In a new academic paper published today, we study the extent that socio-economically disadvantaged children with top Key Stage 2 mathematics scores progress in this subject through school and into university. This is part of our Nuffield Foundation funded study into the outcomes of high-achieving disadvantaged children, with previous papers written on this topic available on this site.
Figure 1 illustrates how, of 31,310 disadvantaged pupils with top Key Stage 2 mathematics scores across two school cohorts, 19,160 (61%) achieved at least a GCSE B/5. By age 17 only 6,745 were recorded as studying mathematics, and just 5,000 took A‑level maths. From those A‑level takers, 2,770 obtained at least a B grade; around 1,105 went on to start a mathematics‑focused undergraduate degree; and roughly 715 achieved a 2:1 or first.
The mathematics pipeline is hence somewhat leaky – tens of thousands of disadvantaged children show clear mathematical ability at age eleven, yet only a tiny fraction go on to achieve at least a 2:1 in a mathematics-focused degree.
Two stages stand out as critical loss points. First, maintaining performance through to GCSE matters enormously: roughly 12,000 of the initial 31,000 high‑ability disadvantaged pupils do not reach a B/5 at GCSE. Second, subject choice at 16 is also decisive — about 14,000 more opt out of advanced mathematics after GCSE. A smaller, but still important, drop occurs when pupils choose university subjects. Taken together, these three losses explain why just 2% of disadvantaged pupils with high Key Stage 2 mathematics scores go on to obtain at least a 2:1 in a mathematics-based degree (physical sciences, mathematical sciences, computer sciences, engineering).
The distribution of losses from this pipeline is not uniform. Figure 2 presents differences in the probability of achieving a GCSE B/5 grade between high-achieving disadvantaged pupils of different ethnicities and genders. Estimates refer to percentage point differences compared with disadvantaged White boys.
The lost potential of disadvantaged White pupils is particularly acute. High‑achieving Asian pupils are, for example, around twenty percentage points more likely than equally high‑achieving White pupils to secure at least a B/5 grade at GCSE. Initially high-achieving disadvantaged Black pupils also outperform their White peers.
Yet, even when disadvantaged White pupils do continue to achieve a strong grade in GCSE mathematics, they are less likely to continue to choose to study this subject post-16. This is illustrated in Figure 3, which presents differences in the probability of taking A-Level mathematics amongst initially high-achieving disadvantaged pupils that achieved the same GCSE mathematics grades. Results are again presented compared to disadvantaged White boys as the reference group. Even when they are equipped to do so, White boys are around 16 percentage points less likely to choose to study A-Level mathematics than their Black and Asian peers.
Together, these results illustrate how many children show clear potential in mathematics at age eleven but fail to develop that early promise. If we want more home‑grown mathematicians, we need interventions that keep bright disadvantaged pupils engaged in mathematics at the moments that matter most. As we have highlighted previously, Key Stage 3 may be a key period when intervention is needed.
This includes during the GCSE to A-Level transition – when many young people who excel at this subject choose not to study it at an advanced level. Schools and colleges should support high‑ability disadvantaged pupils who excel at mathematics with targeted academic support, clearer subject‑choice guidance and stronger pastoral care that sustains their interest, motivation and aspiration in the subject.
Funders: This project has been funded by the Nuffield Foundation, but the views expressed are those of the authors and not necessarily the Foundation. Visit nuffieldfoundation.org.
Want to stay up-to-date with the latest research from FFT Education Datalab? Sign up to Datalab’s mailing list to get notifications about new blogposts, or to receive the team’s half-termly newsletter.
Hi John, this is very interesting, thank you for publishing. Do you have any analysis comparing the Fig 1 “leaky pipeline” of disadvantaged pupils vs non-disadvantaged pupils? is it much less leaky or about the same?
Maths as a subject is not particularly valued in England. Whilst adults might be embarrassed to admit that they cannot read very well, they will openly tell you that they are not very good at Maths and did not enjoy learning Maths at school. Children are influenced by their parents. If their parents do not value Maths skills, then the children are also less likely to. Parents who do not value Maths skills are less likely to support their children by buying them Maths revision books or sending them for extra Maths tutoring. If parents do not have strong Maths skills, they also will not be able to support their children with their Maths homework at secondary school. So whilst these children may have done well at Maths during primary School, they are likely to fall behind during secondary school, losing confidence and therefore interest in the subject. Pupils at secondary school are also very influenced by their peers. Their social life is very important to them. If their friends are not doing so well in Maths and get moved to a lower set, then it is likely they will not mind so much if they end up joining them there. Pupils may be afraid to be seen to be working hard in Maths, due to peer pressure. Secondary schools trying to engage pupils in Maths face more of a challenge when the parents do not support them or value Maths as a subject.
Nothing to do with teaching then?
2%. What percentage of disadvantaged pupils in the top quartile for KS2 maths get any kind of good degree?
At primary school maths is made up of arithmetic (procedural) skills and reasoning skills which are assessed separately. At the end of KS2, given the structure of the SATs, pupils can meet and even exceed the ‘expected standard’ based on good-to-excellent arithmetic skills but average reasoning skills, i.e. they do not need to be great across the board. I research primary school testing, and I have family who teach maths in secondary school, so I have both quantitative and anecdotal evidence here. This means a proportion of the ‘promising mathematicians’ you discuss might not have the solid foundation in mathematical reasoning that’s required for a really secure entry into KS3, despite good procedural skills, so as their secondary school experience progresses they find it more difficult to move forward.