in a real‐life context, which was another important reason why we 

chose this design (Yin, 2009). In an attempt to make participating schools not too diverse 

in  too many  dimensions, we  invited  only  schools  in  Amsterdam31  to  participate  (see 

Appendix 5‐1 for a general description of the participating schools). 

All  six  schools offer  the  two highest  tracks  in  the Dutch  system  for  secondary 

education,  referred  to  as  "academic  tracks"  in  this  chapter,  since  completion  grants 

access to higher education.32 Five schools also offer pre‐vocational tracks that qualify for 

senior vocational education. We labelled schools A‐F in random order. 

                                                       31 Calandlyceum, Comenius Lyceum, Hervormd Lyceum West, Open Scholengemeenschap Bijlmer, Spinozalyceum, IJburg College. 32 Official translations of the Dutch educational system in ISCED classification (International Standard Classification of Education by UNESCO, update1997): 

VMBO: pre‐vocational secondary education, 4 years, ISCED 2, qualifying for senior secondary vocational education 

HAVO: senior general education, 5 years, grade 1‐3 ISCED 2, grade 4‐5 ISCED 3, qualifying for higher education 

VWO: pre‐university education, 6 years, grade 1‐3 ISCED 2, grade 4‐6 ISCED 3, qualifying for higher education 

MBO: senior secondary vocational education, level 1 ISCED 2, level 2‐4 ISCED 3, level 4 qualifying for higher education  

HBO: universities for applied sciences, ISCED 5B 

Page 154: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

128 

In  the  case  of  four  schools  (A,  B,  D,  F),  the  vast majority  of  students  have  a 

migrant background; most of these students were themselves born  in the Netherlands, 

but have at  least one parent born abroad. Three of these four schools (A, D, F) mainly 

enrol  students with  a  Turkish or Moroccan background. One  school  (B) enrols mostly 

students from Suriname, the Antilles, and Ghana. School E attracts many students from 

a well‐educated population in a recently built neighbourhood, as well as students with a 

migrant  background with  parents who  had  fewer  chances  to  receive  good  schooling. 

School C has a currently changing composition of students and enrols a growing share of 

children with middle  class parents, while  the  share of migrant  students  is decreasing. 

The ambition of school C remains, however, to attract students who mirror the diverse 

multi‐ethnic population of Amsterdam. 

The  total number of pupils enrolled  in  the  selected  schools  (school year 2011‐

2012) is 6335. 1765 pupils are enrolled in the largest school, 353 in the smallest one. The 

two smaller schools in our study are growing substantially in student numbers.  

In some  important aspects the six schools may be characterized as schools that 

invest  largely  in "Systemic Equity"  (Scott, 2001; Brown, Benkovitz, Muttillo and Urban, 

2011): they have implemented coherent policies to establish equal opportunities for all 

students, notably those with a non‐academic family background; they welcome diversity 

and  are  investing  in  further  specific  professional  development  of  teachers.  All  six 

participating Principals were committed to the function and place of their school in the 

context  of  societal  themes,  notably  the  full  and meaningful  participation  of migrant 

pupils in society, and their position on the future labour market.  

We  offered  no  cross‐information  between  participating  schools;  it  may  be 

assumed, however, that the Principals themselves have frequent collegial contacts.  

We  chose  to  base  our  study  on  structured  interviews  with  school  leaders 

because,  according,  for  example,  to  the  report  on  leadership  and  student  learning, 

funded  by  the  Wallace  Foundation  (Louis,  Wahlstrom,  Michlin,  Gordon,  Thomas, 

Leithwood, Anderson, Mascall, Strauss and Moore, 2010) "…Leadership explains five to 

seven % of the variation in student learning across schools (not to be confused with the 

                                                                                                                                                                   WO: research universities, ISCED 5A 

The completion of ISCED level 3 is the internationally agreed initial (or basic) qualification for the labour market. School leavers without an ISCED 3 qualification are regarded as early school leavers or drop outs. 

Page 155: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

129 

very  large within‐school effects that are  likely). Five to seven %, however,  is about one 

quarter of the total across‐school variation explained by all school‐level variables, after 

controlling  for  student  intake or background"  (Classroom  factors explain more  than  a 

third  of  the  variation).  The  authors  add,  however,  that  it  is  as  yet  unclear  what 

characteristics  and  practices  drive  this  school  leader's  effect  on  outcomes. With  the 

current  study we  hope  to  contribute more  insight  into  the  questions,  ambitions  and 

practices of the six involved Principals. 

Another  reason  to  choose  Principals  as  our  source  of  information  is  their 

intermediate role in discussing the function of a school in the context of societal themes, 

with external parties, other  schools and  local and national politicians  (Sanders, 2009). 

Importantly, the six school leaders played an important role in the design of the present 

study, where they acted as co‐creators.  

5.3.2 Five separate phases of investigation 

Phase 1: Semi‐structured in‐depth interviews (Research question 1) 

The current research started with a 1.5 hour semi‐structured in‐depth interview 

with  the  Principal.  Each  interviewee was  informed  about  the  overall  purpose  of  the 

current  research;  the  participants  were  also  informed  about  the  method  and  time 

schedule of the project. The interviews were (with permission) recorded. 

The interview consisted of three parts (see Appendix 5‐2 for the complete list of 

questions);  

1. Part 1: A  reflection on  the  school's context,  in a way  that was meaningful and 

relevant in the view of the Principal.  

2. Part 2: A description of  the  influence on  specific  current policies  and  teaching 

practices of distinctive  context  characteristics, or  characteristics of  the  student 

population.  

3. Part 3: When  reflecting on parts 1 and 2 of  the  interview, what are  important 

remaining questions considering further ambitions and developments that are as 

yet  difficult  to  answer?  What  kind  of  scientific  knowledge  (practical, 

organizational, conceptual) would support the Principal in further ambitions and 

the  development  of  school  quality,  and  what  are  the  problems,  so  far,  with 

obtaining this type of academic knowledge? 

Page 156: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

130 

A report of the interview, based on the recorded discussion, was returned to the 

Principal, in order to verify that the text correctly covered the discussion.  

Phase  2:  The  nature  of  school  questions  and  the  search  for  a match with  scientific literature (Research question 2 and 3). 

In  every  interview,  between  three  and  five  main,  complex  questions  were 

identified, in total 21 questions for the six schools together. In this list different types of 

school  questions  could  be  distinguished  (e.g.  conceptual  and  organizational  themes, 

long term policy effects, content of the curriculum, the balance between cognitive and 

non‐cognitive  skills  and  effective  policies  for  teacher  quality)  that  required 

correspondingly  different  types  of  applicable  academic  research  outcomes.  The 

literature search, in general, followed the order of trying to find, in the first place, causal 

evidence,  based  on  randomized  trials  or  quasi‐experiments.  Secondly,  (meta‐)  review 

studies,  carried  out  by  top  ranked  researchers  and  institutes, were  selected.  Thirdly, 

when  school  leaders had questions about good practice  in other  schools, case  studies 

carried  out  by  highly  ranked  researchers  were  selected.  International  comparative 

studies were selected to match questions on practices in other countries. School leaders 

were  informed about the nature of every academic research article we sent,  following 

the above order. Considering  the quality of  selected publications,  the  following  list of 

criteria has been used: 

All  selected articles had  to be published  in  reviewed  journals; books had  to be 

written or edited by academic authors who published highly ranked articles. 

When  case  studies were  used,  studies  in  reviewed  journals were  selected,  or 

case  studies  published  on  programme websites  of  top  ranking  universities  or 

institutes. 

Research  outcomes  should  have  clear  prospects  for  translation  to  school 

practices and ambitions. 

In  the  case  of  research  publications  based  on  educational  systems  in  other 

countries  (this was  the  case  for most  of  the  articles  that were  selected),  the 

relevance  for  the  Dutch  situation  should  be  obvious  (and  was  explicitly 

explained). 

No publications before 1999 were used.  

Page 157: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

131 

Phase 3: Feedback in the format of a letter  

In order to prevent that the Principals from having to read yet another report– a 

recurrent  complaint  expressed  during  the  interviews–  the  matching  academic 

publications were sent  in the format of a  letter, specific for each school, of around five 

pages.  In this  letter the school's questions were summarized and the method used  for 

selecting publications (described above) was also explained. The  links to all the articles 

were attached as well. The  letter briefly motivated the choice of publications, and how 

and why, in our view, these connected in our view to the school's questions. The school 

leader could choose to use the short descriptions in the letter, or study the whole of the 

attached research papers.  

The  Principals  each  received  between  four  and  ten  references  in  total.  For 

practically  all  school  questions  an  extended  list  of  high  quality  research  papers was 

available  in academic databases; the  idea was, however, to send no more than one or 

two state‐of‐the‐art references per question, with the offer that more references could 

be  sent upon  request. We  restricted our answer  to a maximum of  ten  references per 

letter.  

Phase  4:  Intermediate  check  in  three  of  the  six  schools  of  the  applicability  of  the research matches that had been sent (Research question 4) 

The concise style of the letters raised questions by us on the accessibility of the 

texts concerned. In order to test this, with three Principals (A, B and E) an extra meeting 

was planned to reflect on the applicability of the recommended research in the letter.  

The following interview format was followed: 

Did  the  Principal  have  the  opportunity  to  read  the  letter  and  the  attached 

scientific articles? 

Did the articles, in the Principal's view, match his or her initial questions? 

Did the letter raise any further questions, and was extra information needed? 

Phase 5: Evaluative interview (Research question 4, 5 and 6)  

The evaluative  interview  took 1.5 hours  and was  clearly  structured  around  six 

evaluating  questions  on  the  utilization  and manageability  of  the  offered  sources  of 

relevant  knowledge.  These  questions  concerned:  1)  an  impression  of  the  whole 

trajectory  of  the  current  study;  2)  the  usefulness  of  the  academic  knowledge  for 

Page 158: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

132 

discussions with  staff;  3)  the  nature  of  the  use  of  the  presented  research  outcomes: 

instrumental, conceptual, re‐consideration of running processes, use in external debates 

(after Johnson, 1998); 4) the  influence of the scientific  insights on the  interpretation of 

the  school's  context; 5) obstacles  for  the  transfer of academic knowledge  to practice; 

and 6) the options for scaling up the current method (for the full  list of questions, see 

Appendix  5‐3).  The  evaluating  interview was  planned  three months  after  the  school 

leaders had  received  the  letter with  research  recommendations.  The  interviews were 

recorded with permission. 

The  final  interview  partly  mirrored  the  initial  interview:  the  initial  interview 

started with the Principal's reflection on the school context and questions that followed 

from  this  specific  context,  the  final  interview  ended with  the  question whether  the 

recommended scientific insights had influenced the school leader's views on the school's 

environment.  

5.4 Portraits of the participating schools and their main questions 

Below  we  will  describe  the  participating  schools  in  the  following  order  of 

discussion: 

We start with a characterization of every school, and summarize the number and 

type of questions. All questions (numbered Q1, Q2, etc.) and the matches that came out 

of  the  subsequent  literature  review  (numbered  accordingly A1, A2,  etc.)  are  listed  in 

Appendix  5‐4.  Table  5‐1  presents  all  question  clustered  in  themes,  plus  the matched 

publications.  Table  5‐2  lists  the  categories  of  research  publications  that  have  been 

selected per thematic cluster of school questions.  

Furthermore, we describe an intermediate discussion in the case of Schools A, B 

and  E,  shortly  after  the  Principal  had  received  the  scientific matches with  his  or  her 

queries.  

Finally, we quote  the Principal's evaluative  remarks on  the project and present 

an inventory of the use of the suggested research findings, as well as recommendations 

for  the  follow‐up  of  this  project.  Table  5‐3  presents  the  outcomes  of  the  evaluative 

interview, and Table 5‐4 the applications of the research outcomes in school practice. 

Page 159: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

133 

5.4.1 School  A:  An  enriched  idea  of  the  school's  added  value  for  students  with  a migrant family history 

School  A  serves  a  vast  majority  of  pupils  with  a  Turkish  and  Moroccan 

background; the school offers the two highest secondary tracks, and, therefore all pupils 

with a diploma qualify for higher education.  

The Principal's most  important question concerned the segregated character of 

the school, which was situated  in most of the pupils' area of residence. After students 

have  successfully passed  their  final exam  they  typically will enrol  in higher education, 

and may be confronted with a different culture that may be partly unfamiliar to them. 

There might  be  even more  challenges,  the  Principal  expects,  when  his  pupils,  after 

having completed a university degree, enter the higher strata of the labour market. The 

Principal had two main questions: Is it an advantage for pupils to be enrolled in a school 

that  is  specialized  in  upward mobility,  albeit  segregated‐  or  would  pupils  be  better 

prepared  for  a  full  and meaningful  participation  in  Dutch  society, when  they would 

attend a desegregated school?33 Could pupils, furthermore, benefit from more teachers 

who also have a migrant background?  

School A  raised  three main  questions,  that  could  each  be matched  to  several 

research publications. Two questions were of a conceptual, strategic nature, concerning 

segregation  of  schools;  one  question  concerned  policies  for  hiring  teachers.  2  books 

were suggested, 1 study based on causal evidence, 1 international comparative study, 1 

randomized experiment, 1 website. 

The extra discussion, to explore the opportunities for the use of the selected research publications 

School  leader  A  was  surprised  that  the  research  findings  in  important  ways 

differed  from  what  he  had  expected.  His  expectation  had  been  that  desegregated 

schools were  always  the  better  option,  especially  considering  future  chances  on  the 

labour market for his pupils. The Principal, furthermore, valued the presented research 

on the importance of developing multiple identities for migrant youth, and said that he 

felt  encouraged  to  invest  further  in  specialized  teacher  competences with  regard  to 

                                                       33 In the Netherlands all schools are funded equally by the government and low SES pupils, including many pupils with a migrant background, receive extra  funding. Schools are  independent of  local  tax revenues and parents do not pay tuition fee in elementary and secondary education. 

Page 160: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

134 

migrant  students.  The  research  findings had been used  in debates with  local political 

leaders and other school leaders in the district. 

The evaluative interview 

The Principal  stated  that  the project had deepened his  thinking about  the  role 

and  function of his  school  in a community  largely consisting of people with a migrant 

background.  

Based  on  our  discussions  and  the  letter  with  relevant  research 

publications,  I  started  thinking  somewhat differently about my  school.  I 

still think desegregated mixed schools are the preferred option, but  I am 

more inclined to think now about my school as an accelerator for the full 

participation  of  migrant  pupils  in  society.  I  especially  appreciated  the 

participation (NB of the researcher) in the debate with political governors 

of  the  Amsterdam  district  New  West.  Looking  at  segregation  from 

different knowledge‐ based perspectives was helpful;  it made me  realize 

the values of my school, although it is a segregated school. 

I  valued  the  scientific  answers  to  my  questions  and  especially  that 

research  acknowledged  that  schools  are  deeply  contextualized;  even  a 

school  in  the  South  East  district  of Amsterdam may  differ  in  important 

aspects from my school in the New West district.34 

 The  content  of  the  letter  had  not  been  discussed with  colleagues within  the 

school; school leader A considers the complex aspects of equal opportunities for migrant 

children  to  be  his  main  personal  assignment.  The  literature  we  had  suggested  on 

improving opportunities for migrant pupils resulted, according to the Principal, in a more 

positive view on his school "I have a warmer appreciation of my own school, and have a 

higher esteem of our professionalism".  

The  offered  scientific  knowledge  had  influenced  views  on  current  school 

processes,  and  was  used  in  strategic  discussions  with  external  stakeholders.  The 

information was not used for direct instrumental actions.  

                                                       34 All interviews were in Dutch; citations have been faithfully translated literally in English. 

Page 161: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

135 

The method developed in the current study, according to the Principal, could be 

expected  to be of  special  significance  for  starting headmasters.  The Principal made  a 

salient final remark: the fact that the interviewer in the current study, unlike a coach or 

a consultant, did not charge any fee had strengthened his trust in the project. 

