Skip to main content

31 May 2025

The National Curriculum for Computing: a personal view

Simon Peyton Jones profile image
Written by

Simon Peyton Jones

Purpose

Prompted by the National Curriculum Review, this post (originally written November 2024 as input to that review) summarises my personal views about the structure of our national curriculum and (particularly) assessment, insofar as they relate to computing.

One principle of the Review is that it's about evolution rather than revolution.  That's a sensible principle.  But here I have tried not to worry too much about what is "evolution" and what is "revolution", and instead just focus on what are the key challenges we have, and what solutions might we be able to adopt.  An evolutionary approach is right in the short term, because there are immediate problems to solve, but we must not risk compromising our longer-term vision: even the smallest steps need an overall sense of direction.

Crucially, education is about moral purpose.  Alas, education is also complicated, so it is easy for thinking about education to get bogged down, muddied, diluted, mired in that complexity.  A review of the national curriculum should strive to articulate that moral purpose, and locate its detailed recommendations within a compelling, inspiring vision.

1. Computing is a foundational subject, not simply a vocational one

The 2014 National Curriculum fundamentally re-envisioned computing as a foundational subject that every child should learn, just as they do maths, natural sciences, history and geography from primary school onwards.  This contrasts sharply with the earlier view of ICT as a purely instrumental/vocational subject, which young people learn to give them employable skills (I am caricaturing, but only a bit).  

Here are the Aims of the new national curriculum, quoted from the Programmes of Study for computing [my bold]:

The national curriculum for computing aims to ensure that all pupils:

  • can understand and apply the fundamental principles and concepts of computer science, including abstraction, logic, algorithms and data representation
  • can analyse problems in computational terms, and have repeated practical experience of writing computer programs in order to solve such problems
  • can evaluate and apply information technology, including new or unfamiliar technologies, analytically to solve problems
  • are responsible, competent, confident and creative users of information and communication technology

These ambitious aims are as appropriate now as they were a decade ago.  We teach foundational subjects to every single child, not because they are all going to become physicists, doctors, mathematicians, or software engineers, but because we want to equip them with the powerful knowledge that will allow them to make critical and well-informed judgements about the complex world that surrounds them.  If anything, this powerful knowledge is more important in computing than in natural science, because the digital world presses on us so ubiquitously.

We must not lose this North Star vision in the blizzard of a national curriculum review. 

2. It's all about qualifications

Schools, teachers, and students exist in an incentive structure that governs their behaviour. The nation provides the incentive structure; we cannot blame schools for responding to it.  If we want different behaviour, we need to change the incentive structure.

2.1 Qualifications drive behaviour

Our current incentive structure is massively dominated by qualifications. Those qualifications

  • Drive school behaviour
  • Drive teaching pedagogy
  • Drive student behaviour 
  • Drive parental engagement
  • Drive employer recruitment and university admissions

There are sound reasons for this dominance

  • Qualifications allow schools to be measured -- the "accountability system".  This has driven massive improvement in school quality over the last three decades.
  • Qualifications offer a convenient and somewhat-objective proxy for employers and universities to use in selection.
  • Qualifications provide concrete goals for students to work towards, and concrete achievement for them to "bank".

There are plenty of well-known disadvantages too!  (Teaching to the test etc.)

Regardless, because qualifications are so powerful, the national curriculum itself barely matters.  No school is censured for not following the national curriculum.  In theory Ofsted might do so, but in practice, despite Ofsted's more recent focus on curriculum intent, it is extremely reluctant to play a significant role in saying what should be taught.  That said, custom and practice play a huge role.  Ofsted would criticise a school heavily if there was zero PE provision at KS4 for students not taking PE GCSE; but it completely ignores the very same requirement for computing.

Saying "the national curriculum barely matters" is not the same as saying "it doesn't matter what children learn".  On the contrary, we must start from what we want our young people to learn (e.g. see Section 1)!   But having decided what we want them to learn, we must embody that curriculum in the structure and content of our qualifications; doing so is far, far more effective than merely writing it down in a "National Curriculum" that everyone (not just academies) proceeds to ignore.

