While international research gives increasing attention to geographical factors in education, this perspective is lacking in Dutch research and policy. That a geographical perspective is badly needed, is demonstrated on the basis of the proposed policy to promote new schools in the Netherlands. Current housing policies for Dutch schools are described, pointing at disputes between municipalities and school boards who hold shared responsibilities. Next, foreign housing policies for new schools are studied, that is for friskolor in Sweden, free schools in England and charter schools in Texas (USA). Experiences abroad not only testify that very different choices can be made, but indicate that housing policies may in the long run have a substantial impact on segregation and educational inequality. It is also shown that new schools are mainly located in urban areas. It is argued that in addition to costs for new schools, friction costs for existing schools need to be considered. In all, a geographical perspective on education is needed, so as to prevent increasing segregation and social inequality as well as wasting public financial resources.

In 2015 Amsterdam introduced a new primary school admissions system. This system is rooted in various desegregation pilots, including two based on the ‘controlled choice’ approach. In conjunction with one of these pilots located in West Amsterdam, we researched the criteria parents used while thinking about (de-)segregation in primary schools and parents’ attitudes regarding the controlled choice approach. Due to political developments between 2008 and 2011, we considered our data useless for policy discussions. The introduction of a new admissions system, however, gives our data a newfound relevance. This article therefore describes how parents in socio-economically diverse neighbourhoods think about segregation and school choice. The 249 parents we interviewed or surveyed supported schools with a ‘good mix’, but they had diverging opinions about the meaning of such terms. The parents were optimistic with regard to controlled choice, even if this could in some ways limit their own options. We conclude that the political resistance to desegregation at the primary school level cannot be justified by empirically unfounded claims about the perceptions and preferences of parents. We hope that our findings may lead to empirically grounded policy evaluation and policy making.