5.4.2 School B: the advantage of a professional review of academic literature 

School  B  is  situated  in  a  district  of  Amsterdam  with  relatively  high 

unemployment, poverty,  and  single‐parent  families. Around 60% of  the pupils have  a 

Surinamese,  Antillean,  or  Ghanese  background.  The  school  has  a  long  history  of 

extended comprehensive education (the Dutch system typically has early tracking), and 

is committed  to  this principle. The school, however,  faces strong competition  (notably 

regarding talented future students and their parents) from schools in the city centre that 

start with early tracking at age 12 at academic levels.  

The  Principal's  main  questions  refer  to  the  school's  choice  for  extended 

comprehensive  learning:  Is  postponed  tracking  and  stimulating  upward mobility  to  a 

higher  track,  chiefly  benefitting  pupils with  a  non‐academic  family  background,  or  is 

longer comprehensive learning benefitting all pupils? Related to the above, the Principal 

raised questions on  the best balance between cognition on  the one hand– and meta‐

cognitive skills (notably academic reasoning) and personal growth on the other, notably 

for  low  SES pupils.  Like  the  five other participating Principals, Principal B emphasized 

also  the  importance  of  receiving  data  on  the  future  success  of  the  school's  pupils  in 

higher education. 

School  B  raised  three  questions  that  could  each  be  matched  to  research 

publications; the question on the future study success of pupils in higher education after 

leaving school could, however, only be answered in general, not for school B specifically. 

Two questions referred to both the school's role in the local community and subsequent 

choices in current school processes. One question concerned data on later study success 

in higher education of former students. 

The extra intermediate discussion 

Principal B strongly recognized the school's questions in the presented academic 

literature. She was particularly interested in research outcomes that prove that schools 

which offer a broader array of tracks (both professional and academic tracks) do indeed 

Page 162: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

136 

increase  chances  for  upward mobility  (Van  Elk  et  al.,  2011). On  the  other  hand,  she 

recognizes  also, however,  that  this  still might not  change  the  current  trend  that high 

potential  students  tend  to  choose  a  school  in  the  city  centre  that  exclusively  offers 

academic  tracks.  The  selected  research  publications  also  support  the  school's  new 

programmes on higher order  learning  skills. The Principal expressed  the view  that  the 

intermediate explorative discussion made her more aware of the opportunities for use 

of the research findings, and motivated her to discuss the letter with the school's project 

team on diversity in education. 

The evaluative interview 

The school leader pointed out that it is also relevant to learn which questions as 

yet  cannot unequivocally be answered by  research,  for example  the question: Does a 

broad  school  like  hers,  providing  all  tracks  (professional  and  academic),  potentially 

generate better results for all pupils, or might, in fact, early tracking for gifted students 

be more attractive, especially for pupils with more affluent parents? She added that the 

current research project came at the right time, since the school was deeply involved in 

discussions  about  continuing  its  mission  of  comprehensive  learning,  with  special 

emphasis on non‐cognitive skills, or instead making a change to partial earlier tracking.  

It is good to be able to use existing knowledge to plan your strategies. We 

think  too  often  in  education  that  we  cannot  select  an  appropriate 

research  basis  for  our  school  innovations,  since  research  on  the  same 

subject often  seems  to present  contradictory  outcomes. Because of  this 

project we know now which research is relevant in our context, and why. 

Immediately upon arrival, the letter was sent through to the management team 

and  the  school's project group  for diversity, which under  the  leadership of a  recently 

appointed vice‐Principal was reconsidering  the school's mission of postponed  tracking. 

The vice Principal expressed the view that "the presented research inspired our thinking 

and gave us a firm base to stand on". 

The research findings had been used  in different ways, direct and  instrumental, 

for conceptual  thinking,  to assess current developments, and  in external debates. The 

school  leadership  and  the  teachers  cooperating  in  the  project  "diversity",  felt more 

Page 163: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

137 

reassured  after  studying  the  suggested  research,  to  remain  faithful–  but  with  new 

energy and ambitions–  to  the  school's  long history of establishing a balance between 

cognitive  learning and personal growth,  in the context of a diverse community.  It was, 

furthermore,  seen as an advantage  to have a  source of  information  from outside  the 

school. 

The Principal expected that the methodology we developed could be interesting 

for cooperating groups of schools. Seminars and focused discussions on specific themes 

could  largely  benefit  form  a  tailor‐made  selection  of  academic  research.  She  also 

expected herself to have further questions on relevant research outcomes in the future. 

5.4.3 School  C:  The  important  next  step,  translating  research  outcomes  to  teacher practices 

The  school's  educational  philosophy  is  based  on  a  didactical  model  that 

stimulates pupils  to  gradually  take more  responsibility  for  their own  learning process 

(known  as  the  Dalton  pedagogy,  based  on  the work  of  Helen  Parkhurst).  Pupils  can 

choose, for example, to partly make their own choices by taking extra classes in specific 

subjects.  The  school,  moreover,  offers  ample  opportunities  for  cultural  and  artistic 

development, notably  in music. Traditionally, this type of school attracts more middle‐ 

and upper‐class students, with typically well‐educated parents. The Principal confronts 

the  school's pedagogical principles with  contrasting  ideas of  strict and ordered  school 

policies, mainly aimed at cognitive learning, that are in public and political debate often 

regarded as being more effective for pupils with a non‐academic background (which  is 

the case for many migrant students).  

The  main  questions  concerned  a  further  knowledge  base  for  the  school's 

didactical  system,  an  effective  organizational  model  for  teachers  working  with  this 

system,  and  research  on  the  effects  of  grade  repetition  (a  particular  concern  of  the 

school  leader).  three  questions  were  raised;  one  question  could  be  matched  with 

research that approached the school  issue, but was not an exact match. The other two 

questions  could  be  matched  directly.  3  reviews  were  suggested  (one  of  these  an 

inaugural lecture), 1 report, 1 case study, 1 website. 

Page 164: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

138 

The evaluative interview 

The  Principal  emphasized  that  it  was  interesting  to  learn  that  the  profound 

structured  discussion  on  his  own  school  themes  could  indeed  be  linked  to  research 

outcomes,  although  one  question  could  not  be matched  precisely with  the  relevant 

literature.  

It  is  important  to  realize  that  this  research project also  indicates which 

questions still need to be further  investigated, because  indeed we do not 

know  the  answer  to,  for  example,  the  question  whether  my  school's 

philosophy on  learning, results  in better  learning strategies and planning 

later on in higher education. 

I felt the interview was an invitation to try to formulate relevant questions 

concerning my vision and ambitions, related to actual school practice. 

It  is  important  to  follow  up  this  research,  including  the  actual 

implementation in school practice and policies. 

The  critical  remark  was made  that  the most  important  part  of  an  adequate 

utilization of knowledge, in the view of Principal C, had yet to follow: the translation of 

research  findings  to  actual  classroom  practice;  this  translation  process,  in  his  view, 

requires  further  guidance  by  case‐study  research  on  good  practice.  Therefore,  the 

presented research had not yet been used directly. 

The school leader felt that linking school questions to scientific knowledge could 

be an interesting starting point for inter‐collegial seminars and discussions. 

5.4.4 School D: Taking time to focus on the complex questions we encounter 

School  D  has  largely  invested  in  special  programmes,  like  bilingual  education 

(Dutch and English), top sport tracks, and Technasium (technical and science education). 

The school  is committed to maximizing opportunities for  learning for all pupils, and no 

longer  thinks  along  lines  of  ethnic  or  SES  differences.  In  their  view,  they  serve  the 

surrounding  community  and  try  to  relate  to  the  specific  learning  requirements  of  all 

young people  living  in the nearest of the school. They are,  furthermore, committed to 

postponed  tracking,  in order  to grant pupils more  time  for exploring  their  talents and 

Page 165: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

139 

capacities. The director's question  concerned postponed  tracking, extra  courses  for e‐

learning and  its consequences  for  the organizational model of  the  school. Considering 

the fact that most pupils in school D have a migrant family background, the director was 

concerned about the consequences of currently occurring societal prejudice. Finally, the 

director inquired after the most effective (evidence based) financial intervention, tailor‐

made  for  his  school  and  its  environment.  Five  questions were  raised,  one  could  be 

matched with more general  research outcomes, and  the other  four could be matched 

directly. 

The evaluative interview 

The  input  for  further  conceptual  thinking  on  the  school's  role  and  function 

amidst complex societal themes was felt to be an important advantage of this research 

project. The  letter with links to relevant research, moreover, turned out to be a valued 

source of information for discussions with the school leader's successor.  

In the first interview we discussed subjects that are constantly in the back 

of my mind; but it is hard to find the time to really organize my thinking. I 

appreciated this opportunity. The letter that followed up on the interview 

showed  that  high  quality  research  supports–  and  in  some  cases  even 

proves– our visions on  the added value of a broad  school  like ours  that 

offers all secondary tracks and extra programmes as well. It sounded like 

music  in my ears that there  is  indeed evidence that pupils  in professional 

tracks have more chances to eventually complete higher education when 

they start in a broad school. 

I  feel  supported  by  this  project  in  our  choice  to  be  fully  aware  of  our 

ethical, moral and  societal  responsibility– as opposed  to marketing and 

competition driven strategies. 

In external political debates we feel strengthened in our position. 

According  to  the  Principal,  especially  large  complex  themes  require  focused 

meetings and time to work out all consequences, time that  is always too scarce  in the 

everyday complexity of school life. It was mentioned specifically that a knowledge‐base 

Page 166: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

140 

for  standing  practice  may  be  as  important  as  scientific  insights  for  guiding  future 

projects.  

The current method was considered  to be worthwhile  for cooperating  schools, 

clustered, for example, around serving a diverse student body. 

5.4.5 School E: The pleasure of taking time for reflection 

School  E  is  a  recently‐founded  new  school,  offering  ample  opportunities  for 

choices  on  school  organization  and  didactical  principles,  as well  as  the  hiring  of  new 

staff. The Principal  is dedicated  to making evidence‐based decisions  together with  the 

teachers, and emphasizes the school's ambitions in designing innovative learning for the 

21st century. The four raised questions referred to these themes. All questions could be 

matched with  relevant  research  outcomes.  3  publications with  causal  evidence were 

used, 4 (meta‐) reviews, 1 case study, 1 report and 1 book.  

The extra discussion 

Principal E has a background in educational sciences, and is committed to making 

knowledge driven decisions  for all  strategic school  themes. He has  indeed studied  the 

research findings  in the  letter extensively, and feels further supported  in his  innovative 

ambitions  and  practice.  The  Principal  emphasized  that  the  academic  knowledge 

provided  had  inspired  him  to  look  at  current  school  developments  from  new  angles, 

notably considering the balance between skills and cognition. 

The evaluative interview 

According to the Principal, reflecting on the overall design of school  innovation, 

to  consider  limits  and  opportunities  and  to  reflect  on  the  concepts  underlying  one' 

thoughts,  requires  time and  concentration. Therefore, one of  the main advantages of 

the  current  project was,  in  the  view  of  the  director,  the  in‐depth  discussion  on  such 

concepts.  

This was  a  "gift",  to  just  talk  to  someone  in  a  certain  thoughtful way 

about important issues; you do not find the time in the bustle of everyday 

work  to  reflect  on  such  themes;  to  sit  back  and  think  about  the  larger 

design  in your way of working, your  interventions and your concepts, all 

the  things  that  relate  to  this,  what  are  your  impediments,  what 

Page 167: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

141 

information do you need specifically. You do not easily  find  the  time  for 

this, at least not in my practice as a school leader. That requires someone 

who asks questions in a specific way, who from a comparable background 

in  educational  leadership,  and  for  a  specific  purpose,  continues  to  ask 

further  questions  about  the  things  that  are  of  concern  to  you.  I  really 

enjoyed  that. The knowledge sources  that were offered, you do not  find 

easily as a Principal. I will certainly do more with this, notably considering 

the process of  translating  requirements  for  skills  for  the 21st  century  to 

school practice and policies. 

The Principal,  furthermore,  felt  supported  in his vision  that  innovative  learning 

for  the  21st  century  requires  specific  skills  and  further  training  of  teachers.  He  had 

shared  the  letter  with  his  team  and  among  members  of  his  external  network  of 

innovative  schools.    The  information  was  used  for  conceptual  thinking,  and  several 

internal  and  external  meetings.  While  studying  the  literature  on  the  subject  of 

innovative learning he had realized once more, how conservative, in his view, education 

tends to be. And yet, society is changing at an unprecedented tempo, if not only because 

of the phenomenal new digital prospects.  

The school  leader's recommendation was to use our method to support schools 

that  focus  on  the,  in  his  view,  under‐studied  theme  of  urban  education  in  the 

Netherlands. 

5.4.6 School F: Time is the most scarce ingredient 

School F serves a student population consisting of more than 90% pupils with a 

migrant background, among them many youngsters who are the first  in their family to 

enrol  in  secondary  tracks  that  qualify  for  access  to  higher  education.  The  school 

structurally  invests  in  continuous  formative  assessment  of  pupils,  with  the  aim  of 

enhancing  opportunities  for  upward  mobility  to  higher  secondary  tracks.  The  main 

questions  concern extra programmes  (on  top of  the  standard  curriculum)  for migrant 

students, further teacher training, and leadership. Three questions were discussed, that 

could  all be matched  to  research publications. 3 matches with publications on  causal 

evidence were made, 2 reviews, and 1 report. 

Page 168: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

142 

The evaluative interview 

School  leader  F  was  involved  in  two  large  projects  the  school  itself  had 

commissioned,  and  because  of  that  had  not  found  the  time  to  study  the  letter with 

research links connected to the school questions.  

I  realize  that  I did not  find  the  time  to use  the  information  in  the  letter, 

although I think that this might be worthwhile. I would have to study the 

literature  suggested  in  the  letter, however, outside working hours.  I did 

send the information to the colleagues in my leadership team, but nothing 

came out of that. We are kept busy with all the things that happen every 

day in a school. 

Furthermore,  we  commissioned  two  investigations  ourselves,  one 

quantitative  research  project  into  the  characteristics  of  youth  that 

potentially may  choose  for our  school, and another qualitative  research 

into the expectations of parents in this part of the city. 

We are putting a lot of effort in attracting enough pupils for our academic 

tracks, and try to prevent losing many of our talented pupils to schools in 

the city centre. 

The  Principal, moreover,  thinks  that  inspiring  cooperation  with  other  schools 

might be more effective than the use of scientific research outcomes. When a strategy 

has proven to work in another school, the chances are it will work in your school as well. 

She also critically questions the further training of research skills for teachers, and rather 

emphasizes  the  importance  of  peer  learning  and  examples  of  good  practice  in  other 

schools. 

5.5 Findings 

In this Section we will describe our findings, following the order of the research 

questions we presented in Section 5.3.  

The  rich  interviews  taught  us  that  school  leaders  do  indeed  have  important 

questions that in most cases had already occupied their thinking for a long time.  

Page 169: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

143 

Among a  total of 21 questions,  in  three cases  the question could not be  linked 

exactly to a research publication. In all other cases, a match could be made with one or 

several research publications. Considering the three question that could not exactly be 

matched,  in one  case  the question was aimed at  the  long‐term effects of  the  specific 

didactical model of  school C; no  such  research was  found, but we nevertheless  found 

relevant  research  publications  that  well‐approached  the  school  leader's  question.  A 

question on the best options  for  financial  investment  in school D, would need a much 

wider  literature study  to  link options exactly  to  this particular school; our answer was 

more  generally  directed  at  the  paramount  importance  of  teacher  quality.  A  third 

question  that could not directly be addressed was asked by all six school  leaders, and 

concerned the wish  for data on the  future achievements of  former students,  in higher 

education, and on the  labour market. This recommendation will be discussed with the 

ministry of education.  

