From footwear to strategic funding – the curious world of subject-level data

Footwear production is a strategically important subject.

As such, providers (in this case the University of the Arts London, De Montfort University, Leicester College) receive around £243 per student from OfS, per year, for provision returned under this HECoS code.

Until Gavin Williamson’s guidance letter, it was funded at the same level as archeology. Both sat in price group C1, along with most of media studies (promotional media courses were excluded) and creative arts. The latter subjects have been shifted to group C3, receiving a 50 per cent cut in income next year and the promise of further reductions in future years.

Footwear production remains in group C1, along with computing, geographical and information systems, clothing production and – it appears – cinematics and photography. This is based on a comparison of the list of subjects cited in group C1 in the 2020 HESES guidance, and the list in Williamson’s letter.

DfE does have a bit of history in getting confused within the fascinating world of subject specification. It’s a complicated area, but then so is the idea of an academic subject. But both have a hugely significant impact on the funding of a course – these questions might seem theoretical, but they have a huge practical and financial impact.

What is a subject?

How can you tell what a course (or indeed a module) is about? If you are a human being looking to study on said course there will be a human readable description available on the provider website or in a prospectus. The DMU BA (Hons) Footwear Design, for example, “explores the fascinating and complex shoe design process, enabling you to acquire the knowledge and skills that will set you apart from others in the fashion sector”. As such – if you were so minded or if it was your job to do so – you would suggest that it fell within the boundaries of the HECoS code “100110 – footwear production”, which itself falls under the Common Aggregation Hierarchy (CAH) level 3 grouping of “production and manufacturing engineering” (CAH10-01-13) (a part of the big chunk of CAH10 “engineering and technology” subjects).

But a peer under the bonnet shows individual modules on “introduction to fashion/footwear business”, “illustration and visual communication”, and “manufacturing technology”. Much of the course seems to fit well under another HECOS code – “100054 – fashion” – which is a “broad subject which covers every aspect of the fashion process, from initial design concept to the production of a catwalk collection”. A “scope note” to this coding continues “may include pattern-cutting and garment production, CAD, illustration, concept development and communication, design, trend forecasting, and fashion business and marketing”.

There’s also another handy looking code – “100202 – manufacturing engineering” – which covers “The study of the principles of engineering as they apply to the design, installation and maintenance of production-line technologies”. Or how about “100632 – visual communication”, which is “The study of/training in the use of artistic techniques in design to impart information”.

Who decides?

Her Majesties’ Inspectorate of Subject Definitions is not a thing. HECoS codes are allocated to provision by the course provider. And this happens at one of three levels – you can allocate between one and five codes to a course, between one and three to a module, and between zero and five to a qualification. The guidance is that the subjects “must represent directly, or relate very closely to” the qualification’s official certificate or award (for courses and qualifications) or the subject of the module and the module title (for modules – though modules don’t belong to courses).

Data collectors (such as the Office for Students) may occasionally audit data returns where they identify potential problems with the data. If an organisation has been playing fast and loose with definitions or numbers this can be an expensive business – back in 2018 a provider faced a large financial penalty when spending on outreach that was promised didn’t happen. Similar (but less public) funding clawbacks happen when discrepancies are identified in data returns.

If you’re wondering how deep into the weeds people should go on this, the guidance is that good subject coding is “economical” subject coding. The number of codes used should be minimised where possible. This is complicated by the fact that some courses span multiple areas – either by design (a BA (Hons) Business and German) or by the nature of the subject (we could here fairly argue that a BA (Hons) Footwear Design encompasses elements specific to footwear production, manufacturing engineering, fashion, and visual design).

If this is starting to feel disarmingly arbitrary to you, there was a Data Landscape Steering Group that promoted honesty, impartiality, and rigour in data submission and collection. It supported a group of codes of practice, but it wasn’t the HECoS code police and didn’t intervene directly in coding decision making. The group was disbanded in 2020.

Codes and codes

We use subject areas a lot in higher education policy. Subject areas (often at quite broad levels) are used to determine both how students find courses on Discover Uni or UCAS, and how a course is funded. We split data on everything from graduate outcomes to expenditure using subject codes of various types and at various levels. Here’s an incomplete list of where you might find data on footwear production provision:

• Student numbers – the student data in the HESA open data still uses JACS, the predecessor system of HECOS. You can find footwear production at code J445, under the “J400 polymers and textiles” top level subject within the “(J) technology” group. Graduate outcomes also uses JACS. However, if any of the alternate coding strategies I suggested have been used you may also see data aggregated within arts and design (W), engineering (H), or technology (J).
• Financial expenditure – here we use HESA cost centres. Allocation of activity follows the money (for student funding via the module), and as such is based more so on provider structure (in terms of departmental costs incurred in educating students). Depending on the provider in question, we may find “footwear production” data under “120 – mechanical, aero, and production engineering” or “143 – art and design”. Cost centres are specified on module descriptions, which can help in this process.
• Applications – for reasons lost to history, UCAS uses an alphanumeric code of indeterminate provenance and design to signify course content. I found footwear manufacture courses under “WWF7”, “WJ74”, and “W243” – if you follow the theory put forward on the JACS wikipedia page the “W” might signify creative arts and design, the “J” might mean “technology”, and the “F” could indicate a small component of physical sciences. The general suspicion is that the codes are now entirely arbitrary. UCAS reports data on applications using both JACS (historic time series) and HECOS/CAH (going forward).
• Longitudinal Educational Outcomes (LEO) – after an initial decision to use a custom variant of JACS subject groups (separating out stuff like economics and english), LEO data has settled in to using level 2 of CAH. Here “footwear production” and “manufacturing engineering” turn up in engineering (CAH10-01), “fashion” in creative arts and design (CAH25-01).
• Research quality – researchers can be returned via any unit of assessment. For REF 2021 it seems reasonable to look for footwear production research in unit 32 (“art and design” or unit 12 (“engineering”). But again this will depend on decisions made at provider level.
• Teaching quality and course content – the QAA subject benchmark statements, which are a fantastic guide to what you might expect to find on a course in a given subject area, obey a logic all of their own.