Qualifications do not reign entirely supreme.  Schools do, as part of their taught curriculum time, offer PE, RE, and citizenship to all students in all years.   It is interesting to reflect on what incentives make schools do so.  Perhaps we could lean into that?

2.2 Qualification are a lever for change

There are two possible responses to the current dominance of qualifications:

  1. Somehow make qualifications less influential, and the national curriculum more so.
  2. Accept that qualifications dominate, but be much more thoughtful about the behaviours they drive.  There is much that could be improved.

In this response I assume that (2) is the primary way forward, so my comments are focused almost entirely around qualifications and incentives.  Get those right and the rest will follow.

Tantalisingly, qualifications constitute a powerful lever that is in the hands of government. Education is so complicated, and many things are not under government control.   (e.g. You cannot legislate to get more girls to study computer science; it's a socio-technical-cultural problem.)   In contrast, national qualifications are fully under government control.  

Qualifications may be a lever, but it is a lever that is rife with unintended consequences.  We need to be careful about pulling it.  But leaving it as it is today (with its existing unintended consequences) is itself a strategic choice with major consequences.

3. Breadth, agency, and choice

An unintended consequence of our current system is that powerful forces act to narrow the range of subjects that a student studies, too early.  Specifically:

  • In the olden days, students often took four, or even five A levels in Year 12, and dropped one or two of them by Y13, exiting with an AS level.  The loss of AS levels led to earlier specialisation; it is now rare for students to take more than three A levels.
  • The National Curriculum requires that all students study computing up to KS4.  In practice, this requirement is simply ignored (including for schools that are obliged to follow the National Curriculum) because of the pressure to achieve good results at KS4.  Studying digital literacy at KS4 does not contribute to a Progress 8 score, so schools simply don't do it.
  • At primary, the enormous pressure on schools to deliver good SATs results leads them to focus their efforts on English and maths, especially numeracy.  This squeezes out subjects that are not examined in SATs: languages, computing, art, and so on.
  • The English Baccalaureate provides a strong disincentive for schools and students to take a wide range of subjects, including, dance, drama, art, computer science, physical education, religious education, photography, textiles, and more.  See Section 6 on the English Baccalaureate.

4. GCSEs and vocational qualifications

4.1 Two silos

We have somehow come to take for granted that there are two "silos": vocational qualifications (BTecs, T Levels etc) and academic qualifications (GCSEs, A levels).  These silos are deeply embedded in our organisational structure: awarding organisations have complete divisions responsible for each; the Department for Education has entirely different teams for each.   We have literally institutionalised the academic/vocational divide.  No wonder vocational qualifications do not enjoy parity of esteem!

The differences are more than organisational: the rules are different for each.  Notably:

  • The DfE reviews each year which vocational qualifications count for Progress 8.  So an awarding organisation can lose their P8-counting status without warning; and schools have much-reduced incentive to invest in offering qualification if it may (to all intents and purposes) be withdrawn at short notice.  Nothing like that happens for GCSEs.
  • The DfE publishes a Subject Specification for each GCSE, and a variety of awarding organisations then offer qualifications that meet that specification.  The specifications are long lasting -- so much so that they entirely lack a meaningful review and update process -- which allows awarding organisations and schools alike to invest in them.  Not so for vocational qualifications -- see Section 4.2.

 As a result, vocational qualifications continue to be seen as second class.

Yet, seen from 100,000 feet, there is a continuous spectrum of students, ranging from ones who are highly academic and interested in subject disciplines in their own right, to ones who are more interested in the applications of knowledge, and learn best through doing.    The obvious response is to see all KS4 qualifications as lying on a continuum, with one set of rules applying to all of them.