The questions could be clustered around five main themes: 

1. Segregation of schools 

2. Tracking in secondary education 

3. The  balance  between  skills  and  competences  on  the  one  hand,  and  cognitive 

learning on the other hand 

4. Teacher quality 

5. School innovation and leadership  

The  categories  of  questions,  plus  the  suggested  literature  references  are 

presented  in Table 5‐1; for every  literature reference a typology of the publication has 

been added  in  this  table. Table 5‐2  illustrates  the categories of publications  that were 

used per cluster of questions. Table 5‐3 gives an overview of how  the Principals have 

experienced  our method,  how  they made  use  of  the  new  knowledge,  and  how  they 

thought about the feasibility for scaling up the current project. Table 5‐4, finally, shows 

more in detail how the Principals had made use of the suggested research (conceptual, 

process, symbolic, and instrumental).  

Taking into account the different context of every individual school, the same‐ or 

a  largely similar question  (e.g. on  segregation and diversity),  resulted  in different best 

matches with scientific literature.  

Page 170: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

144 

In  five out of six participating schools, the suggested research had mostly been 

applied  for:  1)  reconsidering  current  strategies;  2)  supporting  current  strategies;  3) 

developing new policies; 4)  internal and external debates; and 5) making a motivated 

choice out of contradictory research outcomes on the same subject  (Table 5‐4).  In the 

case of five schools the suggested research was used to further elaborate the role of the 

school  in  its  specific environment;  in  four  schools  the  suggested  research publications 

were used to assess current practices, and four school leaders had used the new insights 

in  external  debates.  One  school  leader  suggested  that  the  translation  of  relevant 

research  to  concrete  teacher practice  could be enhanced by a  follow up project. One 

school  leader  felt  that  illustrative  examples  in  other  schools might  be more  effective 

than academic research literature. 

We  encountered  the  unanticipated  problem,  however,  that  14  out  of  the  25 

publications  were  only  accessible  in  the  university  environment:  universities  pay 

substantial  licence  fees  to  allow  researchers  to  study  scientific  journals.  As  a 

consequence, a considerable part of the publications we selected could not have been 

found by school leaders themselves in the first place (apart from the need to possess the 

research skills to be able to identify appropriate matches between question and answer) 

since schools can never afford these substantial licence fees. 

In  Schools A, B and E,  shortly after  the Principals had  received our  letter with 

links to applicable research, an extra  intermediate meeting was planned to explore the 

options for using this new information. In the case of the other Schools, C, D and F, the 

evaluative interview was the first occasion to discuss the letter. In this last set of schools, 

the evaluative interview much resembled the extra explorative interview held in the first 

three schools. 

   

Page 171: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

145 

Table 5‐1: All school questions clustered in five domains, plus matches with selected scientific publications and the nature of every publication. 

Cluster  Literature reference School

Segregation; dissimilarity school and community culture‐ culture in higher education and the labour market; multiple identities; civic education. 

Banks (2009); book. Bunar (2010), empirical study. Cooper (2011), book. Dobbie and Fryer (2009), causal evidence. Parker, Ninomiya and Cogan (1999), international comparison.  

A, D

Tracking in secondary education‐ comprehensive learning. Intergenerational mobility. 

Brunello and Checchi (2007), international comparison, evidence. Crul and Schneider (2009), comparative study. Hanushek and Woessman (2006), evidence. OECD (2011), report. Pekkarinen, Uusitalo and Kerr (2009), evidence. Van Elk, van der Steeg and Webbink (2011), evidence. 

B, C, D

Balance cognitive learning– non‐cognitive; 21st century skills; Academic skills for pupils from non‐ academic families; Learning for personal growth. 

Heckman and Montera (2009), comparative study. Payne (2010), book. Zimmerman (2002), review. Zohar and Dori (2003), empirical study.  

B, C, E

Teacher quality; innovative organizational models for teacher cooperation. Teacher ethnicity reflecting pupils' ethnicity. 

Boyd, Grossman, Lankford and Wyckoff (2009), evidence. Brown (2007), literature review. Darling Hammond and Youngs (2002), review. Dee (2004), randomized experiment. Jackson and Bruegmann (2009), evidence. Hanushek and Rivkin (2007), evidence. McKinsey (2007), report. Saunders, Goldenberg and Gallimore (2009), quasi‐ experiment. Van der Grift (2010), inaugural lecture.  

A, C, D, E, F 

Innovative school organization and leadership; digital learning; interactive learning with other actors in society. Extra programmes e.g. sports, music, bilingual curriculum. 

Ala‐Mutka, Punie and Redecker (2008), report. Ancess (2000), case study. Cheung and Slavin (2011), meta‐analysis. Nilsson (2008), case study. Notten (2009), descriptive. Sanders and Harvey (2002), case study. Tefera, Frankenberg, Siegel‐Hawley and Chirichigno (2011), practical guide. Välijärvi and Sahlberg (2008), review.  

C, D, E, F 

Table 5‐2: Categories of selected research publications, per cluster of questions. (Note that the same publication may be suggested to several schools; e.g. 5 meta‐review studies were matched 15 times with a school question). 

Cluster of  Questions 

Category of  research  

Segregation Tracking Cognitive Learning and 

Skills 

Teacher Quality 

Organization and Leadership 

Total

Causal Evidence, Randomized (quasi‐) Experiment 

1  4 5   10

(Meta‐) Review    1 2 2  5

Case Study    5  5

(International) Comparative Study 

1  1 1   3

Report    1 1 1 1  4

Book  2  1   3

Empirical Study  1    1

Page 172: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

146 

Table 5‐3: Overview of the outcomes of the evaluative interview. Schools A, B, and E had an extra intermediate exploratory discussion. 

School  Overall impression 

Informationshared with others 

How was the suggested research used 

Influence on views on the school context 

Method interesting for other schools 

A  Scientific basis strengthens strategies. Knowledgeable sparring partner appreciated. 

No. Themes concern my personal goals and struggles. 

Conceptual thinking has been influenced. 

Alternative views on the added value of school.  

Scientific basis strengthens debate. Use method for starting Principals. 

B  Relevant selection of literature useful, albeit not all school questions could be answered directly.  

Yes. Discussed with school board and with teachers. 

Recommendations have strengthened the school in the complex redefining of strategies.  

Better understanding of complex relationship between context and school. 

Research input worthwhile for groups of cooperating schools. 

C  Connection between research and practice offers important new insights. 

Yes. With vice‐Principal and friends. 

Scientific basis supports current project to re‐define school didactics. 

Research not yet used. 

Follow‐up important: translate research to teacher practice. 

D  Grateful for insight into research base for complex school questions. 

Yes. With colleagues school leadership and in meetings with other schools.  

Influence on conceptual thinking most important and appreciated. 

Supported development of strategies. 

Use method for clusters of cooperating schools. 

E  Valued reflective discussions with researcher who is also experienced in school practice as well. 

Yes. In my team, network, and at the Ministry of Education. 

Many links to current school innovation. Inspired further reading and conceptual thinking. Support for ambitions. 

Supported ambition for learning for future society. Helpful overview of strategic indicators. 

Specialize in Urban Education, under‐ studied concept.  

F  Due to pressing school developments, no time yet to study the letter. 

Send to management team and teacher project. Not discussed. 

‐ ‐ Illustrative good practices in other schools more relevant. 

   

Page 173: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

147 

Table 5‐4: Application (after Johnson, 1998) of research outcomes by school leaders after three months.  

Type of Application    Frequency 

1. Conceptual  1a: rethinking strategies  3 

  1b: new projects  1 

  1c: school's role and societal issues  5 

2. Process  2a: support current practices  4 

  2b: used in discussions with board of overseers  1 

  2c: support current decision‐making  2 

3. Symbolic  3a: use in external debates  4 

  3b: use in political debates  1 

4. Instrumental  Direct use in teacher practice  2 

5.6 Conclusions and Discussion 

In the current explorative study we have made an inventory of existing questions 

on  the  subject  of  improving  educational  quality  in  six  urban  schools,  and  tested  a 

practical method  for matching these themes to relevant and manageable outcomes of 

existing  high‐quality  research.  We  have  described  both  the  characteristics  of  the 

questions and the categories of scientific publications that matched these questions. In 

the  course  of  the  project, we  found  that  taking  into  account  the  specific  context  of 

schools and a timely and precise match between questions and answer may add to an 

effective use of existing research in schools. Furthermore, we described the ways school 

leaders made use of the presented academic knowledge, and mention the obstacles we 

encountered  concerning  the  use  of  academic  knowledge  in  schools.  We  made  an 

inventory of the evaluative remarks that the participating Principals made of the current 

project and their suggestions for expanding the project to other schools. On the basis of 

these  suggestions, we  reflect  on  the  feasibility  of  an  intermediate  function  between 

educational research and practice, and report on concrete follow‐up.  

5.6.1 Principals have questions that can be linked to research outcomes 

Considering our  first  research question, we have  found  that  school  leaders do 

indeed have  important questions, which  in their view demand an academic knowledge 

basis. As we expected, we found that research publications only rarely reach the school. 

Typically,  school  leaders  learn about new educational developments  through  thematic 

reports  or  publications  on  government  policies.  However,  in  most  cases,  these 

publications and policies do not coincide with current school themes.  

Page 174: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

148 

Out of  a  total of 21  school questions,  all but  three questions  could  indeed be 

matched with existing research publications that could be translated to school strategies 

and  practice  (Research  question  2,  Table  5‐1).  For  the  three  questions we  could  not 

match exactly, we found an approximation in two cases, and we will try to put the third 

question up to the higher level of the research agenda of the Ministry of Education (the 

question addressed long‐term monitoring data on former pupils). 

5.6.2 Comparable questions, different approaches 

Considering Research question 3 (the nature of questions and answers), the main 

questions could be clustered  in  five domains  (Table 5‐2): 1)  segregation of  schools; 2) 

early  tracking; 3)  the balance between  cognition  and  skills; 4)  teacher quality;  and 5) 

innovation.  Although  schools  may  have  comparable  questions,  depending  on  their 

context  the most appropriate answer may be different. To  illustrate  this, we  consider 

the example of the questions  in the cluster "segregation", concerning the best strategy 

for  serving  students  from  non‐academic  families,  notably  students  with  a  migrant 

background.  This  type  of  question was  raised  by  all  school  leaders.  In  one  case we 

suggested  research  that  demonstrates  the  positive  effect  of  close  cooperation  with 

parents,  because  this  particular  school  was  located  in  a  neighbourhood  with  many 

migrant parents who  themselves did not have  the opportunity  to enrol at  the  level of 

academic  learning when  they were  young.  In another  school, however, we presented 

research  into  the beneficial effects  for all students of additional prestigious courses  in 

academic thinking, because this particular school ran the risk of losing its most talented 

pupils  to  schools with  a higher  SES population  in  the  city  centre.  Yet  another  school, 

convinced of  the added value of a diverse  student body, wanted  to attract more  first 

time  academic  learners,  whereas  the  school  was  increasingly  attracting  middle  and 

upper class pupils because of its special emphasis on the Arts. In this case, we suggested 

research into magnet schools that invest in the arts to enhance opportunities, especially 

for children with a non‐academic background. 

As  Table  5‐2  shows,  most  of  the  publications  we  selected  presented  causal 

evidence (10). The choice for a review or a case study publication was usually motivated 

by  the  fact  that  the  Principal  had  specifically  asked  for  illustrations  of  the  use  of 

alternative  strategies.  Note  that  the  same  publications may  have  been  suggested  to 

Page 175: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

149 

more  than one  school. They  valued,  furthermore, publications  that offer  an overview 

and analysis of large numbers of articles on a specific theme, because this offered them 

more comparative insights. 

5.6.3 Actual applications of academic knowledge and impediments 

Our overall finding is that a match with existing school questions can– according 

to  the  interviewed  Principals–  in  important  ways  contribute:  to  developing  school 

strategies;  offer  support  for  current  policies;  inspire  communication with  the  school 

board, and strategic discussions with neighbourhood and municipal (political) governors 

(Research question 4). After only three months, in the case of 17 out of 21 of the school 

questions, five out of six Principals had already used the provided evidence for further 

school  development  and  innovative  programmes  or  as  a  knowledge  base  for  current 

processes, notwithstanding the short period of time of three months between receiving 

our research suggestions and the final interview.  

As Table 5‐4 illustrates, in most cases the new knowledge and insights had been 

used  to  support  and  inspire  conceptual  thinking  (9  times),  especially  considering  the 

school's role  in societal  issues  like segregation. In 7 cases, the research outcomes were 

used to guide and support current school processes. In 5 cases the information was used 

in debates with external stakeholders. Instrumental, direct use occurred only 2 times.  

The  main  hurdles  in  knowledge  dissemination  (Research  question  5)  we 

identified are  the  following:  the Principals  indicated  that  specialized professional  skills 

are  needed  to match  upcoming  questions  to  relevant  literature,  skills which  are  not 

always  available  in  schools. Moreover,  the participants  confirmed  that  the  structured 

initial  interview had been essential  in  the articulation of  significant  themes. However, 

we also encountered a more serious problem: we found that about 50% of the papers 

we  matched  to  school  questions  were  not  accessible  outside  the  network  of  the 

university. Access to scientific papers, in most cases, requires the payment of substantial 

licence fees, which schools cannot afford.  

5.6.4 Contributing to the use of existing research in school practice  

Our  study  suggests  that  a  bridging  function  between  academe  and  secondary 

schools could enhance the use of academic knowledge in schools, notably since schools, 

on  the one hand, and  research  institutes, on  the other, obviously have different goals 

Page 176: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

150 

and organizational incentives (Research question 6). The involved school leaders were of 

the opinion that  it would be worthwhile to expand the method we developed to other 

schools. They stressed the importance of our tailor‐made approach, and were in favour 

of  an  intermediate  flexible  network  function,  rather  than  a  bridging  institute. More 

specifically, they suggested the development of courses  in knowledge based  leadership 

for starting school leaders, and seminars for experienced school leaders to study more in 

depth  main  school  questions.  Furthermore,  the  involved  school  leaders  suggested 

initiating  clusters  of  cooperating  schools,  in  order  to  create  opportunities  for 

comparative research based experiments. One of the Principals advocated the start of a 

cooperative  cluster  especially  for  urban  schools,  since–  in  his  view–  urban  education 

appears  to be an understudied  theme  in  the Netherlands. Finally,  it was  suggested  to 

follow‐up  the  current  project  with  research‐based  strategies  for  the  translation  of 

research outcomes from theory to actual classroom practice. 

Important  steps  have  been  made  by  the Ministry  of  Education,  Culture  and 

Science,  and  by  the  Netherlands  Organization  for  Scientific  Research,  to  start  new 

cooperatives  between  researchers  and  schools  in  designing  a  joint  agenda  for 

educational research. Additionally, the current research contributes to a more effective 

use of  research  in  school practice, by  linking already existing  research publications  to 

school  themes.  By  using  existing  research,  school  questions  can  be  addressed  in  a 

relatively short period of time. 

Several outcomes of the current research  initiated direct follow up. 1) A cluster 

of innovative schools has asked for a programme to be started to link research to school 

developments as a  structural new way of working; and 2) A  leading  institute offering 

advanced  courses  in  educational  development  for  school  leaders  and  teachers  in  all 

levels  of  education  is  interested  in  scaling  up  the method we  developed. Moreover, 

opportunities  for  advanced  courses  for  school  leaders  aimed  at  applying  research  in 

school practice are currently being discussed.  

In  summary,  the  method  we  developed  what  may  be  a  first  step  to  make 

research  actually work  in  schools, when  the  following  requirements  for  its  effective 

transfer from theory to practice are taken into account: 

Schools should be understood  in their specific context  in order to match theory 

accurately to practice. 

Page 177: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

151 

A timely match between actual questions and research answers is a prerequisite. 

School leaders expect a motivated selection and a workable amount of research 

findings  to  support  their  work;  neither  general  reports,  nor  long  lists  of 

publications to choose from, fit their demands. 

The participating Principals were mostly  interested  in academic knowledge that 

guides  their  conceptual  thinking  and decisions on  school  strategies,  related  to 

the school environment– both the surrounding community and societal themes. 

