Exploring the explosion in franchise and partnership activity

There’s a clear need for more regulatory oversight of franchise and partnership teaching arrangements, but – as regulators are finding – there’s no easy way to track which students are being registered, taught, or physically located at which provider.

Knowing where students study feels like a straightforward matter – indeed “where do HE students study” is one of the top level questions posed in HESA’s Student open data collection. If you click on that, it takes you to an up-to-date (2022-23 academic year) summary of student numbers by registering provider.

But as we’ve learned from concerns raised by the Office for Students, the Student Loans Company, the National Audit Office, and (frankly) Wonkhe there is a bit more to it than that. And it is not currently possible to unpick this to show the number of students at each provider – for any given value of “at” other than registering – using public data. But we can do it for the number of courses.

The forgotten open data set

Yes – I’ve started the year abusing the Unistats open data release (it’s the only open data release that lets you find details of courses was the clue). And you sort of, kind of, can unpick some of these relationships using it. After a fashion.

It is worth unpacking our terms a bit:

• A student’s registration provider is the provider that returns that student as a part of their official data returns. If the registration provider has degree awarding powers, this is generally the provider that awards the qualification the student is working towards
• A student’s teaching provider is the provider where a student is actually taught – in the instances where this is different to the registering provider this usually happens via a partnership arrangement of some sort.
• A student’s location provider is the actual place a student is taught – usually, this is the same as a teaching provider, but not always – for example a “university centre” based at an FE college counts as the university in question doing the teaching, but the location would be the FE provider that hosts the centre.
• We’ve also got to deal with the idea of an awarding provider, the place that actually awards the degree the student is working towards. In the main this is the same as the registering provider, but where the registering provider can’t award the degree in question it will be someone else by arrangement.

How does this appear in Unistats? You are probably familiar with the notion of the UK Provider Reference Number (UKPRN): a unique identifier for educational settings. In the unistats data we get something called PUBUKPRN, which identifies where a course is primarily taught. We get something called UKPRN, which identifies where students on a given course are registered. And we get LOCUKPRN, which identifies the location a student is taught at – where this a location with its own UKPRN that is not the same as the PUBUKPRN.

Limits to sector growth visualisation

What’s missing – we don’t get a UKPRN for the awarding body. Not in public data. It is collected (OTHERINST) and used on the Discover Uni website but it is not published for me to mess with. Not yet, anyway.

So what I can show you is the number of courses at each combination of registering, teaching, and location provider. This shows instances where students may be registered at one provider but taught at another (your classic partnership arrangement), and the evolving practice of declaring unilateral branch campuses (where students are registered and taught at one provider, but located at a different one).

That latter one explains the explosion of London “campuses” that area really an independent provider (that may cater to multiple institutions in a similar way). The whole thing gives some indication of where a given provider is involved in franchise/partnership activity – but only where this is shown on unistats.

What I found

First up – sorted by registering provider. You can filter by registering provider if you don’t want the whole list (whyever not?), and I’ve included a wildcard filter for course titles. This is instead of a subject filter – each course in unistats is meant to have a subject associated with it, but this tends not to happen for the kinds of courses we are interested in here:

[Full screen]

And the same thing, sorted (and searchable) by teaching provider:

[Full screen]

Many limitations, but a space to keep asking

The limitations here are huge. What we really want is the number of students registered on each course – we can get this at various levels of aggregation but not reliably at a single course, single cohort level. HESA’s policy of rounding and aggregating to avoid identifying individual students (and of course the historic nature of much of the data presented on unistats) means that most of the information in the data set (entry qualification), NSS, graduate destinations is combined across multiple years and numerous related subjects.

In the usual unistats fashion “courses” are a unique combination of qualification aim, title, and mode for each location. So the handful of remaining providers that do defined “joint” degrees (two or more subjects) look like they are spectacularly busy. And, as always, the overall quality of data isn’t brilliant so there will be stuff that doesn’t look right (pro tip: yes tell me, but also tell HESA).

Here then, is a partial explanation as to why regulators and others have been slow to respond to the growth in franchise, partnership, and other joint provisions: almost by definition the novel things providers are up to don’t show up in data collection. That’s kind of the point.

But is a simple list of available courses, where they are taught, what (and whose) awards they lead to, what subjects they cover, and how many students are on each too much to ask? It appears so.