4.2 GCSE subject specifications

Our current GCSE arrangements are these:

  • The DfE publishes a subject specification for each GCSE.  These specifications are extremely long lived and hard to change.   Here, for example, is the subject specification for computer science GCSE, which has been in place unchanged for over a decade, despite well-understood shortcomings.
  • Awarding organisations publish qualifications that meet the subject specification.  Ofqual checks that they do so.

For vocational qualifications and for International GCSEs (IGCSE), the system is different. The DfE publishes guidance covering guided learning hours, assessment methods, grading, progression, overlap with GCSEs etc; but does not specify content in detail.   The awarding organisations are free to offer qualifications that meet these guidelines.  Applying this same approach to national GCSEs would make the system far more agile and flexible.

For example, currently there is a DfE-specified GCSE in computer science; but computer science is only part of the subject described in the National Curriculum programmes of study for computing.  Weirdly, computing is the only subject whose GCSE offer does not cover its own National Curriculum.  With a more liberal regime, awarding organisations would be able to offer, say, 

  • a GCSE in Computing covering the whole National Curriculum subject
  • a GCSE in Applied Computing that focuses on computing through the lens of its applications

Awarding organisations, schools, and teachers would decide what a good balance is, rather than DfE prescribing it with a decade-long time constant.

England is unusual in having multiple awarding organisations: many countries have only one, and one that is closely allied with the department for education.  But since we have a multiple-AO system, let us exploit that opportunity by encouraging AOs to come up with exciting and innovative qualifications, rather than micro-managing them.

Would this mean that the government is abdicating responsibility for a very important area where the risk of market failure is high?  No: the government would still play the role of regulator -- it would just be stepping back from micro-management with a decade-long pace of change.

Care would need to be taken

  • To avoid a confusing plethora of qualifications.  For example, there are no fewer than 86 active digital technical awards (including Level 3).
  • To ensure the quality of approved qualifications. In suggesting that we treat GCSEs and vocational qualifications as a continuum, I am not advocating lowering expectations, returning to the bad old days of "ICT is dead easy and is worth 4 GCSEs".   There must be a quality threshold, with clear criteria, and transparent assessment.

But these are problems we would like to have, and they seem eminently soluble.

4.3 Key Stage 4 qualifications for computing do not serve our students needs

Our current KS4 qualifications offer in computing does not meet our young people's needs.   It is helpful to recognise three cohorts of students

  1. The software developers of the future.  These students are like me: they enjoy creating new software by programming, they love the intellectual richness of computer science.  They make up only 15% of the cohort, in which girls are very under-represented.
  2. The professionals of the future.  An engineer makes deep use of computing, but for the purpose of engineering.  An artist may make deep use of generative AI, but for the purposes of art.  Similarly game designers, architects, doctors, data scientists, and so on. This "deep use" means they need a foundational understanding of how computation and data work; but their motivations are driven by applications rather than by excitement about computing itself.  I guess this is around 30% of the cohort.
  3. The citizens of the future.  Every single person must be digitally literate, to survive and thrive in a world pervaded by digital technology.  That includes students who aren't really interested in computing at all, and want to do GCSEs in art, dance, Latin, PE, DT, and religious studies.  And fair enough!  I'm guessing that this is 50% of the cohort.

The National Curriculum embraces all three groups (Section 1).  Given the pivotal importance of qualifications, argued above, we absolutely must provide meaningful qualification pathways for all three; but sadly we do not:

  • Group (A) is well served by GCSE Computer Science.
  • Many in Group (B) would enjoy a KS4 qualification in computing, but they are served only vocational qualifications, which do not enjoy parity of esteem with GCSEs.  Moreover they were designed under a requirement to have no overlap with GCSE, and which have therefore been eviscerated of technical content.  So they do not serve group (B) well at all.
  • Group (C) is not really served at all.  Most schools offer no computing whatsoever to students not taking a GCSE CS or vocational qualification; that is, for a majority of students computing education ceases entirely at the end of KS3, regardless of the fact it remains a statutory obligation at KS4.