According to the Principals, in‐depth discussions between researchers and school 

leaders  which  explore  options  for  making  use  of  academic  knowledge  can 

contribute in major ways to the transfer of research from theory to practice. 

The start‐up of a flexible network consisting of school leaders and academics will 

be further elaborated 

A serious obstacle needs further study: academic articles, in many cases, are not 

freely accessible outside the university environment because of the  licence fees 

required to access scientific journals. 

5.6.5 Contributing to research into the valorisation of scientific knowledge 

The  findings  in  the  current  case  study may  also  contribute  to  research  in  the 

domain of the valorisation and utilization of knowledge. In the current study we tried to 

avoid  epistemological  discussions  (except  an  occasional  reflection)  on  the  intricate 

differences  between  the  culture  of  academe  and  school  practice. We  focused, more 

concretely, on designing an intermediate practice that acknowledges the characteristics 

of both worlds, and yet sets out to establish a connection that makes research actually 

work  in schools. The participating Principals supported the  feasibility of expanding our 

method  to  secondary  education  at  large.  As  to  the  question  what  competences 

professionals in an intermediate network function should have, principals expressed the 

view  that working experience  in both worlds  should be a  strict  requirement;  this may 

include  both  school  leaders  with  research  experience  and  researchers  who  actually 

worked  inside  schools  for a  longer period of  time. All participants agreed  that  flexible 

tailor‐made networks consisting of scholars, professionals  in schools and policy makers 

could  in  important ways add to the effective use of scientific knowledge  in educational 

practice.  

Page 178: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

152 

5.6.6 Limitations and further research 

An  obvious  limitation  of  our  study  is  the  short  period  of  time  (three months) 

between sending the response to the questions raised in the first interview, and the final 

evaluative  interview.  The  results  of  knowledge‐based  interventions  may  require 

monitoring over a  longer period of  time. Furthermore, as one of  the Principals stated, 

the utilization of academic knowledge might benefit from follow up programmes on the 

actual translation of research to classroom practice. We consider that further research 

into  the possible effects on  school outcomes of  knowledge‐based  interventions  to be 

our most important subject for follow‐up research. Clusters of cooperating schools may 

offer  interesting opportunities for the set‐up of experiments. Additionally, we consider 

our study among six schools as a pilot for a project involving a larger number of schools. 

Finally,  an  interesting  future  experiment  would  be  to  investigate  whether 

different  researchers, who  are  also  experienced  in  secondary  school  practice, would 

make comparable matches of a given question with existing research publications. 

   

Page 179: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

153 

5.7 Appendices 

Appendix 5‐1: General characteristics of the six participating schools.  

School  Tracks  Extra Profile Number of pupils School year 2010/11 

Commitment  

A  Senior general, Pre‐ university 

Extra opportunities for migrant pupils in Years 1 and 2; Assisted homework, extra reading and maths 

495, substantial growth 

Dedicated to supportingpupils who will be the first in their (migrant)family to continue at university level 

B  Pre‐ vocational, Senior general, Pre‐ university 

Combined comprehensive Years 1 and 2, to promote learning between all pupils 

1615, stable Balancing personal growth and high learning goals for a very diverse student body. 

C  Pre‐ vocational, Senior general, Pre‐ university, Gymnasium 

Dalton school,extra opportunities in the Arts, Professional theatre and music studios  

1205, growing Pupils learn to take own responsibility for learning. Partially tailor‐made curriculum according to pupils' choice. 

D  Pre‐vocational, Senior general, Pre‐ university, Gymnasium 

Sports class (top sport potentials) Art & culture class; Technasium; Bi‐lingual track in English and Dutch 

1765, growing To be a strong context‐oriented school and serve all pupils who live around the school. 

E  Pre‐ vocational, Senior general, Pre‐ university 

Innovative school, using evidence‐based research for the design of an interactive learning community 

353, newly‐founded school, growing fast 

Creating a community of learners in open interaction with the school's context and society at large 

F  Pre‐ vocational, Senior general, Pre‐ university, Gymnasium 

Bi‐lingual track in English and Dutch.; Cooperation with primary schools for high potential pupils  

902, growing Committed to stimulatingupward mobility to a higher track for a majority of pupils with a migrant background  

Notes:  For ISCED qualifications see footnote 3. Gymnasium is a pre‐university track with extra courses in the classical languages Greek and Latin. Dalton is a didactical concept based on the ideas of Helen Parkhurst. 

Appendix 5‐2: Format of the first interview 

Before  the actual  interview  started,  the  interviewees were  informed about  the 

purpose of  the study:  to  identify core questions concerning  the school's performance, 

Page 180: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

154 

that are  regarded  to be  school or context  specific. Secondly  it was explained  that  the 

questions would be the starting point for a review of the relevant academic  literature; 

the quality  standards  for  the  selected articles were  clarified, notably  the  requirement 

that  the  research  findings  could  be  translated  to  school  practice  or  strategic 

developments.  The  interviewee  was  informed  that  we  would  provide  an  answer  by 

letter  to  the  raised  questions  consisting  of  no  more  than  six  pages.  Finally,  the 

evaluating interview was announced, three months after the letter had been received. In 

Schools A, B and E an extra meeting, shortly after the  letter was sent, was planned. All 

interviews were  recorded with  permission.  The  confidentiality  of  the  discussion was 

guaranteed. 

Phase 1 

For  the  record,  a  brief  overview  of  the  school's  administrative  school 

characteristics was asked (number of pupils, staff, offered secondary tracks, extra 

programmes, composition of the student body). 

How would you describe, in ways that are meaningful to you, the school context? 

Could  you  describe  where  your  pupils  come  from,  how  many  live  close  the 

school, how many pupils does the school attract from other city districts or from 

other towns? 

Could you describe which context, staff, or pupil characteristics have  influenced 

school policies? 

Phase 2 

Could you elucidate how context and pupil characteristics have been translated 

to specific, demonstrable school practice and strategies? 

What kind of knowledge, data and information supported these decisions? 

Were stakeholders (e.g. parents, the governing board, local political government) 

in any way involved in these discussions? 

Does the school have a special didactical concept, a specific mission, or motto? 

Phase 3 

Did  the characteristics of  the  school and  its context  that we discussed,  lead  to 

specific views on your role as a school leader. 

Page 181: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

155 

With reference to school quality improvement as a continuous process, what are 

the most significant themes, questions or obstacles on your agenda? Could you 

select those themes that are within your span of control as a Principal, but that 

require,  for example, because of  their  complexity, or because  these are highly 

specific  concerns  for  your  school,  additional  knowledge  and  insights. What,  in 

your view, makes it difficult to acquire appropriate knowledge?  

Could you select three to five questions that you consider to be crucial?  

The interview was concluded with the question whether the Principal had missed 

something  in  the  dialogue,  or wanted  to  add  information.  At  the  end  of  the 

interview the time schedule was briefly repeated. 

Appendix 5‐3: Questions for the evaluating interview 

1. Could  you  describe  how  you  experienced  this  research  project  (associations, 

images, metaphors)? 

2. Did the matches with academic knowledge influence your thinking or your work? 

Could you describe  if and how you could apply this knowledge? Did you discuss 

the matches  of  scientific  outcomes with  the management  team,  teachers,  or 

external parties? 

3. When you actually did use the information that was offered, and in what way did 

you use this? a) Direct practical application, b) conceptual thinking, c)  influence 

on current school processes, d) used in discussions with external parties. 

4. In  interview 1 we started with your reflections on the school's context and how 

these context characteristics  lead to school strategies and practices. In this final 

interview we ask  the mirror of  this  initial question: did  the presented matches 

with research influence your views on the school's context? 

5. Would  it  be worthwhile  to  scale  up  this method?  If  so,  do  you  have  critical 

remarks and suggestions? Could you describe your views on the characteristics of 

an intermediate role between academic research and schools? 

Page 182: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

156 

Appendix 5‐4: Questions and matches per school 

School A 

Q1   Is it an advantage for pupils to be enrolled in a school that is specialized in upward 

mobility,  albeit  segregated,  or  would  pupils  be  better  prepared  for  a  full  and 

meaningful  participation  in Dutch  society  if  they were  to  attend  a  desegregated 

school. 

A1  Banks (2007), Cooper (2011), Dobbie and Fryer (2009), Parker, Ninomiya and Cogan 

(1999).  

Q2   Should recruitment policies aim at hiring more teachers with the same background 

as pupils? Would teachers with the same migrant background enhance the learning 

and identity building of migrant pupils? 

A2  Dee (2004).  

Q3  Which  examples  could  inspire  the  school  in  translating  specific  school 

characteristics to class room practice?  

A3   The  Centre  for  Multicultural  Education  of  the  University  of  Washington  was 

suggested  as  a  source  for  illustrative  programmes  and  academic  publications  on 

multicultural  learning,  (http://education.washington.edu/cme/).  Furthermore,  the 

Stanford Centre for Opportunity in Education (http://edpolicy.stanford.edu) and the 

Dutch "Echo" programme with university students acting as mentor  for secondary 

school pupils (www.echo‐net.nl/) were advised as examples of interesting networks 

and good practice. 

School B 

Q1  Is  postponed  tracking  chiefly  benefitting  pupils  with  a  non‐academic  family 

background,  or  is  comprehensive  learning  benefitting  all  pupils?  Related  to  this, 

how can the best balance be created between cognition on the one hand‐ and non‐

cognitive skills and personal growth on the other hand, notably for low SES pupils?  

A1   Hanushek and Woessman (2006), Payne (2010), Van Elk et al. (2011).  

Q2   In  the  Principal's  view,  the  development  of  academic  learning  skills,  excellent 

language  skills,  and  academic  reasoning  is  an  especially  important mission  for  a 

secondary school with a  large number of pupils with a non‐academic background. 

Page 183: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

157 

Does  research  support  this view, has  research been  carried out on  the effects of 

different classroom practices for academic learning strategies? 

A2   Brunello and Checchi (2007), Crul and Schneider (2009), Zimmerman (2002), Zohar 

and Dori (2003). 

Q3   This third question was raised by all other  interviewees: the school would be very 

interested in longitudinal data on study success and access to the labour market of 

its  own  former  students;  this  was  considered  to  be  crucial  information  for 

evaluating the quality of teaching, didactical practices, and strategic choice. 

A3   Privacy  laws and regulation severely  limit this type of school based  information on 

individual  former  students.  Since  all  six  Principals  brought  up  this  issue,  this 

question, obviously, will be followed up by further exploration of options with the 

Ministry of Education. 

School C 

Q1   Has  research  been  carried  out  that  demonstrates  that  the  school's  specific 

didactical system (Dalton), and the emphasis on the Arts, can potentially reinforce 

the  emancipation  and  upward  mobility  of  underrepresented  groups  in  higher 

education and the higher strata of the labour market?  

A1   In the present study, this turned out to be a difficult question to answer precisely. 

We  choose  to  take  as  a  proxy  for  the  school's  specific  didactical  system,  the 

example  of  magnet  schools  (Notten,  2009).  Furthermore,  the  aforementioned 

Center  for  Multicultural  education  at  the  University  of  Washington  provides 

research findings on the  importance of cultural education and the arts. The OECD, 

in  its  publication  Education  at  a  Glance  (OECD,  2011),  finally,  argues  that  the 

success  of  the  Finnish  system may  be  explained  by,  among  other  indicators,  the 

combination of comprehensive education until age 16 and flexible ability grouping. 

This combination seems related to school C's model. 

Q2   The second question is also related to the school's didactical model: Does research 

offer  insights  into  organizational  development models  in  which  teachers  take  a 

major  role  in  exploring,  implementing  and  monitoring  innovative  teaching 

methods? 

A2  Ancess (2000), van der Grift (2010). 

Page 184: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

158 

Q3   Third, the Principal was interested in research on the effects of grade repetition and 

examples of successful other practices to respond to the learning tempo of pupils. 

A3  (Välijärvi and Sahlberg, 2008). 

School D 

Q1   The  director  was  particularly  interested  in  a  firm  research  base  for  postponed 

tracking, in the context of learning in diverse groups.  

A1   Hanushek and Woessman (2006), Pekkarinen, Uusitalo and Kerr (2009)Van Elk, Van 

der Steeg and Webbink (2011). 

Q2   Considering that most pupils at School D have a migrant background, the director 

was  concerned  about  the  effect  of  prejudice  and  stigmatizing  that  pupils  may 

encounter.  He was  especially  interested  in  research  on  the  long‐term  effects  of 

stigmatization. 

A2  Banks (2007), Bunar (2010). 

Q3   Suppose  the  school  had  extra  financial  resources,  what  would,  according  to 

scientific  research,  be  the  most  effective  intervention  for  further  quality 

development?  

A3   It turned out to be very complex to answer this question on the allocation of extra 

resources  specifically  for  this  particular  school. Academic  literature  puts  a  focus, 

however,  on  the  quality  of  teachers  as  paramount:  Brown  (2007),  Jackson  and 

Bruegmann (2009), McKinsey & Company (2007). 

Q4   The  school  is well equipped  for  computer  supported  learning: What are  the best 

research‐based  practices  that  could  guide  the  school  in  the  optimal  use  of  e‐

learning?  

A4  Ala‐Mutka, Punie and Redecker (2008), Cheung and Slavin (2011). 

Q5  Which  studies  can  further  inform  the  school  about  effective  strategies  for 

innovating  the  organization  of  learning,  notably  interactive  learning  with  the 

school's environment?  

A5  Ancess (2000), Nilsson (2008). 

Page 185: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

5 A tailor made transfer of scientific knowledge to school practice 

159 

School E 

Q1   The  main  questions  concerned  the  balance  between  innovative  learning,  while 

maintaining  the  high  cognitive  standards  of  the  national  final  exam,  and, 

furthermore, a research base for interactive learning with the environment.  

A1  Heckman and Montera (2009), Payne (2010). 

Q2   Secondly, the Principal was interested in a knowledge base for effective leadership 

in innovative schools  

A2  Sanders and Harvey (2002). 

Q3   Since  school  E  is  organized  in  smaller  unities with  sub  teams  of  teachers, more 

insight is required in best evidence for effective teacher cooperation and training.  

A3   Brown  (2007),  Jackson  and  Bruegmann  (2009),  McKinsey  (2007),  Saunders, 

Goldenberg and Gallimore (2009).  

Q4   The hiring of new teachers is an important strategic advantage of a newly founded 

school with  a  constantly  growing  student  body; what  teacher  qualities  correlate 

positively to learning outcomes?  

A4   Boyd, Grossman,  Ing, Lankford and Wyckoff (2009), Darling‐Hammond and Youngs 

(2002), Hanushek and Rivkin (2007), Van der Grift (2010). 

School F 

Q1   The  Principal  inquired  after  illustrative  extra  programmes  at  other  schools, 

additional  to  the  standard  curriculum,  aimed  at  creating  extra  opportunities  for 

pupils, like the example of magnet schools in the U.S. 

A1   Tefera, Frankenberg, Siegel‐Hawley and Chirichigno (2011). 

Q2   Research convincingly demonstrates that no other factor has such a large effect on 

learning outcomes as teacher quality.  In School F, where teachers constantly work 

on the upward mobility of their students, this may be even more  important. What 

elements  of  teacher  quality  matter  especially,  and  what  research  skills  should 

teachers  develop  in  order  to  effectively  monitor  the  effects  of  their  teaching 

methods? 

A2  Ancess (2000), Darling‐Hammond and Youngs (2002) Hanushek and Rivkin (2007). 

Page 186: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

160 

Q3   Furthermore, a question is raised on the relatively under‐used opportunity in Dutch 

education to use the influence and authority of the Principal in selecting personnel, 

and his or her role in stimulating effective cooperation among teachers.  

A3  Boyd, Grossman, Ing, M., Lankford, H., Wyckoff, J. (2009). 

Page 187: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

161 

6 Conclusions and Discussion 

I know that I am human and may have erred. However, I have 

taken great pains not to err, and to ensure above all that whatever 

I have written should be entirely consistent with the laws of the 

land, with piety, and with morality. 

