For the LLE to work we need to get serious about student identifiers

The proposed lifelong learning entitlement (LLE) is, as I never tire of telling people, potentially transformative. It’s the biggest (and arguably best) idea in post-compulsory education for quite some time.

But I’m acutely aware there’s a lot of detail that needs sorting out. For instance, it seems clear that the policy aim is for all adults in England to be able to access the equivalent of four years of higher education fee loans (as things stand, a cool £37,000) for any form of post-compulsory course.

This prompts a number of questions:

• How do we keep track of how people have used this entitlement, not least how much money they have left to spend?
• How do we manage eligibility for people who have accessed higher education in the past (via any funding system) – how do we count this against the LLE entitlement?
• How do we track eligibility via any rules around equivalent or lower qualifications (ELQs)?
• And how do we understand – locally and nationally – how people are using their LLE and how this links to national skills priorities or proposed course offers?

Detail like this should have been consulted on at the point that the LLE was announced. It should have been included in the background documentation of the Skills and Post-16 Education Bill. It should already be being built into data collection plans (hello Data Futures!). It hasn’t been, and it isn’t.

To address questions like this, and more, we need to dip a toe into how we see people within higher education data. We need, in other words, to raise the whole question of identifiers.

School days

Like much in life, identifiers are simultaneously more simple and more complicated at school. Up to the age of 14, every single person in the English state education system has a unique identification number – the Unique Pupil Number (UPN). One letter and 12 numbers allows schools, local authorities, and central government (in particular the National Pupil Database) to share information about young learners, their progress, attainment, and needs.

If you don’t remember yours from school the reason is that you never knew it – schools are not allowed to tell you. The number is destroyed when pupils leave state funded schooling, at the age of 16 or older. The system is not comprehensive – if your entire school career was outside the state sector (except if you have special educational needs or have been permanently excluded) you won’t have had a UPN.

The National Pupil Database that this identifier enables is astonishing in scope: it is very much your “permanent record” of legend. Attainment at each Key Stage – and even regarding your year 1 phonics screening and early years profile – rub shoulders with exclusion and attendance data, demographics, addresses, and care details. It’s all stuff that would be hugely useful to universities in understanding and supporting learner needs. But universities never get to see it.

Next steps

After the end of compulsory schooling there are a number of other identifiers in use that depend on your chosen destination. The canonical example is the Unique Learner Number (ULN – ten digits). It is not, as the name suggests, a post-compulsory version of the UPN as it is only really used in records to do with attainment and registration. Available from age 14 onwards it covers everything Ofqual-regulated in terms of education, training, and skills provision encountered after the end of compulsory education. It enables the population of your Personal Learning Record (PLR), which you can access by filling in a word document based form and email or post it to DfE.

Before you rush for that envelope, know that the information is only for relatively recent qualifications – covering QCF, A levels, GCSEs, BTECs, Diplomas, and Functional Skills provision since 2012 in England, 2014 in Wales, and 2015 in Northern Ireland. Until 2016, you could access your record via a whizzy website, and even add your own reflections. No more.

University life

But what if you decided that higher education was your next step? The original plan was to also use the ULN in higher education, providing a single identifier for everything post-compulsory. It was sensible, practical, and seemingly without drawbacks. So it didn’t happen.

Before we dive into why, let’s have a quick look at the various identifiers you might come accross in higher education data:

• The venerable HUSID (13 digits, here’s how to make one, and here’s how to make the proposed 17 digit version in Data Futures), is the magic that ties your HESA student records together, and may also have formed a chunk of your student ID card number.
• Some providers use their own identifiers, which are also collected nationally by HESA
• Some providers are happy to supply ULNs, so do so. HESA grab these when they can – and these are compulsory for apprentices.
• Applied via UCAS? You’ve got a UCAS number (two letters, two numbers, or one letter and three numbers – previously nine numbers). Some providers use this internally, and it can be supplied to HESA.
• Applied otherwise in Scotland? You’ve got a Scottish Candidate Number (nine digits), collected by HESA.
• If you had any interactions at all with the Student Loans Company, you have an SLC customer reference number (13 digits). You will also have a Student Support Number (SSN, 13 characters, alphanumeric), which also covers any financial support you might get from other places, but is based on each funding interaction you have with SLC and is not (strictly speaking) a personal identifier. The SNN is collected by HESA for the majority of students.
• And this is just the common ones – there are many, many, more in use by PSRBs, employers, awarding bodies, and overseas agencies.

