The system rationing entry to England’s more effective schools could have a big impact on social mobility. Our research provides the first detailed national dataset since widespread academisation.
By Simon Burgess (University of Bristol and IZA), Estelle Cantillon (Université Libre de Bruxelles, FNRS, and CEPR), Mariagrazia Cavallo (University of Bristol), Ellen Greaves (European University Institute and University of Bristol), and Min Zhang (University of Westminster)
Families can express a preference (‘choose’) for their child to attend any state school in England, but schools set the rules for who gets in if they are over-subscribed. On 1 March each year, there is joy or heartbreak as parents find out whether their preferred school also ‘prefers’ them.
In our newly-published paper, we study the rules that schools set, finding that there is huge variation in a complex system, and that parents in some areas have very little information to navigate it. We find that despite explicit financial incentives to admit disadvantaged pupils, very few schools decide to give them priority in admissions.
In contrast, most schools choose to give priority to pupils according to where they live, which can help to sustain an outcome in which students from the poorest families are assigned to the least effective schools. Overall, the picture is not positive for social mobility, but the few schools adopting innovative approaches offer hope.
Schools differ in their effectiveness. To illustrate, comparing regular state schools within a 5-mile radius of University of Bristol School of Economics, there is a wide range of school effectiveness (using the Progress 8 measure from the last pre-covid year), from +0.53 (‘well above average’) to -0.69 (‘well below average”). That gap is over one grade per subject in each of 10 subjects, so schools are clearly very different.
That effectiveness in turn helps shape students’ qualifications and in turn their future. This means that which students can access the most effective schools, and which end up in the least effective, matters for the distribution of life chances.
In England, school allocation is based on the preferences of families, and the over-subscription criteria of schools. These criteria include, for example, whether there is a sibling already at the school, or the family lives within the catchment area, or perhaps a religious requirement. They are therefore important in influencing educational inequality, social mobility and the transmission of privilege.
Intuitively, admissions arrangements matter more for schools with high performance, which are more likely to be over-subscribed, and where the design of criteria might have the most influence on social mobility.
We study the diverse admissions arrangements across all secondary schools in England, as published in 2020. This is the first national study since work for 2012, since when the increase in academisation has meant most schools now determine their own admissions rules.
Historically, local authorities have been responsible for setting admissions arrangements (uniformly) for most schools in their area; for example, in 2004 only around a third of schools had the power to set their own admissions arrangements – now that figure is 90%.
This almost total decentralisation naturally produces a great diversity of over-subscription criteria and a highly complex system. To a degree, this honours the different missions that schools may follow, but it also creates a complex puzzle for parents trying to make important decisions. All the rich variety of admissions arrangements is detailed and illustrated in the paper. Here we summarise the main points.
How common different admissions criteria are in secondary schools in England:
- Legally required priorities – 3245 (99.91%)
- Sibling – 3103 (95.54%)
- Geographical – 2862 (88.12%)
- Social and medical need – 1571 (48.37%)
- Child of staff – 1420 (43.72%)
- Feeder schools – 1245 (38.33%)
- Religious criteria – 491 (15.12%)
- Test – 351 (10.81%)
- Pupil premium – 170 (5.23%)
- Fair banding – 103 (3.17%)
- Child of armed forces – 69 (2.12%)
- Random allocation – 33 (1.02%)
Source: Authors’ dataset of secondary school admissions arrangements (3,244 schools). Secondary school admissions arrangements collected from local authority and school websites for entry to the 2020-2021 school year.
Note: An observation is a school – criterion pair. Each school is counted more than once if the school has more than one criterion. Admissions criteria are grouped to aggregate criteria types. Percentages will add to more than 100. ‘Special need’ groups special circumstances, medical need, and international students (one school). ‘Quota’ and ‘fair banding’ are coded as the presence of these features across any school admissions criteria.
It is obvious from this data that all schools use several priorities, including the legally required ones of students with an education, health and care plan (EHCP) and students in the care of the local authority (looked after children, LAC).