Baruch de Spinoza, 1670 

6.1  Conclusions 

This  thesis  has  explored  success  in  secondary  education  among  pupils with  a 

migrant background in the four major cities in the Netherlands. This quest commenced 

from  four  different  points  of  departure:  ethnic  segregation  of  the  pupils'  area  of 

residence, the stability of school careers, patterns of school choice, and the transfer of 

relevant and practicable research outcomes to schools with many migrant pupils.  

Local  and  national  data  sets,  different  statistical  techniques,  and  case  study 

research based on  in‐depth  interviews have been used  in an effort  to  investigate  the 

achievements  of  migrant  pupils  in  a  multidisciplinary  way.  In  this  last  chapter,  the 

findings  and  conclusions  from  the  previous  chapters  are  combined  and  discussed  as 

concluding  research  findings  that may  imply  suggestions  for  policies  at  the  national, 

Page 188: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

162 

municipal,  and  school  level  aimed  at  speeding‐up  the  closing  of  the  educational 

achievements gap between migrant pupils and native Dutch pupils. 

The  association  between  the  percentage  of  16  year‐old  pupils  of Moroccan 

descent at zip‐code level and enrolments in secondary academic tracks is negative. 

In this first study, to set the scene, we took a closer look at the widespread, yet 

not  clearly  underpinned  notion  in  public  parlance  and  political  debate,  that 

neighbourhood segregation hinders migrant pupils to find their way to higher strata of 

secondary  education.  The  literature  on  the  negative  effects  of  segregation,  notably 

considering  large metropolitan areas  in  the U.S., describes a strong negative effect on 

educational  outcomes.  Other  literature,  however,  reports  the  beneficial  effects  of 

closely‐knitted migrant communities and neighbourhood schools.  

Instead of considering the ethnic composition of the total population at zip‐code 

level, we chose to take a different approach: we measured specifically the share of 16 

year‐old pupils of Moroccan descent among all 16 year‐olds living in the same area. We 

decided on this option because national statistics point at very large increases in average 

attainment  levels  between  immigrant  generations  of  grandparents,  parents  and  the 

current young generation. For this reason, we wanted to observe, as closely as possible, 

this  youngest  generation  separately.  We  used  bivariate  linear  (OLS)  and  LOWESS 

regressions  to  consider  the  percentage  of  16  year‐old Moroccan  second  generation 

youth  at  zip‐code  level,  and  their  actual  enrolment  in  academic  secondary  tracks. 

Overall,  this association  is negative:  for every 10%  increase of Moroccan 16 year‐olds, 

there is a 2% estimated decrease in enrolments in academic tracks. The data set we used 

for this study did not cover other socio‐economic variables besides ethnicity; therefore, 

we  could  not  interpret  further  the  high  dispersion  of  data  points  we  found  that 

suggested  that  factors  other  than  ethnic  neighbourhood  composition  and  ethnicity 

affect school success.  

 

Switching between  secondary  schools burdens pupils;  it  reduces  the odds of 

obtaining  a  diploma  or move  up  to  a  higher  track,  and  increases  the  odds  of  later 

dropout. Moroccan pupils living in three disadvantaged zip areas switch schools most 

frequently. 

Page 189: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

6 Conclusions and Discussion 

163 

We started this second project with an exploration of a very detailed municipal 

data set (The Amsterdam Social Development Office, DMO), without formulating specific 

questions  in  advance, with  the  objective  to  detect  any  patterns  that might  point  to 

phenomena that may have been overlooked so far. The data set covers the educational 

history of 16‐22 year‐olds in three disadvantaged zip areas in Amsterdam. Already, a first 

visual inspection of the data files raised questions on the remarkably high number of re‐

enrolments in secondary schools for considerable numbers of pupils.  

The  literature  unequivocally  illustrates,  and  presents  causal  evidence,  that  all 

switching– even for positive reasons–  is harmful for pupils, and even more harmful for 

low SES pupils. This harmful effect is attributed to the disruptive effect of, for example, 

losing  classmates,  familiar  teachers, and a  curriculum  the pupil  is used  to.  In  the  first 

part of this research we measured switching  in straight counts of schools attended per 

pupil– measured after they had exited secondary school– and the relationship to school 

success  (diploma)  or  dropout. Our  findings  point  at  the  same  negative  effect  as  the 

literature  predicts:  migrant  students  switch  more  often,  and  repeated  switching  is 

progressively linked to a considerable decrease in the percentage of pupils who obtain a 

diploma,  and  increasing  dropout.  Motivated  by  these  findings,  we  repeated  our 

measurements with  the national educational data source BRON at a different point  in 

time  in  the  pupils'  school  career  (secondary  Year  3)  and  on  the  larger  scale  of 

Amsterdam.  In  this  second  part  we  looked  specifically  at  the  connection  between 

switching and upward or downward  track changes. We used a dummy‐coded variable 

for  switching  here,  and  did  not  count  the  number  of  schools  attended.  In  the  three 

socio‐economically‐challenged  areas  all  pupils  switch  more  often,  except  for  native 

Dutch pupils. We are somewhat puzzled by our  findings  that  in  the case of all groups, 

but more  so  among Moroccan  pupils,  a  share  of  switchers move  to  another  school 

within  the  same cluster; we do not know whether  this  type of  switching  is associated 

with  secondary  track  change,  or  that  pupils  with  behavioural  problems more  often 

migrate within a cluster. Additionally, the vast majority of switchers continue in another 

school at the same track level. This raises the same question about the motive for school 

change. Especially  in the case of migrant pupils, upward mobility to a higher secondary 

track  is an  important extra opportunity to reach the  level that best fits their capacities. 

Staying enrolled in the same school, is convincingly connected to higher percentages of 

Page 190: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

164 

upward mobility; considering pupils with a Moroccan background, almost one  in every 

five pupils moves up to a higher track  in the same school (in Amsterdam), while this  is 

only 7.6% after switching.  

 

88.7% of pupils living in the four largest cities exercise the right to free school 

choice and do not select the nearest school. Migrant pupils on average prefer a school 

with more migrant  students  than  the nearest  school; native Dutch pupils  show  the 

opposite  preference.  Unexpectedly,  the  odds  of  moving  up  to  a  higher  track  are 

slightly  lower for Moroccan pupils who select another school than the one closest to 

their residence. SES, rather than ethnicity, drives school choice. 

In the literature about the unequal opportunities for low SES pupils to find access 

to high quality schools, much  is expected of extended options for school choice.  In the 

Netherlands  this  choice  is  completely  free,  there are no  financial barriers of any  sort, 

and the Inspectorate assesses the sufficient quality of all schools, although there may be 

quality  differences  above  this  basic  quality  level.  This  situation  provides  interesting 

opportunities  to analyse which  factors drive school choice  in  this  freely accessible and 

transparent  school market  (all  school  quality  reports  by  the  Inspectorate  are  on  the 

Internet). We were permitted to use the  latest updated version of the rich educational 

data source BRON, and could merge these data with the socio‐economic characteristics 

of neighbourhoods, and with school quality standards by the Inspectorate. As a proxy for 

the  selectivity of  choice, we used  two distance measures:  the distance  from home  to 

school, and the difference in distance between the nearest and the preferred school. We 

used  multi‐level  regressions  analyses  (OLS),  also  in  the  case  of  the  dummy‐coded 

variable nearest/non‐nearest school, and  the distance difference  (always  ≥ 0); we also 

carried out both a Logit and Tobit regression, but preferred the OLS regression because 

we  found no disturbing differences between  the  two  techniques, and decided against 

the complex interpretation of odds ratios in the case of the Logit and Tobit regressions.  

We found that migrant pupils on average travel smaller distances to school than 

native Dutch students; this may correspond, however, with the fact that more migrant 

pupils live in poorer, more densely populated areas, with more schools. Neighbourhood 

SES has a two‐way effect on distance to school:  in  low SES areas, migrant pupils travel 

shorter distances, but native Dutch pupils tend to escape the area and travel further to 

Page 191: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

6 Conclusions and Discussion 

165 

school. In contrast, in high SES areas, native Dutch pupils tend to travel a small average 

distance and more often choose the nearest school with a population that mirrors the 

neighbourhood population. Whereas we found that a higher school average SES attracts 

native Dutch pupils, somewhat unexpectedly, we found hardly any association between 

distances  travelled and  school quality  indicators,  like  the average exam  score and  the 

school's  percentage  of  upward mobility  to  a  higher  track. We were  surprised  by  our 

finding  that migrant  pupils  tend  to  prefer  a  school  further  away  from  their  area  of 

residence, which  had  an  even  higher  percentage  of migrant  pupils  than  the  nearest 

school. Moreover, in their case, a more selective choice to a school at a greater distance 

was only slightly associated with a lower percentage of pupil upward mobility to a higher 

track. Additionally, our analyses demonstrate  that migrant pupils on average attend a 

school with a higher mean percentage of upward mobility  than  the schools chosen by 

native  Dutch  pupils.  However, migrant  pupils  themselves move  up  less  often  to  an 

academic track. 

 

Secondary  school Principals have  strategic and complex conceptual questions 

that can be matched with the high‐quality practicable outcomes of academic research. 

Principals express the view that  intermediaries who know the reality of schools, and 

are also academic  researchers, could bridge  the gap between secondary schools and 

academe. 

Overall,  the  literature  on  the  applicability  of  scientific  knowledge  for  school 

practice  demonstrates  that  this  transfer  is  rather  problematic.  Complex  issues  arise, 

considering  for  example  the  different  types  of  knowledge  –academics may  prefer  to 

isolate features  in education, while schools consider the complexity of everyday school 

life– and the difficulties school leaders may encounter in translating research findings to 

concrete classroom practice.  

We made  an  attempt  to  study  in  detail,  together with  six participating  school 

leaders of secondary schools with a  largely diverse student body, the requirements for 

an  accurate  match  between  existing  research  and  school  questions.  We  used  the 

method of  in‐depth,  semi‐structured  interviews  to  identify  school questions,  searched 

for a match with recent, high‐quality, practicable research publications, and carried out 

Page 192: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

166 

a  first  evaluation  of  the  actual  use  of  the  suggested  research  in  practice,  after  three 

months– admittedly a short period of time. 

All  but  three  out  of  21  questions  could  be  linked  to  high  quality  research 

outcomes; we found that most questions were of a conceptual nature, considering the 

best strategies for the school amidst its surrounding context of a multi‐ethnic society (all 

involved  schools  were  located  in  Amsterdam).  The  Principals  emphasized  that  they 

repeatedly had repeatedly encountered the problem that scientific research may report 

conflicting  findings  related  to  the  same  issue,  for  example  on  the  subject  of  school 

segregation: would  desegregating  schools  be  the  preferred  option? Or would  schools 

with a majority of migrant pupils be a better choice, since such school can specialize  in 

curriculum content (e.g. special attention to language proficiency) and the counselling of 

migrant  pupils  (e.g.  repeated  formative  assessments  to  enhance  upward mobility  to 

higher tracks)? After three months, five out of the six school leaders expressed the view 

that they had made use of the suggested research findings, notably considering strategic 

decisions, and communications with the School Board and external parties (for example, 

the municipality).  

We  encountered,  however,  the  rather  serious  technical  problem  that  only  a 

limited  number  of  research  publications  are  freely  accessible  via  academic  search 

machines: in many cases a licence fee has to be paid to a scientific journal in order to get 

access  to  full  text  reading;  such  fees  can  only  be  paid  by  universities  (or  other  large 

research institutions), and not by single schools. 

Principals  would  welcome  an  intermediate  function  between  schools  and 

academe, carried out by people who are well‐experienced in both worlds. 

6.2 Policy Implications 

6.2.1 A different perspective on segregation 

The  findings  presented  in  this  thesis  imply  considerations  for  policies  at  the 

government,  municipal  and  school  level.  Considering  the  societal  discourse  on  the 

segregation  of  schools  and  neighbourhoods,  our  results  show  a  relatively  small,  but 

significantly negative association between the percentage of 16 year‐old migrant pupils 

at  zip‐code  level  and  their  enrolments  in  the  two  highest  academic  secondary  tracks 

(Chapter 2). However, we also found a similarly small, but significant positive association 

Page 193: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

6 Conclusions and Discussion 

167 

between upward mobility  to a higher  track and choosing a  school closer  to  the home 

address (Chapter 4). In combination, these findings do not convincingly support policies 

to  desegregate  schools  by  centralized  school  assignments,  and  suggest  instead  that 

investing  in  increasing  the  chances  for  upward mobility  in  schools  located  in migrant 

neighbourhoods, may be worth  considering.  Interestingly, our  analyses of patterns of 

school choice reveal that migrant parents and pupils prefer to travel further to a school 

with an even larger percentage of migrant pupils than the school closest to their home. 

Finally,  our  investigation  of  the  usability  of  academic  knowledge  in  the  case  of  six 

secondary schools with a diverse student body shows that the Principals of these schools 

are  strongly  committed  to  using  the  best  scientific  knowledge  basis  for  the  specific 

function  of  their  school  vis‐à‐vis  the  migrant  communities  they  serve  (Chapter  5). 

Together, the outcomes of the different studies  in this thesis seem to make a case  for 

high‐quality, specialized schools that are well connected to migrant communities in the 

surrounding residential areas. 

6.2.2 Stable school careers 

Although we encountered a rather serious problem concerning school switching, 

in  retrospect  we  think  that  our  choice  to  start  with  a  visual  observation  of  the 

Amsterdam municipal data set, without any research question in advance, turned out to 

be  an  interesting  strategy. We  found  that  school  switching,  which  according  to  the 

literature,  is  in all cases harmful  to students, occurs more often among migrant pupils 

who  live  in socio‐economically disadvantaged neighbourhoods. Furthermore, we found 

that  switching  reduces  the  odds  of  upward mobility  to  a  higher  track,  an  important 

instrument for the narrowing of the educational achievement gap between migrant and 

native  Dutch  youth.  The  reduction  of  switching,  contrary  to  other  factors  that  are 

connected  to  the  achievement  gap  (e.g.  the  attainment  levels  of  parents)  seems  a 

manageable  strategy.  It may  be  considered,  at  the  level  of  the  government  and  the 

Inspectorate,  that  school attainment  levels  should be mentioned more prominently  in 

public  reports  on  the  quality  of  all  schools,  and  inform migrant  parents  in  particular, 

about the importance of this indicator. At the municipal level, the reduction of switching 

could  be  linked  to  large‐scale  programmes  to  prevent  dropout,  since  switching  is 

Page 194: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

168 

strongly  related  to  later  dropout.  Especially  those  schools  that  cooperate  in  clusters 

under the same board can relatively easily decide on policies that may reduce switching.  

6.2.3 Validation of academic knowledge 

A better  insight  into school questions, making  the  link  to manageable  research 

findings,  and  considering  usability  in  school  practice,  resulted  in  concrete 

recommendations by school  leaders for an  intermediate function between schools and 

academe. Therefore cooperation with an institute for school leadership is already being 

currently explored. At the same time, at the national level new initiatives are also being 

developed concerning the knowledge base for educational policies. Both developments 

may benefit from the mutual exchange of goals and strategies. 

6.2.4 The crucial importance of data 

Considering  that computing power  is no  longer a  limiting  factor, we hope  that 

our work based on the national database BRON, as the Amsterdam municipal database, 

may  also  bring  further  aspirations  to  make  national,  local,  and  school  data  sets 

compatible.  Schools  create  many  interesting  data  files  themselves,  but  may  need 

(financial)  support  to  further develop  the architecture of  their database and  invest  in 

professional skills to align their data to national data analyses. In this way schools could 

specifically monitor the performance indicators they consider to be relevant for decision 

making. 

The  BRON  data  set  is  being  developed  in  the  first  place,  for  the  analysis  of 

national  costs  and  the  outcomes  of  the  educational  system.  The  data  set  offers, 

however, enormous new opportunities for academic research that may add to not only 

measuring but also understanding the complexity of everyday school processes, with the 

goal to enhance equal opportunities for all pupils. 