This matters.  KS4 has a huge back-wash into KS3: in practice computing stops for many at the end of Year 8, and even when it does not, it is often relegated to one 50-minute lesson a fortnight.  That just isn't enough to provide an engaging education.  We should not blame the schools for this; they are simply following the incentive structure that our qualifications and accountability measures provide.

In my view, the biggest single obstacle to computing education in England is the unbalanced landscape of computing qualifications at Key Stage 4.  It's not the only challenge of course -- teacher recruitment and retention comes next -- but it is tragic that the biggest obstacle is one of national policy.

What can be done?  It's not that hard:

  • We must ensure that there are meaningful and motivating qualification pathways for children in groups (A), (B), and (C).
  • We need GCSE pathways that suit students in groups (A) and (B). 
  • We need some kind of non-GCSE digital-literacy qualification for group (C); see Section 5. 

There has been quite a bit of constructive debate in CAS about GCSE pathways for groups (A) and (B)., which I do not want to repeat here:

  • One possibility would be a sister GCSE in Applied Computing.  
  • Another might be to replace GCSE Computer Science with a broader GCSE in Computing, covering the whole of the national curriculum; Kemp and Berry make the case here
  • Other constructive contributions have been made by Greg King (here and here) and Dave Gwilt (here).
  • Another more radical possiblity (sketched in 4.2 above) might be to cease prescribing precisely which GCSEs should exist and harness the creativity of the awarding organisations.

5. Literacy in maths, English, and computing at KS4

It is outright immoral that we force every child to take GCSE Maths; that 30% fail it (one in three); that they are forced to retake it; and that failure rates for these resits are still higher. It is hard to imagine a more effective way to turn students off maths.

We have to acknowledge that the (rightly demanding) Maths GCSE simply does not meet the needs of a significant cohort of our young people.  Yet they must leave school with a level of mathematical literacy.   We know that mathematical literacy has an important influence on their life chances (point to evidence); but we do not know that maths GCSE is the most effective way to exert that influence, and the malign effects of repeated failure for a third or our young people are intolerable.

The situation in computing is different.  Only 15% of the cohort take GCSE computer science, and another 5% or so take vocational qualifications.  That leaves a huge majority who simply stop studying computing at the end of KS3.  And yet again we know that students must leave school with a level of digital literacy.

Conclusion: we should complement high-stakes, demanding GCSEs with a suite of literacy qualifications.

  • Lower stakes
  • Pass/fail not graded
  • Levels: bronze/silver/gold
  • Not taken as part of the end-of-KS4 blizzard.  Could be taken when ready: earlier (e.g. in Year 9), or later.
  • Not just skills -- must be underpinned with knowledge.  Must not be a qualification in the ability to take screen-shots.
  • Emphatically not a replacement for GCSE(s) in computing.
  • Explicit employer buy-in essential. Part of the incentive to make the qualification attractive would be explicit endorsement by employers.
  • Big focus: inclusion, every child

Other key points:

  • Any such literacy suite must be seen as an integrated whole, not tackled piecemeal as point-solutions for maths, for computing, or for English.
  • A literacy suite complements GCSEs; it does not replace them.   In particular, digital literacy qualification is absolutely not a replacement for a GCSE track in computing, any more than a maths literacy qualification would replace a maths GCSE.   

Open questions:

  • How can schools and students be incentivised to adopt these qualifications?
  • Could these quals be framed as "half-GCSEs", about which there is some prior art?
  • How would this literacy suite differ from the current Functional Skills qualifications?  We should not reinvent the wheel!   But Functional Skills today is very much seen as “a second-best for children who are failures at GCSE”.  We need to reposition literacy as something that every single child needs, not just a back-stop for failures.

Side note. The current Functional Skills qualifications in English and Maths are available at Level 1, 2, and 3.   But weirdly for IT, Level 2 is not allowed; only Level 1 and 3 are allowed.  This seems strange and counter-productive.  End of side note.