Beyond those, some schools have only one more criterion, and others have another seven, with fine gradations between the criteria. Most schools choose to have a distance tie-breaking rule to break ties between pupils in the same priority group. Our paper provides deeper analysis of this data.
By far, the most common types of criteria are having siblings already in the school, and geographical criteria. By geographical criteria, we mean criteria that specify certain locations that are privileged in terms of entry to the school; this includes catchment areas and simple distance or travel time.
Other widely discussed criteria such as religious requirements (15% of schools) and entry tests for selective schools (10% of schools) are by comparison far less common.
The near-ubiquitous sibling criterion is non-substantive: this criterion simply intensifies the importance of getting the first child into the desired school on the basis of some other criteria. It does not inherently favour any group, and just acts as a multiplier for the impact of other criteria.
Almost 90% of schools use oversubscription criteria based on a measure of a family’s location relative to the school. Furthermore, this criterion is near the top of schools’ ordering of criteria, meaning that it will often be decisive in determining who gets in.
The use of a distance-based tie-breaker is also extremely common. This means that the scope for indirect selection via house prices is high: where you live matters enormously for which school you get into.
Of course, there are also benefits from having a school’s intake based on location. This might help to create a strong community feeling, in which children who learn together also play together. But whilst being clear about that, we should not forget the impact of location-based criteria on access to high-performing schools.
There is a trade-off, and society needs to determine the best point on that trade-off. In our view, that best point is very unlikely to be all the way at one end of the trade-off, as it is at present.
In contrast to the potential negative effect of geographical admissions priorities, the government included in its 2014 School Admissions Code a channel for schools to explicitly prioritise students from disadvantaged backgrounds: those students eligible for the pupil premium (based on the receipt of welfare benefits).
A very striking finding from our data is that this is almost completely absent: used by only 5% of schools, and, for reasons discussed in the paper, is really only meaningful in a few dozen schools (out of over 3,250).
Given the often-progressive ethos of many schools, this is fascinating and puzzling. Neither the additional funding allocated to schools for each eligible student nor schools’ possible social goals of improving diversity appears to be sufficient for schools to explicitly prioritise the admission of pupils eligible for the pupil premium.
We find that a few schools are showcasing more innovative admissions arrangements. These include random allocation of some places to some applicants without reference to distance (104 schools), a test-based entry to assure a mixed-ability intake (`Banding’, 103 schools), and a meaningful use of the pupil premium criterion (around 40 schools).
Geographical criteria can also be modified to be more inclusive, for example reserving places for pupils out of the catchment (23 schools) or across catchment areas (35 schools). Even straight-line distance tie-breaking rules can be modified: for example by spreading the distance measurement points across the city.
These schools are showing that other approaches to admissions are possible and may offer exemplars for other schools to follow. We find that free schools are notably more likely to use innovative admissions.
There is an intriguing possibility that this flows from specific guidance from the Department for Education to new free schools on how to set admissions criteria that gives some prominence to the use of the pupil premium criterion, although it might simply reflect the ‘fresh start’ these schools have.
Our study highlights and illustrates the complex environment facing parents. In some areas of England, parents do not have good enough information about schools’ admissions arrangements.
For example, 33% of local authorities do not publish school admissions guides that contain the admissions arrangements for all schools in their area, 63% do not provide full information about schools’ over-subscription in the previous academic year, and almost 10% of local authorities contain schools where catchment area information is not possible to find without contacting individual schools.
Information can be incomplete, unclear or incorrect – it seems like a relatively easy policy win to correct these shortcomings.
This paper is the first part of a research programme funded by the Nuffield Foundation. We are very interested in measuring the causal impact of these admissions arrangements on the chances of children from disadvantaged backgrounds attending highly effective schools.
Our future work will focus on this, and specifically on the implications for the test score gap between poor and more affluent students, and on school segregation.