6.3 Limitations and further research 

However much I would have liked to think otherwise during the long and at times 

almost desperate journey towards a doctorate, finishing a thesis is a mere debut in the 

world of academic research, not a completion. 

 There  are  obvious  limitations  to  the  studies  that  have  been  described  in  this 

thesis,  and,  more  important,  this  work  has  generated  new  questions  for  further 

Page 195: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

6 Conclusions and Discussion 

169 

research.  First,  the  scale of  the  research  carried out may have been  a  limitation:  the 

study  into switching resulted  in promising  findings  that could be scaled up  to  the  four 

major  cities  in  the  Netherlands.  Furthermore,  the  reduction  of  switching  offers 

possibilities  for  an  experimental  research  design,  which  would  involve  comparing 

schools with well ‐defined practices to reduce switching and other schools.  

Second,  the  nature  of  the  explored  research  questions may  require  follow‐up 

research: analyses of distance to school as a proxy  for the selectivity of school choice, 

have raised further questions , for example concerning the preference of migrant pupils 

and their parents for a school with a high percentage of migrant pupils: Do these pupils 

expect  that  schools  with  a  vast  majority  of  migrant  pupils  might  offer  specialized 

learning opportunities for migrant pupils, or might it be the case that parents prefer for 

their children to stay  in the same culture as at home, or even avoid stigmatization  in a 

school where their child might belong to a minority?  

And  third,  in  the  case  of  the  study  into  the  transfer  of  scientific  research 

outcomes  to school practice,  time was a  limiting  factor: we evaluated after only  three 

months  whether  the  Principals  had  made  use  of  insights  from  academic  research, 

whereas the search for the effects on school processes of a more solid knowledge base 

for school practice, requires monitoring over far longer periods of time.  

Page 196: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit
Page 197: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

171 

References 

Ala‐Mutka, K., Punie, P., Redecker, C. (2008). Digital Competences for Lifelong Learning 

(Report). European Commission. 

Allen, R. (2007). Allocating pupils to their nearest secondary school: The consequences 

for social and ability stratification. Urban Studies,  44 (4), pp. 751‐770. 

Ancess, J. (2000). The reciprocal influence of teacher learning, teaching practice, school 

restructuring, and student  learning outcomes. Teachers College Record, 102 (3), 

pp. 590‐619.    

Andersson, E., Malmberg, B., Östh, J. (2012). Travel‐to‐school distances in Sweden 2000‐

2006: changing school geography with quality  implications. Journal of Transport 

Geography, pp. 35‐43. 

Anger,  S.,  Heineck,  G.  (2009).  Do  smart  parents  raise  smart  children?  The 

intergenerational  transmission  of  cognitive  skills.  Journal  of  Population 

Economics 23 (3), pp. 1105‐1132. 

Archbald,  D.A.  (2004).  School  choice, Magnet  schools,  and  the  liberation model:  An 

empirical study. Sociology of Education, 77 (238).  

Bakker,  J.  (2012).  Cultureel‐etnische  segregatie  in  het  Onderwijs:  achtergronden, 

oorzaken en waarom te bestrijden? Landelijk Kenniscentrum Gemengde Scholen. 

Banks,  J.A.  (2007).  Educating  citizens  in  a multicultural  society.  New  York:  Teachers 

College Press. 

Page 198: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

172 

Banks,  J.A.  (Ed.).  (2009).  The  Routledge  International  Companion  to  Multicultural 

Education. New York, NY: Routledge. 

Bauder, H. (2002). Neighbourhood effects and cultural exclusion. Urban Studies, 39 (1). 

Bauer,  P.,  Riphahn,  R.T.  (2006).  Timing  of  school  tracking  as  a  determinant  of 

intergenerational transmission of education. Economics Letters, 91 (1), pp. 90‐97. 

Baysu, G., de Valk, H. (2012). Navigating the school system in Sweden, Belgium, Austria 

and  Germany:  school  segregation  and  second  generation  school  trajectories. 

Ethnicities, 20 July 2012. Retrieved from DOI: 10.1177/1468796812450857. 

Belfield,  C.R.,  Levin,  H.M.  (Eds).  (2007).  The  price  we  pay.  Economic  and  social 

consequences  of  inadequate  education. Washington D.C.:  Brookings  Institution 

Press.  

Björklund, A., Salvanes, K.G. (2010). Education and family background: Mechanisms and 

policies. IZA (Institute for the Study of Labour) Discussion paper 5002. 

Boyd, D., Grossman, P., Ing, M., Lankford, H., Wyckoff, J. (2009). The influence of school 

administration on  teacher  retention decisions. Retrieved August 31, 2011  from 

http://www.stanford.edu/.  

Brown, M.  R.  (2007).  Educating  all  students:  creating  culturally  responsive  teachers, 

classrooms, and schools. Intervention in School and Clinic, 43 (1), pp. 57‐62. 

Brown, K., Benkovitz,  J., Muttillo, A.J., Urban, T.  (2011).  Leading  schools of excellence 

and equity: documenting effective strategies in closing the gap. Teachers College 

Record, 113, No. 1, pp. 57‐96. 

Brunello, G., Checchi, D. (2007). School tracking and equality of opportunity. The Journal 

of Economic Policy. pp. 781‐861.  

Bunar,  N.  (2010).  Multicultural  schools  in  Sweden  and  their  communities:  social 

predicaments,  the power of  stigma, and  relational dilemmas. Urban Education, 

46 (2), pp. 141‐164. 

Bunar, N. (2010). The geographies of education and relationships in a multicultural city: 

enrolling  in  high‐poverty,  low‐performing  urban  schools  and  choosing  to  stay 

there. Nordic Sociological Association: Acta Sociologica. 

Cabrera, A.F.,  La Nasa,  S.M.  (2001). On  the path  to  college: Three  critical  tasks  facing 

America's disadvantaged. Research in Higher Education, 42 (2), pp. 119‐149. 

Page 199: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

References 

173 

Cheung,  C.K.,  Slavin,  R.E.  (2011).  The  effectiveness  of  education  technology  for 

enhancing reading achievement: a meta‐analysis. www.bestevidence.org. 

Coleman, J.S., Campbell, E.Q., Hobson, C.J., McPartland, J., Mood, A.M., Weinfeld, F.D., 

York,  R.L.  (1966).  Equality  of  educational  opportunity.  Washington  DC: 

Government Printing Office. 

Commissie  Nationaal  Plan  Toekomst  Onderwijswetenschappen  [Committee  for  a 

National Programme  for  the  Future of  Educational  Sciences]  (2011). Nationaal 

Plan Onderwijs/Leerwetenschappen.  Report retrieved from  www.vsnu.nl 

Cooper,  C.  R.  (2011).  Bridging multiple worlds.  Cultures,  identities,  and  pathways  to 

college. New York: Oxford University Press. 

Crul,  M.,  Holdaway,  J.  (2009).  Children  of  immigrants  in  schools  in  New  York  and 

Amsterdam: The  factors shaping attainment. Teachers College Record, 111, No. 

6, pp. 1476‐1507. 

Crul,  M.,  Schneider,  J.  (2009).  Children  of  Turkish  Immigrants  in  Germany  and  the 

Netherlands:  The  impact  of  differences  in  vocational  and  academic  tracking 

systems. Teachers College Record, 111, No. 6, pp. 1508‐ 1527. 

Crul, M., Schneider, J., Lelie, F. (Eds). (2012). The European second generation compared. 

Does the integration context matter? Amsterdam: Amsterdam University Press. 

Cullen,  J.B.,  Jacob,  B.,  Levitt,  S.  (2000).  The  Impact  of  School  choice  on  Student 

Outcomes.  An Analysis of the Chicago Public Schools.  NBER working paper 7888. 

Cambridge, MA: National Bureau of Economic Research. 

 Darling‐Hammond, L., Youngs, P. (2002). Defining "highly qualified teachers": what does 

"scientifically‐based research" actually tell us? Educational Researcher, 31 (9), pp. 

13‐25. 

De Witte, K. (2010). Dropout from secondary education: All's well that begins well. TIER 

working paper series 10/01. Retrieved from http://www.tierweb.nl  

Dee, T. S. (2004). Teacher, race, and student achievement in a randomized experiment. 

The Review of Economics and Statistics, 86(1): 195‐201. 

Dobbie,  W.,  Fryer,  R.G.  (2009).  Are  High  Quality  Schools  Enough  to  Close  the 

Achievement Gap? Evidence from a Social Experiment in Harlem. (NBER Working 

Paper  15473). Retrieved  from National Bureau of  Economic Research website: 

http://www.nber.org.papers/w15473 

Page 200: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

174 

  

Dronkers,  J.  (2010). Positive but also negative  effects of  ethnic diversity  in  schools on 

education  performance?  An  empirical  test  using  cross‐national  PISA  data. 

Inaugural  Lecture  at Maastricht  University.  Maastricht,  the  Netherlands:  Océ 

Business Services.  

Durlauf,  S.N.  (2004).  Neighbourhood  Effects.  In  Henderson,  J.V.,  &  Thisse,  J.F.  (Eds). 

(2004). Handbook of regional and Urban Economics, Volume 4. Amsterdam, the 

Netherlands: Elsevier. 

Education Council of the Netherlands (2011). Ruim baan voor stapsgewijze verbetering 

[Make way for gradual improvement]. The Hague: Onderwijsraad. 

Elk, R.  van,  Steeg, M.van  der, Webbink, D.  (2011). Does  the  timing  of  tracking  affect 

higher education completion? Economics of Education Review, 30, 1009‐1021. 

Furlong,  J.  (2010).  BERA  at  30.  Have  we  come  of  age?  British  Educational  research 

Journal, 30:3, pp. 343‐358. 

Gevers, J.K.M.  (1996). Over  iets te  ingewikkelds. Presidential Lecture at the opening of 

the  Academic  Year  1996  at  the  University  of  Amsterdam.  Amsterdam: 

Vossiuspers AUP. 

Gibbons,  S., Machin,  S.,  Silva, O.  (2008). Choice,  competition,  and pupil  achievement. 

Journal of the European Economic Association, 6 (4), pp. 912‐947. 

Gijsberts, M., Vervoort, M., Havekes, E. and Dagevos, J. (2010). Maakt de buurt verschil? 

(Does neighbourhood make the difference). The Netherlands  Institute for Social 

Research. The Hague: Textcetera. 

Glaeser, E.L., Resseger, M.G., Tobio, K.  (2008). Urban  Inequality.  (NBER Working Paper 

14419).  Retrieved  from  National  Bureau  of  Economic  Research  website: 

http://www.nber.org.papers/w14419. 

Gunderson,  L.G.,  Holling,  C.S.  (2002).  Panarchy:  Understanding  transformations  in 

human and natural systems. Washington D.C: Island Press.  

Grift, W.J.C.M.  van  der  (2010). Ontwikkeling  in  de  beroepsvaardigheden  van  leraren. 

Rijksuniversiteit Groningen, Oratie. 

Groot, W., Maassen van den Brink, H. (2003). Investeren en terugverdienen. The Hague: 

Sectorbestuur Onderwijsarbeidsmarkt. 

Page 201: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

References 

175 

Hanushek, E. A., Rivkin, S.G.  (2007). Pay, working  conditions, and  teacher quality. The 

Future of Children, 17 (1), pp. 69‐86. Princeton University.  

Hanushek, E.A., Kain, J.F., & Rivkin, S.G. (2001). Disruption versus Tiebout improvement: 

the  costs  and  benefits  of  switching  schools.  (NBER  working  paper  8479). 

Retrieved  from  National  Bureau  of  Economic  Research  website: 

http://www.nber.org/papers/w8479 

Hanushek, E., Woessman, L.  (2006). Does educational  tracking affect performance and 

inequality?  Difference‐in‐difference  evidence  across  countries.  The  Economic 

Journal, 116 (March), C63‐C76.  

Harris,  R.,  Johnston,  R.  and  Burgess,  S.  (2007). Neighbourhoods,  ethnicity  and  school 

choice:  Developing  a  statistical  framework  for  geodemographic  analysis. 

Population Research and Policy Review, 26 (6/6), pp. 553‐579. 

Hartog,  J.,  Zorlu,  A.  (2009).  Ethnic  segregation  in  the  Netherlands:  analysis  at 

neighbourhood level. International Journal of Manpower, 30, (1/2), pp. 15‐25. 

Heckman,  J.  J.  (2008).  Schools,  Skills,  and  Synapses.  (NBER  Working  Paper  14064). 

Retrieved  from  National  Bureau  of  Economic  Research  website: 

http://www.nber.org.papers/w14064. 

Heckman, P.E., Montera V.L.  (2009). School reform:  the  flatworm  in a  flat world:  from 

entropy to renewal through  indigenous  invention. Teachers College Record, 111 

(5), pp. 1328‐1351. 

Heineck,  G.,  Riphahn,  R.T.  (2007).  Intergenerational  transmission  of  educational 

attainment  in Germany: The  last five decades.  Institute for the Study of Labour, 

Discussion Paper no. 2985.  Retrieved from http://ssrn.com/abstract=1012343. 

Herweijer,  L.  (2009). Making  up  the  gap: Migrant  education  in  the Netherlands.  The 

Netherlands Institute for Social Research. The Hague: Textcetera.  

Heus, M.  de,  Dronkers,  J.  (2008).  Can  Educational  Systems  Compensate  for  Societal 

Features?  The  Effects  of  Educational  Systems  and  Societal  Features  of  both 

Countries  of  Origin  and  Destination  on  the  Scientific  Literacy  of  Immigrant 

Children  in Western Countries. Paper presented at the  International Conference 

on  Survey Methods  in Multinational, Multiregional,  and Multicultural  Context. 

Berlin 2008. 

Page 202: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

176 

Ho  Sui‐Chu,  E., Willms,  J.D.  (1996).  Effects  of  parental  involvement  on  eighth‐grade 

achievement. Sociology of Education, 69 (2), pp. 126‐141. 

Inspectorate  of  Education  (2011).  Onderadvisering  van  allochtone  leerlingen?  [under‐ 

advising of migrant students?]. Report. 

Jackson, K. C., Bruegmann, E.  (2009). Teaching  students and  teaching each other:  the 

importance  of  peer  learning  for  teachers.  NBER Working  Paper  Series,  paper 

15202. Cambridge, MA: National Bureau of Economic Research. 

Johnson, R. B.  (1998). Toward a  theoretical model of evaluation utilization. Education 

and Programme Planning, 21, pp. 93‐110. 

Kerbow, D., Azcoitia, C., Buell, B. (2003). Student mobility and local school improvement 

in Chicago. The Journal of Negro Education, 72, No. 1, pp. 158‐ 164. 

Konstantopoulos,  S.,  Borman,  G.  (2011).  Family  background  and  school  effects  on 

student achievement: a multilevel analysis of the Coleman data. Teachers College 

Record, 113, No. 1, pp. 97‐132. 

Ladd,  H.F.,  Fiske,  E.,  Ruijs,  N.  (2009).  Parental  choice  in  the  Netherlands:  Growing 

concerns  about  segregation.  Paper  prepared  for  the  National  Conference  on 

School Choice, Vanderbilt University.  

Ladd,  H.F.,  Fiske,  E.B.,  Ruijs,  N.  (2010).  Parental  Choice  in  the  Netherlands:  growing 

concerns about segregation. Sanford Working Papers Series (SAN10‐02). 

Landry,  R.,  Amara, N.,  Lamari, M.  (2001).  Climbing  the  ladder  of  research  utilization: 

evidence  from  social  science  research.  Science  Communication.  SAGE 

Publications. Retrieved from http://scx.sagepub.com/content/22/4/396 

Leigh  McGarrigle,  J.  (2010).  Understanding  processes  of  ethnic  concentration  and 

dispersal.  South  Asian  residential  preferences  in  Glasgow.  Amsterdam,  the 

Netherlands: Amsterdam University Press. 

Levin, H. M.  (2011).  The economics of education. Albany Government  Law Review, 4, 

pp.395‐425. 

