Sorting matters: choice and selection in English schools
19 May 2006
Parents naturally care about the peer groups their children have at school.
Our research shows that the degree of choice has a powerful impact on school sorting over and above differences in residential sorting. We also find that sorting is much higher in selective LEAs than in the rest of the country.
In principle, we can distinguish three ways of assigning children to secondary schools: elite schooling, neighbourhood schooling and choice-based schooling. In England, elite schooling means selection into grammars or secondary moderns. In non-selective LEAs, the de facto system is a mix of neighbourhood schooling – children go to the nearest school – and choice-based schooling – children go to the school chosen by their parents.
Elite schooling was explicitly designed to sort pupils by their ability at age 11. This necessarily produced a system very highly segregated by pupil test scores and – because of its correlation with test scores – segregated by family background. Part of the drive for comprehensive schools was precisely to end this segregation. There would no longer be a two-tier system, but all pupils would attend their local secondary school.
But many have argued that segregation has been maintained through the operation of housing markets. Neighbourhood schooling means that proximity to a school matters, thus raising house prices around good schools. This in turn means that only the richer parents can live there, and school segregation re-emerges.
The fact that a small number of LEAs have kept selective education means that we can compare sorting outcomes between areas. It turns out that the degree of sorting is much higher in selective LEAs, which we define as those where more than 10% of pupils attend grammar schools.
Analysis of our data shows a dramatic difference between the selective and non-selective LEAs. In the former, there is a very clear difference in scores between the top set of schools, and the bottom schools. In the latter, while there are differences between schools, they are far less marked. So there are important differences in the degree of pupil sorting between areas that have retained selection by ability and areas with comprehensive schooling. Segregation has not simply been re-introduced through house price differences.
The degree of choice has a powerful impact on school sorting over and above differences in residential sorting
Turning to the role of choice, since the Education Reform Act of 1988, English schooling has had some elements of a choice-based system. There are currently proposals to increase the scope for choice. We investigate the relationship between the degree of choice and the degree of sorting. This is a complex issue. On the one hand, increased choice might be expected to reduce pupil sorting. If all pupils in an area have the same chance to go to any school, then on average each school would have a mix of different children.
On the other hand, it might be that an area with more choice might produce a more finely segmented outcome. Parents’ demand for high quality schooling can translate into a demand for good peer groups, and choice allows some to achieve this. But popular schools cannot take all pupils that apply (indeed, some may be put off applying by the low chance of success) and have to assign scarce places. So ‘choice’ becomes choice by both parents and schools, and the overall outcome depends on the admissions process. In this context, choice by one has effects on others – one child going to a particular school reduces the chance of others going.
School sorting might simply reflect neighbourhood sorting, which might arise for a number of different reasons. To examine the potential role of school choice, we compare the sorting of pupils in their neighbourhoods – residential segregation – with the sorting of the same pupils in their schools. We use the same segregation index as before for schools, compute a similar index for residential segregation and then look at the ratio of the two. We interpret higher school segregation relative to neighbourhood segregation as indicative of the ‘re-sorting’ process involved in school choice.
We measure the degree of choice as the number of schools within easy reach. Overlaying school locations with a road network, we construct 10-minute drive-time zones around schools. We simply count the number of schools within this area and use this as our measure of the extent of choice in a local area.
The results show that a greater degree of choice is associated with greater sorting – by both ability and disadvantage – and the effect is strong. Differences in choice lead to substantial differences in each of the three measures of segregation.
This can be understood in two ways. First, more choice means that parents can select from a greater number of schools and hence find one close to their ideal, conditional on where they live. Or second, more choice means that given the desired school, parents have greater scope to live where they choose, not necessarily right next to the school.
This is not just being careful with language. It is clear that school assignment rules have implications for residential segregation as well as school segregation. Greater school choice allows more dispersion of where people choose to live, and thereby potentially more heterogeneous communities, at the potential cost of more segregated school communities.