A national student record system?

As part of an AHUA essay competition that appears to have attracted entries from a fair few Wonkhe authors the UK’s premier exponent of registrarism as lifestyle choice – True Crime on Campus author Paul Greatrix – has argued for what we might describe as an anti-competitive big state solution to the problem of student data.

Because the team at HEPI typically are keen on things like that, Paul has been offered two blog posts to expound on the benefits of such a scheme. The argument as I understand it goes:

• Every higher education provider needs a student records system, and currently every one procures their own – even though only a handful of vendors offer such a system.
• Implementing big IT projects in a university or college is hard and annoying (as most provider data teams know only too well from the extended run up to Data Futures).
• A common system means it is easier for regulators and government to access up-to-date information, and easier for students to move from provider to provider – and even into and out of the sector.
• So centrally procuring one bare-bones SRS (student record system – though with his love of administration Paul prefers SAMS, a student information management system) would save money and time, and give us all better data to play with.

As the man himself says:

What’s not to like?

Well…

So Ucisa – an interest group made up of the kind of people who like to think about university systems a scary amount – collects data on what systems people use. There is a segment leader for student records, but it is also notable that 44 per cent of survey respondents (for 2021-22) do not use it. A browse through this fascinating report shows that there are different patterns (and different market leaders) for other systems – such as timetabling, estates, customer record management, virtual learning environments, IT service management, and survey tools.

I mention this because all of these systems need to interface directly with your student record system. As much as I’d like to think that universities are sensible enough to avoid proprietary standards and vendor lock-in, the chances are that there are several potential combinations of software that simply do not work. Each higher education provider has its own blend of systems (for reasons both practical and historic), and has invested time and effort in making them all play nicely together. Swapping out – or even updating – one system has a huge knock-on effect, requiring further procurement (“oh, of course, you need our new system to do that”), substantial in-house development and testing, and (sorry) the kind of back-office duct tape and string “data factories” that are often what holds university systems together.

So – our first question needs to be: are we just swapping out everyone’s SRS, or the whole kit and caboodle? And, either way, who is going to pay for all the work involved in taking an existing system that may on occasion approach adequacy, throwing it in the bin, and implementing something completely new in a reasonable timeframe so that timetables can be populated and VLE access granted in time for some arbitrary deadline.

The changes people have need to make just to work with Data Futures requirements (of which more anon) have been long and painful – not least because one of the main users of that data can’t make it’s mind up what those requirements are, or admit that other use cases for the same (or similar – unless we’re standardising across the whole tertiary sector?) data exist. Even assuming that we can have a stable data specification (again, Data Futures suggests not), is now really the time to spend scare university resources on systems architecture? We’re talking millions of pounds for a larger provider, and although Martha Horler at Futureworks could probably implement something like this in a spare 20 minutes between nostalgic rounds of Quake Deathmatch not every smaller higher education provider has that kind of expertise.

Level playing field

No IT development starts from a blank piece of paper. And it would be remiss of me not to mention that we already have a national student record system that allows us to track student progress and understand trends in performance – it is called HESA Student. It sounds trite, but it collects the majority of the data Greatrix is asking for already (though it doesn’t publish all of it) If you think that is a bit vanilla, and you want to much about with more granular student performance data, Jisc has a system that you can buy into if you chose to (and chose to comply with data protection policies).

Most of the student and regulatory-facing benefits that Paul suggests could be achieved by doing something I’ve been arguing for for a while – implementing the 2002 MIAP recommendation and mandating the use of the Unique Learner Number (ULN) for all post-16 learning. We’re pretty good as a nation at tracking learners through various government systems based on existing HESA records and data-matching – a single, universal, identifier, offers that magic to higher education providers too, providing the key to many of the problems that the LLE will face (and the next government will have to deal with after the election).

There are already an array of student identifiers (SLC customer numbers, provider-specific codes, HESA student numbers, UCAS identifiers, candidate numbers) and we have been unable in two decades to agree to use a single already-existing standard that would benefit literally everyone. The fact that the assembled brains of the government, regulator, and higher education sector have stumbled here suggests that agreement while specifying, procuring, testing, and implementing an actual standardised piece of software (or even agreeing a few standards for systems interoperability) may be beyond them.