Louis, K., Wahlstrom, K.L., Michlin, M., Gordon, M., Thomas, E., Leithwood, K., Anderson 

S.  E.,  Mascall,  B.,  Strauss,  T.,  Moore,  S.  (2010).  Learning  from  leadership: 

Investigating  the  links  to  improved student  learning. Final report of research  to 

the Wallace  Foundation. University of Minnesota: Center  for Applied Research 

and Educational improvement. Retrieved from www.cehd.umn.edu/CAREI/ 

Page 203: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

References 

177 

Margolis,  J.  (2010). Why  teacher quality  is a  local  issue  (and why Race  to  the Top  is a 

misguided  flop).  Teachers  College  Record,  date  published  June  23,  2010. 

Retrieved from http://www.tcrecord.org ID Number: 16023. 

Maroulis, S., Gomez, L. M. (2008). Does "Connectedness" matter? Evidence from social 

network analysis within  small‐school  reform. Teachers College Record, 110, pp. 

1901‐1929. 

Massey, D. (2001). The Prodigal Paradigm Returns: Ecology Comes Back to Sociology. In 

Booth, A., Crouter, A.C.  (Eds).  (2008). Does  it take a village? Hove, East Sussex: 

Psychology Press. 

May, S.  (2009). Critical multiculturalism and education.  In Banks,  J.A.  (Ed.).  (2009). The 

Routledge  International  Companion  to Multicultural  Education.  New  York,  NY: 

Routledge. 

McIntyre, D. (2006). Bridging the gap between research and practice. Cambridge Journal 

of Education, 35:3, pp. 357‐382. 

McKinsey & Company  (2007). How  the World's Best‐performing  School  Systems Come 

out on Top (Report).  

McKinsey & Company (2009). The economic Impact of the Achievement Gap in America's 

Schools. Australia: New Media. 

Miller  Marsh,  M.,  &  Turner‐Vorbeck,  T.  (Eds),  (2010).  (Mis)understanding  families: 

Learning  from  real  families  in our  schools. New York, NY, Columbia University: 

Teachers College Press. 

Morrell, E., Noguera, P. (2011). A framework for change: a broader and bolder approach 

to school reform. Teachers College Record, ID Number 16503. 

Murnane, R.J. (2009). Educating urban children. In Inman, R.P (Ed.). Making cities work. 

Prospects for urban America. Princeton NJ: Princeton University Press.  

Nakagawa, K., Stafford, E., Fisher, A., Matthews, L. (2002). The "City Migrant" dilemma: 

Building community at high‐mobility urban schools. Urban Education 37: 96. 

Nilsson, M. E.  (2008). University‐school collaboration based on complementary needs. 

Journal for Educational Change, 9, pp. 349‐356. 

Noguera,  P.A.  (2004).    Social  Capital  and  the  Education  of  Immigrant  Students: 

Categories and Generalizations. Sociology of Education, 77 (2), pp. 180‐183. 

Page 204: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

178 

Notten, A. (2009). Desegregation  in schools– How can  it be achieved? Journal of Social 

Intervention: Theory and Practice, 18 (4), pp. 83‐93. 

Nutley,  S.,  Walter,  I.,  Davies,  H.  (2002).  From  knowing  to  doing:  a  framework  for 

understanding the evidence‐into‐practice agenda. Discussion paper. University of 

St. Andrews: Research Unit for Research Utilization. 

OECD  (2010),  PISA  2009  Results:  What  Students  Know  and  Can  Do  –  Student 

Performance  in  Reading, Mathematics  and  Science  (Volume  I).  Retrieved  from 

http://dx.doi.org/10.1787/9789264091450‐en 

OECD (2010). OECD Reviews of Migrant Education:  Netherlands. Paris: OECD  

OECD  (2010). Reviews of Migrant Education: Country Note  for  the   Netherlands. Paris: 

OECD. 

OECD (2011). Education at a Glance. Paris: OECD. 

Oosterbeek, H.  (2001). Voortschrijdend  Inzicht  [Progressive  Insight].  Inaugural  lecture. 

Amsterdam: Vossius Press. 

Opdenakker, M.C., Van Damme, J. (2001). Relationship between school composition and 

characteristics of school process and  their effect on mathematics achievement. 

British Education Journal, 27 (40), pp. 407‐432. 

Oreopoulos, Ph.  (2007). Neighbourhood Effects  in Canada: A Critique. Paper presented 

at the 2007 Statistics Canada Socio‐economic Conference. 

Palardy,  G.J.  (2008).  Differential  effects  among  low,  middle,  and  high  social  class 

composition  schools: a multiple group, multilevel  latent growth  curve analysis. 

School Effectiveness and School Improvement, 19 (1), pp. 21‐49. 

Parker,  W.C.,  Ninomiya,  A.  and  Cogan,  J.  (1999).  Educating  world  citizens:  toward 

multinational curriculum development. American Educational Research  Journal, 

36:117. 

Payne, C.M. (2010). So much reform, so little change: The persistence of failure in urban 

schools. Cambridge, MA: Harvard Education Press. 

Pekkarinen, T., Uusitalo, R. Kerr, S. (2009). School tracking and intergenerational income 

mobility: evidence  from Finnish comprehensive school reform. Journal of Public  

Economics, 93, pp. 965‐973. 

Rumberger, R. W.,  Larson,  L. A.  (1998). Student mobility and  the  increase  risk of high 

school dropout. American Journal of Education, 107, (1), pp. 1‐35. 

Page 205: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

References 

179 

Rumberger,  R.W.,  Palardy,  G.J.  (2005).  Does  segregation  still matter?  The  impact  of 

student composition on academic achievement  in high school. Teachers College 

Record, 107, (9), pp. 1999‐2045. 

Rumberger, R.W., Thomas, S.L.  (2000). The distribution of dropout and  turnover  rates 

among urban and  suburban high  schools. Sociology of Education, 73, No.1, pp. 

39‐67. 

Sanders,  M.  (2009).  Collaborating  for  change:  how  an  urban  school  district  and  a 

community‐based  organization  support  and  sustain  school,  family,  and 

community partnerships. Teachers College Record, ID Number 15304. 

Sanders, M. G., Harvey, A.  (2002).  Beyond  the  school walls:  a  case  study  of  Principal 

leadership for school‐community collaboration. Teachers College Record, 104 (7), 

pp. 1345‐1368. 

Saunders, W.M.,  Goldenberg,  C.  N.,  Gallimore,  R.  (2009).  Increasing  achievement  by 

focusing  grade‐level  teams  on  improving  classroom  learning:  a  prospective, 

quasi‐experimental  study  of  Title  I  schools.  American  Educational  Research 

Journal, 46 (4), pp. 1006‐1033. 

Scheffer,  P.  (2000).  Het  culturele  drama  [The  cultural  drama].  NRC  Handelsblad  29 

January 2000. 

Schön, D.A.  (1983). The  reflective practitioner: How professionals  think  in action. New 

York: Basic Books. 

Scott,  B.  (2001).  Coming  of  age.  Intercultural  Development  Research  Association 

Newsletter, pp. 1‐8. 

Shewbridge, C., Kim, M., Wurzburg, G., Hostens, G.  (2010). OECD Reviews of Migrant 

Education: Netherlands. Paris, France: OECD. 

Slavin, R.E. (2002). Evidence‐based education policies: transforming educational practice 

and research. Educational Researcher, 31:7, pp. 15‐21. 

Spinoza,  Benedict  de,  (1670)  and  Jonathan  Israel,  Editor,  (2007).  Theological‐Political 

Treatise. Cambridge, UK: Cambridge University Press. 

Stuart Wells, A., Jellison Holme, J., Tijerina Revilla, A. and Korantemaa Atanda, A. (2009). 

Both sides now: The story of school desegregation's graduates. Los Angeles, CA: 

California Press. 

Page 206: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

180 

Swanson, C. B., Schneider, B. (1999). Students on the move: residential and educational 

mobility in America's schools. Sociology of Education, 72, (1), pp. 54‐67. 

Tefera,  A.,  Frankenberg,  E.,  Siegel‐Hawley,  G.,  Chirichigno,  G.  (2011).  Integrating 

suburban schools: how to benefit from growing diversity and avoid segregation. 

University  of  California  in  Los  Angeles:  Civil  Rights  project. 

www.civilrightsproject.ucla.edu. 

Thrupp,  M.,  Lauder,  H.,  Robinson,  T.  (2003).  School  composition  and  peer  effects. 

International Journal of Educational Research, 37, pp. 483‐504. 

Tiebout,  C.M.  (1956).  A  Pure  Theory  of  Local  Expenditures.  The  Journal  of  Political 

Economy, 64, pp. 416‐424. 

Tweede kamer der Staten Generaal, Vergaderjaar 2006‐07, 30 995, nr. 5. Brief van de 

minister  van Wonen, werken en  Integratie  (2007). Aanpak Wijken. The Hague: 

SDU Uitgeverijen. 

Välijärvi, J., Sahlberg, P. (2008). Should "failing" students repeat a grade? Retrospective 

response from Finland. Journal of Educational Change, 9 (4), pp. 385‐389. 

Vanderlinde,  R,  Van  Braak,  J.  (2009).  The  gap  between  educational  research  and 

practice:  views  of  teachers,  school  leaders,  intermediaries  and  researchers. 

British Educational Research Journal, 36:2, pp. 299‐316. 

Vigdor, J., Ludwig, J. (2010). Neighbourhoods and peers  in the production of schooling. 

In Brewer, D.J. and McEwan, P.J. (Eds.). Economics of Education. Amsterdam, the 

Netherlands: Elsevier. 

Welie,  E.A.A.M.  van,  Hartog,  J.  and  Cornelisz  I.  (2013).  Free  school  choice  and  the 

educational achievement gap. Journal of School Choice. Forthcoming. 

Wells,  R.  (2010).  Children  of  immigrants  and  educational  expectations:  The  roles  of 

school composition. Teachers College Record, 112 (6), pp. 1679‐1704. 

Werfhorst, H.G. van de, Hofstede, S. (2007). Cultural capital or relative risk aversion? The 

British Journal of Sociology, 58 (3), pp. 391‐415. 

Werfhorst, van de, H.G. and Van Tubergen, F. (2007). Ethnicity, schooling, and merit  in 

the Netherlands. Ethnicities, 7, pp. 416‐443. 

Yin, R.K.  (2009). Case  study  research. Design and methods.  Thousand Oaks, CA:  SAGE 

Publications Inc. 

Page 207: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

References 

181 

Zimmerman, B.  J.  (2002). Becoming a  self‐regulated  learner: an overview. Theory  into 

Practice, 41 (2). 

Zohar, A., Dori, Y.  (2003). Higher order  thinking  skills  and  low‐achieving  students:  are 

they mutually exclusive? The  Journal of  the  Learning Sciences, 12  (2), 145‐188.

Page 208: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit
Page 209: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

183 

Samenvatting 

Dit proefschrift bestaat uit  vier  studies naar de positie  van  leerlingen met een 

migranten achtergrond  in het voortgezet onderwijs, die wonen  in de vier grote steden 

Amsterdam,  Rotterdam, Utrecht  en Den Haag. De  onderwijsprestaties  van migranten 

leerlingen, en de segregatie van scholen en buurten, zijn al enkele decennia onderwerp 

van verhit debat  in Nederland. Crul, Schneider en Lelie (2009)  laten zien dat de kansen 

voor  een  12‐jarige  migranten  leerling  (hun  studie  gaat  voornamelijk  over  Turkse 

leerlingen) om geconfronteerd te worden met schoolproblemen die uiteindelijk kunnen 

leiden tot uitval, vrijwel even groot zijn als de kansen om uitzonderlijk succesvol te zijn in 

het  onderwijs,  en  het  opleidingsniveau  van  de  ouders  verre  te  overtreffen.  In  het 

maatschappelijke  debat  echter,  lijkt  deze  succesvolle  groep  vrijwel  aan  het  zicht 

onttrokken  te  zijn. Het  is  zorgwekkend  dat  niet  zelden  in  categorische  termen wordt 

gesproken  over  problemen  van  bijvoorbeeld  Marokkaanse  leerlingen,  die  zouden 

kunnen suggereren dat hun Marokkaanse afkomst de hoofdoorzaak van hun problemen 

is;  tevens  lijkt  er  geen  rekening  mee  te  worden  gehouden  dat  de  leerlingen  met 

schoolproblemen,  niet  dezelfde  individuen  zijn  als  de  succesvolle  leerlingen.  Verder 

laten nationale statistieken (Ministerie van Onderwijs, Cultuur en Wetenschappen) een 

gestage toename van het percentage migranten leerlingen op HAVO en VWO zien. 

Page 210: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

184 

Het zal duidelijk zijn dat een opleiding die geen recht doet aan het niveau van je 

capaciteiten,  en  die  leidt  tot  een  positie  op  de  arbeidsmarkt  waarin  je  je  ambities 

onvoldoende kunt verwezenlijken, voor  iedere  jongere een groot persoonlijk drama  is. 

Daarnaast echter, vormen  jongeren met een migranten achtergrond een groot aandeel 

van het  toekomstige arbeidspotentieel; zeker  in een vergrijzende samenleving hebben 

wij hen hard nodig. Tenslotte  tonen Groot en Maassen van den Brink  (2003) en Levin 

(2011)  overtuigend  aan  dat  geen  investering  een  grotere  opbrengst  genereert  dan 

investeringen in het onderwijs, en dat de kosten van falen zeer hoog zijn. Een belangrijke 

bijdrage  van  hun  werk  is  het  pleidooi  om  de  opbrengsten  van  het  onderwijs  te 

berekenen  over  het  totale  werkende  leven  van  een  individu,  en  niet  slechts  als  de 

kosten per capita per jaar in het funderend onderwijs. 

De  vier  studies  in  dit  proefschrift  onderzoeken  1)  of  kenmerken  van  de  buurt 

waar  migranten  jongeren  wonen,  samenhangen  met  hun  niveau  in  het  voortgezet 

onderwijs, 2) hoe een stabiele school carrière, niet onderbroken door een overstap naar 

een  andere  school,  samenhangt  met  succes,  3)  hoe  school  keuze,  gemeten  als  de 

reisafstand  naar  een  school  die  wordt  verkozen  boven  de  dichtstbijzijnde  school, 

verschilt  tussen  diverse  etnische  groepen,  en  4)  hoe  de  transfer  van  toepasbare 

wetenschappelijke  kennis  scholen met  veel migranten  leerlingen  kan ondersteunen  in 

hun werk.  

In  hoofdstuk  2  wordt  beschreven  hoe  het  percentage  16‐jarige  Nederlands‐

Marokkaanse  jongeren, gemeten op het niveau van het 4‐cijferige postcodegebied van 

hun woonadres,  samenhangt met  deelname  aan  HAVO  of  VWO.  Deze metingen  zijn 

uitgevoerd voor alle postcodegebieden in de vier grote steden, en daarnaast apart voor 

postcodegebieden  met  geaccumuleerde  sociaaleconomische  problemen.  Tenslotte  is 

onderzocht  hoe  de  deelname  aan  HAVO  of  VWO  onder  Nederlands‐Marokkaanse 

leerlingen  samenhangt  met  het  totale  percentage  16‐jarigen  met  een 

migrantenachtergrond binnen het postcodegebied waar zij wonen. Ten behoeve van dit 

onderzoek is een data set samengesteld uit CBS data (Centraal Bureau voor de Statistiek) 

en data beschikbaar gesteld door DUO (Dienst Uitvoering Onderwijs, voorheen CFI). 