Standards in public life

At this point, you might be wondering why we need all these identifiers in higher education. A similar question could, in fact, be asked more generally – why do we have:

• A national insurance number (two letters, six numbers, and a letter) when we go to work, or draw benefits?
• A unique taxpayer reference number (ten digits) when we pay tax?
• An NHS number (ten digits) when we access any form of health or social care?
• A passport number (nine digits) when we travel?
• A driving license number (actually 16 alpha-numeric characters) when we use a motor vehicle?
• And so, inevitably on.

Why not just have one big number that we could use to identify ourselves in any interaction with the state? Why not print it on a card? In other words – why not have ID cards?

There’s a section of the population that would never countenance such a thing, for all it would simplify things. Brought up on bank holiday viewings of The Great Escape, they would see anything that would potentially allow an official to ask to see their papers as an imposition against their liberty and freedom. The Prime Minister provides, inevitably, a perfect exemplar:

But I tell you this. If I am ever asked, on the streets of London, or in any other venue, public or private, to produce my ID card as evidence that I am who I say I am, when I have done nothing wrong and when I am simply ambling along and breathing God’s fresh air like any other freeborn Englishman, then I will take that card out of my wallet and physically eat it in the presence of whatever emanation of the state has demanded that I produce it.

More sensibly, there is a concern that we may choose to or be required to provide information to one body that we do not wish to provide to another. You will be happy that the NHS keeps a full record of your health interventions, but you may feel that you don’t want the DVLA to know about the time you fell off your bike and broke your arm. You may be happy that the Passport Office knows you have left the country for two weeks, but not that the Department of Work and Pensions does. And similar issues do apply in higher education.

Managing Information Across Partners (MIAP)

There’s a corollary to that, of course. How do you avoid having to give the same information to multiple sources, multiple times? Why – in other words – spend hours filling in forms at every conceivable agency when you could just do it once?

In 2002, the government established the MIAP programme to sort out the tangled world of personal data in education. Everyone in post-compulsory education was to be allocated a Unique Learner Number (ULN) – which would allow them to access their own records via the web-based service (the Learning Records System) that eventually became the form you had to send into DfE.

The aspiration was that learners could also add their own records and reflections on participation and achievement – that could be updated, shared with employers, and used in applications for further training. This record would also act as a spur to developing personalised and joined up information, advice, and guidance. It also addressed some of the glaring gaps that lead to the carnage that was the Individual Learning Account (ILA) in the early 00s – we urgently needed to get better at identifying providers in the post-compulsory space, and the learners that study at them.

Some of the underpinnings of this work are still used to this day – a lot of very sensible work on common data definitions mean that things are slightly less chaotic than they could be. For me, the highlight was the establishment of the UK Register of Learning Providers and most importantly the UK Provider Reference Number – one of the reasons working with provider level data is so much fun and so straightforward.

But the ULN itself, and the Learning Records System? It was launched in 2008 and still exists to this day – though the learner-focused web platform has long gone. If you have ever taken a qualification post-16 that was on the old Qualifications and Credit Framework or the new Regulated Qualifications Framework, you’ll have one. If you are over 30, went to an independent school and then straight to university, were educated overseas or in Scotland – you won’t. Either way, your university education most likely won’t be on it.

Big brother wasn’t at university

Given the blizzard of criticism at the launch of the ULN – this is no mean feat. It made (in a first for education data personal identifier specifications) the front page of the Telegraph.

The plans for Unique Learner Numbers have sparked fears of moves towards a Big Brother state and raised suspicions that ministers are trying to introduce a national identity card system by stealth.

In the House of Commons, Damian Green was equally unimpressed:

Will the Secretary of State confirm that the charmingly named “Unique Learner Number” on page 66 is, in fact, an identity card, and can he tell the House what level of compulsion he plans for the use of that card?

The ULN was mandated for the school and FE sector – something that apparently did not bring about an Orwellian surveillance state or spark the development of Skynet. But it was never mandated in higher education – something that you may ascribe to provider autonomy until you recall the large number of data conventions that are enforced by regulation and by administrative convenience.

Instead, the blame can be placed on the way ULNs were mandated for use elsewhere. Remember we noted that you wouldn’t know your UPN because you were never told it at school? ULNs have worked, in practice, in the same way – meaning that the flow of ULNs into higher education (it’s fair to assume most young people entering HE will hold a Level 3 qualification) has depended on the passage of information between schools and colleges on the one hand and universities on the other. Which hasn’t always been great. This is why the ULN hasn’t been used widely in HE – as the coverage hasn’t been great, and other identifiers have the benefit of wider and more established use. It’s just a shame nobody else uses any of the common HE ones.