What about OfS high cost subject funding?

The former HEFCE price groups currently used by OfS to develop high cost subject funding allocations do not map cleanly to any method described above. Annex G of the HESES2020 guidance specifies the subjects in each group in three ways:

• By general definition – a set of specific definitions highlights sandwich year students, medicine and dentistry, veterinary science, pre-registration nursing and allied health, social work, and education. These set out a range of definitions based on qualification aim, clinical or non-clinical nature of provision, staff background, and even the average cost of delivery per full-time equivalent (derived from TRAC(T) returns).
• By CAH and HECOS – using a range of hierarchy levels, and even individual HECOS codes (which is how we know footwear manufacture is in group C1)
• By LDCS codes – a frankly horrifying looking alternative to CAH/HECOS used in further education.

What’s always fascinated me here is that this guidance treats subject areas as absolute rather than as being based on a provider-level professional decision made in coding. You need to return full-time equivalent students to the right price group by “year of study”, which is a concept not used in HECOS code allocation elsewhere. Here’s an example of how the guidance suggests you do that:

A provider offers a foundation degree in Climate Science and Climate Change, over two years. The course has been allocated two HECoS codes by the provider which correspond to different price groups:

• • 100379 (climate science) – CAH26-01-06 – price group B
• • 101070 (climate change) – CAH26-01-02 – price group C2.

In the first year of the course, students mainly study climate science. In the second year of the course, there is a larger focus on climate change. Over the entire duration of this course, student activity relating to climate science makes up 40 per cent of the course, with climate change making up the remaining 60 per cent. Students in both years of this course would be reported with 40 per cent of their FTE in price group B and 60 per cent in price group C2.

To do this, you might have to go down to module codes or how they map to cost centres (remember you can use up to three HECOS codes per module) and select the most appropriate ones – and then think about the proportion of student activity over the whole course based on what modules they would be expected to study.

Wait… what?

Do we know what modules every student is going to do over a whole course? No.

Can we accurately discern a difference between “100379 – climate science” (“The study of secular weather conditions and climates”) and “101070 – climate change” (“The study of the modelling and analysis of current and past climate variability and causes of change on all timescales. Includes the understanding and modelling of future climate change, the detection and attribution of past change and the prediction of impacts of climate change on the environment.”)? Probably not.

But if we use one code, we get £1,500 per FTE student (group B) – if we use the other we get £0 (group C2). If we use a mixture we get a number between the two. Or, going back to our footwear design course – it could be either in C1 (“footwear production”), C3 (“fashion”), or B (“manufacturing engineering”). So what might happen?

Provider data people are an ethical and upstanding bunch; they’re some of my very favourite people in higher education. But any provider, especially in these straitened times, is going to want to maximise income. Expect those module titles to get another tweak ahead of next academic year.

A cause celebre

Only about 40 students are currently studying “footwear production” according to HESA. There’s four courses at three providers in the whole country. I’m genuinely delighted that Gavin Williamson sees it as a strategic priority subject.

But the Secretary of State isn’t as big a fan of media studies. Hansard records the “lairy uncle drunk at wedding” energy with which he reported that he was “slashing the taxpayer subsidy for such subjects as media studies” (I can’t even bring myself to type that without adding that PwC has calculated the entertainment and media industry could be worth up to £76bn to the UK economy by 2021).

By “media studies” in CAH terms this is the whole of CAH24 moving from C1 to C3, as follows

• (CAH24-01-01) information services
• (CAH24-01-03) publishing
• (CAH24-01-04) journalism
• (CAH24-01-05) media studies
• (not (CAH24-01-02) publicity studies as it was already in group D for some reason)

This includes subjects as diverse as librarianship, broadcast journalism, film studies, digital media and online publishing.

Here’s who is teaching students in these areas and how many students are involved. Note that I’ve used JACS codes P100, P300, P400, and P500 as an approximate match, and that these numbers are headcount not FTE. That said, if you wanted to multiply a provider’s number for FT first degree students by £121, you’d get a decent idea of the amount of OfS funding that might be lost per provider just to make Gavin Williamson look good in the bar of the Codsall Conservative Club:

[Full screen]

But here – and within the Secretary of State’s list – we do not include “internet technologies”, “business information systems” or “multimedia computing science” (still in group C1 under computing), or “cinematography” or “photography” (both within CAH25-01-04 – cinematics and photography, the only creative arts and design topic that was not mentioned in Williamson’s guidance and thus seemingly still in group C1). It’s all a question of emphasis – a matter which course teams will be considering with some interest.

Data categorisation can be seen as a violent act, in the sense of destroying difference in pursuit of convenience – but this is more often a case of categories being a convenience based on consent. The violence comes when these conveniences are used as a means to target the politically unfashionable. An act like this destroys the integrity of the data we have; it collapses the fragile structure of custom and commonplace, and aligns data creation against data use. The subject descriptions we are familiar with are not strong or simple enough to use as a weapon in a culture war. They will bend or collapse in the interests of peace.

Meanwhile, students already on these courses may find that, from next academic year there’s less money to cover the cost of running and maintaining the kit and facilities they signed up to use. Or fewer academics and support staff available to teach them. Students will be trampled underfoot. Which will, of course, be the fault of providers.