De associatie tussen de etnische compositie op postcodeniveau en deelname aan 

HAVO  of  VWO  (als maat  voor  onderwijssucces) werd  geanalyseerd  door middel  van 

LOWESS  regressies  (Locally  Weighted  Scatterplot  Smoothing)  en  lineaire  regressies 

Page 211: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

Samenvatting 

185 

(OLS). De  resultaten  laten  zien dat er over het geheel een negatief  verband  is  tussen 

deze  twee  variabelen,  de  dispersie  van  de  datapunten  is  echter  substantieel.  De 

gemiddelde  deelname  aan  HAVO  of  VWO  onder  Nederlands Marokkaanse  leerlingen 

was    in   2009 20.5%,  tegen 45.1% onder Nederlandse  leerlingen  (Bron: Ministerie van 

Onderwijs,  Cultuur  en  Wetenschappen).  Met  de  kanttekening  dat  deze  nationale 

gemiddelden niet zondermeer vergeleken kunnen worden met de deelpopulatie die  in 

dit onderzoek is onderzocht, valt op dat alleen in postcodegebieden met rond de 40% of 

meer Nederlands‐Marokkaanse jongeren als aandeel van de populatie van 16‐jarigen, de 

deelname  onder  dit  landelijke  gemiddelde  komt.  De  verschillen  tussen  alle 

postcodegebieden‐ en probleemwijken  in het bijzonder, zijn marginaal. De bevindingen 

in  dit  hoofdstuk  lijken  geen  stevige  onderbouwing  te  bieden  voor  een  actief 

desegregatie  beleid  rond  schoolplaatsing.  Om  verder  te  onderzoeken  welke  andere 

factoren, naast etniciteit, schoolkeuze en niveau  in het voortgezet onderwijs sturen,  is 

een breder onderzoek uitgevoerd, dat in hoofdstuk 4 wordt gepresenteerd. 

In  het  maatschappelijke  debat  over  de  onderwijs  prestaties  van  migranten 

leerlingen,  met  name  Nederlands‐Marokkaanse  leerlingen,  stonden  enkele  postcode 

gebieden  in Amsterdam West  in het centrum van de belangstelling. Door de aandacht 

voor veel  zorgen en problemen  in dit  stadsdeel,  leken ook hier  succesvolle migranten 

leerlingen  geheel  buiten  beeld  te  blijven,  of  werden  zij  geportretteerd  als  de 

uitzondering  die  de  regel  bevestigt. Om  een  beter  beeld  te  krijgen  van  jongeren  die 

wonen  in  postcode  1061,  1062  en  1063  in  Amsterdam  West,  werd  een  data  set 

geanalyseerd  die  beschikbaar  werd  gesteld    door  de  Dienst  Maatschappelijke 

Ontwikkeling  (DMO)  van  de  gemeente  Amsterdam.  Deze  data  set  bevat  een 

gedetailleerde  schoolloopbaan  historie,  met  alle  exacte  in‐  en  uitschrijfdata  in  het 

onderwijs, van alle 16‐22  jarigen woonachtig  in de bovengenoemde drie postcodes op 

de  peildatum  31  juli  2009.  Bij  de  start  van  dit  onderzoek  waren  er  geen 

onderzoeksvragen  vooraf,  maar  gingen  we  op  zoek  naar  eventuele  opvallende 

fenomenen  in  de  data. Onmiddellijk,  al  bij  eerste  beschouwing  van  de  data,  viel  het 

opvallend  grote  aantal  in‐  en  uitschrijvingen  op  een  school  op,  in  het  geval  van 

substantiële aantallen  individuen; dit  inspireerde onze hoofdvraag, hoe  switchen naar 

een  andere  school  de  resultaten  van  leerlingen  beïnvloedt.  De  tellingen  die worden 

gepresenteerd  in  dit  hoofdstuk  laten  verschillende  frequenties  van  switching  tussen 

Page 212: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

186 

etnische  groepen  zien,  en  een  substantieel  verschil  tussen  migranten  leerlingen  en 

Nederlandse  leerlingen.  In het geval van Marokkaanse  leerlingen behaalde 87.4% een 

diploma  op  HAVO/VWO  niveau, wanneer  zij  hun  gehele  school  periode  op  dezelfde 

school bleven. Onder de Marokkaanse  leerlingen die hun derde school voor voortgezet 

onderwijs verlieten, had nog slechts 39.1% een diploma behaald.  

Om  switchen  in  de  drie  onderzochte  postcodes  te  kunnen  vergelijken  met  

switchen  in de  rest van Amsterdam,  is een  tweede  serie analyses uitgevoerd op basis 

van  de  BRON  data  (burger  service  nummer  en  onderwijsnummer;  DUO,  Ministerie 

OCW). Deze analyses laten zien dat met name Nederlands Marokkaanse leerlingen die in 

de drie postcodegebieden wonen, vaker switchen dan Marokkaanse leerlingen in de rest 

van Amsterdam Over het algemeen switchen migranten leerlingen vaker dan leerlingen 

van  Nederlandse  origine,  hoewel  er  verschillen  zijn  tussen  etnische  groepen.    In 

ongeveer  een  derde  van  de  gevallen  switchen  leerlingen  binnen  een  groep  van 

samenwerkende scholen.  

De  BRON  data maken  het mogelijk  om  de mobiliteit  naar  een  hoger  of  lager 

niveau  in  het  voortgezet  onderwijs  te  meten,  als  de  leerling  in  de  derde  klas  zit. 

Substantieel  meer  leerlingen  die  opstromen,  blijven  op  dezelfde  school.  Onder  de 

switchers  neemt  de  opstroom  af,  en  neemt  de  afstroom  toe.  Voor  Marokkaanse 

leerlingen  in Amsterdam,  is het verschil  in opstroom  tussen  leerlingen die niet‐ en die 

wel  switchen,  bijna  12  procentpunt.  In  tegenstelling  tot  de  hoge  kosten  die  gemoeid 

gaan  met  het  terugdringen  van  drop‐out,  zou  het  reduceren  van  switching  vrijwel 

kostenneutraal  kunnen  zijn,  omdat  het  succes  vooral  afhangt  van  goede 

overeenkomsten tussen scholen, en transparantie over het beleid om  leerlingen al dan 

niet vrijwillig over te plaatsen naar een andere school. 

Hoofdstuk  4  presenteert  een  studie  naar  patronen  van  schoolkeuze  in  het 

voortgezet  onderwijs.  In  Nederland  is  de  schoolkeuze  vrij,  (behoudens  enkele 

ontwikkelingen  rond  loting  bij  over‐intekening),  en  zijn  er  geen  financiële 

belemmeringen voor ouders, omdat het funderend onderwijs volledig door de overheid 

wordt gefinancierd. Ten behoeve van dit onderzoek hebben wij een zeer rijke BRON data 

set  gebruikt, met  de  gegevens  van  16.071  leerlingen  die  in  2008  in  groep  8  van  de 

basisschool zaten, en ten tijde van ons onderzoek in de meeste gevallen in de derde klas 

van het voortgezet Onderwijs zaten. Voor een deel van de leerlingen waren ook de CITO 

Page 213: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

Samenvatting 

187 

score en het advies van de docent van groep 8  in het basis onderwijs beschikbaar. De 

BRON data set met uitgebreide (volledig geanonimiseerde) gegevens op het individuele 

niveau,  werd  gekoppeld  aan  sociaal  economische  gegevens  op  buurtniveau  van  het 

Sociaal Cultureel Planbureau (SCP) en het Centraal Bureau voor de Statistiek (CBS)  , en 

kwaliteitsgegevens op school niveau van de Inspectie van het Onderwijs. Op deze wijze 

konden  wij  schoolkeuze  patronen  analyseren  op  basis  van  regressies  met  controle 

variabelen op het  individuele‐, buurt‐ en school niveau. 

Wij keken achtereenvolgens naar 1) de associatie tussen de gemiddelde afstand 

naar  school  en  individuele‐  en  sociaaleconomische  variabelen  voor  verschillende 

etnische groepen, 2) mogelijke verschillen tussen de groep die de dichtstbijzijnde school 

kiest en de groep die een andere school prefereert, 3) in het geval van leerlingen die niet 

de  dichtstbijzijnde  school  kiezen,  het  verschil  in  afstand  tussen  de  gekozen  en  de 

dichtstbijzijnde  school,  4)  het  verschil  in  kenmerken  tussen  de  gekozen  en  de 

dichtstbijzijnde  school en 5) de associatie  tussen  schoolkeuze en opwaartse mobiliteit 

naar een hoger niveau in het voortgezet, op individueel niveau.  

Onze resultaten  laten zien dat migranten  leerlingen gemiddeld minder ver naar 

school  reizen  dan Nederlandse  leerlingen.  Echter,  in  postcodegebieden met  een  lage 

gemiddelde SES (Sociaal Economische Situatie) reizen migranten leerlingen nog minder, 

maar  Nederlandse  leerlingen  juist  verder.  Daarentegen  in  rijke  buurten,  is  de 

gemiddelde afstand naar  school onder Nederlandse  leerlingen  laag. Enigszins  tot onze 

verrassing,  vonden  wij  geen  systematische  verschillen  tussen  de  groepen  die  de 

dichtstbijzijnde‐ of een andere school kozen. Maar er bleek een opmerkelijke verschil in 

school    preferenties  tussen  Nederlandse  leerlingen  en migranten  leerlingen  die  een 

school verder weg kozen: Nederlandsen leerlingen kozen gemiddeld een school met een 

lager  percentage migranten  leerlingen,  een  hogere  school  SES,  en  een  hoger  school 

eindexamen  gemiddelde‐  in  vergelijking  met  de  dichtstbijzijnde  school.  Bijvoorbeeld 

Nederlands‐Marokkaanse leerlingen daarentegen, kozen een school met een nog hoger 

percentage migranten leerlingen dan hun dichtstbijzijnde school, zelfs ten koste van een 

lagere  school  SES.  Mogelijk  hebben  wij  hier  patronen  van  zelfverkozen  segregatie 

zichtbaar gemaakt, een inspiratie voor vervolgonderzoek. Hoewel het verschil klein is, is 

de  kans  op  opwaartse  mobiliteit  (vooral  voor  migranten  leerlingen  een  belangrijk 

Page 214: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

   

188 

mogelijkheid  om  de  prestatiekloof  te  versmallen)  statistisch  significant  iets  groter 

naarmate leerlingen dichter bij huis naar school gaan.  

 In  hoofdstuk  5  tenslotte, wordt  een  onderzoek  gepresenteerd  naar  de wijze 

waarop  toepasbare  resultaten  van  bestaand  wetenschappelijk  onderzoek,  gekoppeld 

kunnen worden aan actuele vragen van scholen met veel migranten leerlingen. Het idee 

voor dit onderzoek is ontstaan tijdens mijn promotie traject: vrijwel dagelijks verschijnen 

er wetenschappelijke publicaties over groepen  leerlingen die nog onvoldoende kansen 

hebben gekregen om op de juiste school, of het juiste niveau terecht te komen, gegeven 

hun  capaciteiten.  Veel  van  deze  inzichten  uit  wetenschappelijk  onderzoek  zouden 

vertaald  kunnen worden naar  innovatieve handelingspraktijken  in  scholen, en  zouden 

schoolleiders  en  docenten  kunnen  ondersteunen  in  het  nemen  van  strategische 

beslissingen. Echter, uitkomsten van wetenschappelijk onderzoek bereiken scholen niet 

of nauwelijks.   

Samen met  6  schoolleiders  in  het  voortgezet  onderwijs  in  Amsterdam,  is  een 

methode  ontworpen  en  getoetst  om  schoolvragen  te  verbinden  wetenschappelijke 

kennis.  1)  Op  basis  van  een  semigestructureerd  diepte‐  interview  met  individuele 

schoolleiders,  werden  belangrijke  schoolvragen  geïdentificeerd;  2)  in  de 

wetenschappelijke  literatuur werd gezocht naar een match met deze  thema's; 3) deze 

wetenschappelijke  inzichten  werden  gemotiveerd  per  brief  gerapporteerd  aan  de 

schoolleiders;  4)  na  3  maanden  werd  nagegaan  of  deze  aangedragen  kennis  was 

toegepast.  De  resultaten  laten  zien  dat  schoolleiders  inderdaad  belangrijke  vragen 

hebben,  die  in  vrijwel  alle  gevallen  konden  worden  gekoppeld  aan  toepasbare 

wetenschappelijke  inzichten.  De  meeste  vraagstukken  waren  van  conceptuele  of 

strategische aard, bijvoorbeeld over de  functie van de  school  in de  context van grote 

maatschappelijke vraagstukken  rond diversiteit; er waren nauwelijks vragen van meer 

instrumentele  aard,  wat  mogelijk  samen  kan  hangen  met  het  feit  dat  alleen 

schoolleiders  werden  geïnterviewd.  Schoolleiders  pleitten  voor  een  intermediaire 

functie  tussen het onderwijsveld en de  academische wereld, die naar hun  inzicht het 

beste  vervuld  zou  kunnen  worden  door  mensen  die  beide  werelden  van  binnenuit 

kennen.  Verder  hechtten  verschillende  schoolleiders  aan  vervolgonderzoek  naar  de 

concrete vertaling van wetenschappelijke inzichten naar de lespraktijk in de klas. 

Page 215: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

Samenvatting 

189 

Tijdens ons onderzoek kwamen wij een vrij ernstige belemmering voor scholen 

op het spoor: veel publicaties in wetenschappelijke tijdschriften zijn slechts beperkt vrij 

toegankelijk via het  internet, en kunnen alleen binnen de muren van een universiteit‐ 

waar  aanzienlijke  licence  fees worden betaald‐ worden  geraadpleegd. Deze beperking 

maakt  het  moeilijk  voor  schoolleiders  om  zelf  te  zoeken  naar  relevante 

wetenschappelijke publicaties. 

Page 216: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit
Page 217: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

191 

Biography 

Liesbeth van Welie graduated in Biology from Radboud University (Nijmegen, the 

Netherlands), with a specialization in Aquatic Ecology. 

 She  started  her  career  as  a  teacher  in  Biology  and  a  few  years  later  was 

appointed as Principal of an  innovative newly‐founded school for secondary education.  

Her second appointment as Principal was at a diverse  inner‐city school  in Amsterdam, 

where she devoted her work to educating migrant pupils for successful access to higher 

education.  

Next,  she was  invited by  the Board of  the University of Amsterdam,  to  start a 

large programme for improving the quality of academic teaching and learning, as Senior 

Vice‐President  for  Education.  During  these  years,  she  engaged  in  an  extended 

international  network  in  the  field  of  quality  assessment  of  higher  education, 

internationalization and organizational change. Her next step was advisor to the Board 

at Maastricht University,  followed by  two  years  as  senior  consultant  at  a  consultancy 

firm, where her portfolio consisted of assignments at several Dutch universities. 

For  several  years  she  held  the  position  of  Chief  Inspector  of  Secondary  and 

Higher  Education,  where  after  the  Secretary  General  of  the  Ministry  of  Education, 

Culture and Science, invited her to make an organisational design for the enhancement 

of evidence‐ and information‐based policy making at the government level. 

Page 218: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit
Page 219: THEY WILL GET THERE!prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit

 

193 

Top Institute for Evidence Based Education Research (TIER) 

The  Top  Institute  for  Evidence  Based  Education  Research  (TIER)  is  an  inter‐university 

institute that conducts research to develop evidence based education. The institute has 

three partners: the University of Amsterdam, Maastricht University and the University of 

Groningen and is located in Amsterdam, Maastricht and Groningen. 

 

TIER contributes to the  improvement of the quality of education  in the Netherlands by 

promoting an evidence based approach as a guiding principle  in education policy and 

practice. It accomplishes this by developing (cost) effective education interventions that 

are  grounded  in  sound  scientific  research.  TIER  research  is  funded by  the Ministry of 

Education,  Culture  and  Science  and  the  participating  universities  through  NWO  (The 

Netherlands  Organisation  for  Scientific  Research)  and  complies  with  the  quality 

standards and evaluation procedures used by NWO.  

 

The following books recently appeared in the TIER Research Series: 

 

I. C. Haelermans  (2012), On  the productivity and  efficiency of  education.  The 

role of innovations in Dutch secondary education 

II. L. van Welie (2013), They Will Get There! Studies on Educational Performance 

of Immigrant Youth in the Netherlands