The issue came back up during the HEDIIP project that led to the development of Data Futures. In 2015 a published blueprint called for HESA and UCAS to be able to verify ULNs, for SLC, HESA, and UCAS to routinely and by default capture and share ULNs, for HE providers to be able to mint new ones where needed, and for HESA to provide HE qualification data to the PSR. Again, it isn’t Orwellian, it isn’t hugely complex – but it wasn’t done.

Back to the LLE

We need, clearly, to have a single identifier to support people in using their lifelong loan entitlement within the rules of the scheme. To me, the ULN is the obvious candidate – it is already mandated in skills, adult learning, and the FE sector. And there is no immediate barrier to using it in higher education. It would be the culmination of nearly two decades of data architecture development to support a more joined-up post-compulsory sector. It doesn’t meet all the use cases that the various identifiers already in use address – but no single identifier ever could.

Using administrative data matching (in the same way as something like LEO connects student records with taxpayer data using names, dates of birth, and addresses) we could match current and recent students back to their ULN in most cases, and allocate new ones where this is not possible. But a universal entitlement – making the LLE available as promised to fifty-five year old retraining diesel engineers and such like – requires going back in time long before the start of the ULN. And this is where we get into serious trouble.

Legacy data

There’s a service called Prospects HEDD (run by Jisc) that allows employers to check to see whether job candidates have the qualifications they claim to. For £12 you could (with my consent) check up to see if I really got my degree or have just been bluffing for 20 years. But for people a bit older than me this relies on records held at providers (even the university statistical record only goes back to 1972 and does not include former polytechnics). Some providers, but not all, have digitised their historic student records and allow Prospects HEDD to sell the ability to search them. But quite a lot of student records are still only in hard copy.

The various changes of form and ownership among colleges and polytechnics are not the best way to ensure comparable data is available. If you are designing a scheme that allows people to claim an entitlement that needs to take into account previous government-funded access to qualifications – and this would be the only fair way to address this – this presents a problem. Boris Johnson had three years at Oxford at the taxpayer’s expense – he should have access to a years worth of loans, just like someone who just graduated would.

Going forward, the ULN (and matching HESA records to the ULN) would cover most of this – which is why getting ahead of the data specification game is so urgent for DfE right now. But this does not address the other part of the LLE promise – allowing the current workforce to access training via loans.

The nuclear option

I’m never a fan of inventing new identifiers, and I sigh inwardly when people invent their own codes to identify universities in data (in particular I hate it when people think they can just standardise university names, which is maybe a rant for another day). But rather than fix the existing mess of HE identifiers to line them up with the rest of the post-compulsory sector, the temptation to invent a new one must be ever-present in DfE.

For example, you could ignore all the skills planning possibilities of rich data, and all the dreams of fairness in the system, and just standardise on the SLC Customer Reference Number. If you’ve already had a fee loan from SLC, you can’t get the full four years any more – otherwise the four years is there for you (and Boris Johnson) as it is currently. Restricting access to Equivalent Level Qualifications (ELQs) would be silly in terms of upskilling, but it would mean SLC could use the same systems as it does currently to spot people who already have degrees trying to borrow to do another one.

It cuts across the identifiers problem by treating the LLE like a bank account (the kind of account that individuals could use to fund their learning, say…). It’s pragmatic, but we would lose the ability to take a strategic look across learner choices and aspirations to sit alongside parallel insights from employers locally and nationally. All we would know is who spent what, where. Adding a Clearing Plus variant on top of this isn’t skills planning.

If this really is the route, a lot of the benefits that LLE could bring to an England battered by Covid-19 and years of rapid changes to employment patterns and expectations would be lost. It would be more expensive, less comprehensive, and ride roughshod over the ways in which data is currently collected and used in higher education – driving up provider overheads and bureaucracy. It would, in other words, be a waste.

In conclusion

There is never an easy answer in policy – and in policy as far-reaching as the proposed LLE this is doubly true. Anything DfE does come up with, be it mandating the ULN or coming up with something entirely new, will be controversial and difficult.

There are some great data brains in higher education – more than anything I want to see the sector have the ability to feed into the design of this system at an early stage and a fundamental level. We can, in other words, make this work – if we are ever given the chance to.