References: OCR report, Maths Futures report, Raspberry Pi/Manchester MBacc applied computing qualification, Adrian Mee's "Five-strand model"

6. Review, and relax or abandon the English Baccalaureate 

The English Baccalaureate (EBacc) has commendable intentions, but in practice it has turned out to be (in my view) over-prescriptive .

Students for whom the EBacc is in reach are driven to study

  • Maths (1 GCSE, counts double for Progress 8)
  • English (2 GCSEs)
  • Science (3 GCSEs)
  • A foreign language (1)
  • History or geography (1)

Again driven by Progress 8 accountability measures, most schools let their students do 9 GCSEs, which leaves a knife-fight for a single timetable slot between

  • Dance, Drama, Art
  • Computer Science
  • Design and Technology, Photography, Textiles etc
  • Physical Education
  • Religious education

Side note about computer science.  The EBacc requires that pupils take either Combined Science (covering the natural sciences) or Triple Science (any three of physics, chemistry, biology, or computer science).   In practice, however, no school offers the possibility of studying two natural sciences plus computer science.  There is a good reason for this: it is a National Curriculum requirement that all students study physics, chemistry, and biology at KS4.  In practice, therefore, the fact that computer science counts as a "science" in EBacc has zero effect.  Two different parts of national policy (the EBacc and the national curriculum) are pulling in different directions.   End of side note.

The balance is wrong here, and the choices feel arbitrary.  For example

  • Why does science take a 3-GCSE "footprint", for students taking triple science?
  • Why is English (two GCSEs) twice as big as maths (one GCSE)?
  • Why are so many subjects fighting for a single slot?
  • Why history or geography; it could be "a humanity" including e.g. RE?

The main point is this: students should have more agency in these decisions.

Side note: these "footprint" observations are further complicated by the weightings of Progress 8.  End of note.

The simplest solution is to abandon the EBacc.  The main reason to retain it would be to prevent mass flight into "easier" GCSEs, but they no longer exist following the reforms to GCSEs.

7.  High-stakes accountability measures

One of the difficulties with qualifications is that they used to serve multiple purposes.  In particular:

  • To certify a student's level of attainment
  • As an "accountability measure" to judge the quality or effectiveness of a school.

Accountability measures are important.  A Department for Education that made no attempt to evaluate which schools are doing a good job, and which a less good job, would itself not be doing a good job.  Moreover, Progress 8 is a far, far better measure than its predecessors (based on absolute attainment), because Progress 8 (attempts to) measure how much value a school has added, accounting for the prior attainment of its incoming cohort.

The stakes are high.   Accountability measures drive school behaviour.  And, ineluctably, any high-stakes accountability measure has unintended consequences.  For example, the very fact that the stakes are so high acts as a powerful driver for reliability of measurement, and that in turn drives a system strongly based on written exams.

  • This undermines trust in teachers.  Instead of trusting their judgement, we only trust exams.  There is a reason for this -- it is human to err -- but the erosion of trust is real and corrosive.  It is the exact opposite of the message we would like to convey: that we trust our teachers' moral purpose, expertise, and professionalism.
  • Practical work and other non-examined assessment became so asphyxiated by safeguards in many GCSEs (including computer science) that it has been abolished in favour of written exams.  Again, the reasoning is sound, but the loss is real, and painful.
  • Reliable exams need highly scaffolded questions. Teachers spend far too long teaching exam technique and "what the examiner expects from this kind of question" rather than teaching the subject itself.   No one wants that; but we have placed teachers in a system in which not to do that would be to their students' disadvantage.  So they grit their teeth and do it.

We cannot abandon accountability measures of some kind.  But I think a clear-eyed and above all honest accounting of their unintended consequences might, just might, lead to ideas to make some of these consequences less painful and pernicious.

8. The KS4 exam load is too high

There is a widespread consensus that the exam load at the end of Key Stage 4 is too high.   See, for example, the Sept 2024 OCR report

I agree, and it's a very important question, but I don't have anything new to add to what others are already saying.