REF2021: much still to be decided

Image: Shutterstock

The rules for the next Research Excellence Framework were supposed to be done and dusted a long time ago. Instead, we are looking at yet another consultation exercise covering institutional eligibility, staff submission, and the hotly contested question of output portability. These are questions that HEFCE and the other UK funding councils have been very publically wrestling with since the publication of the Stern Review of Research Funding.

What the REF primarily does, lest we forget, is assist the funding councils in allocating funding (QR) for research. For decades, the stated funding policy has been one of concentration – the very best research should attract the most funding, which leads inexorably to a smaller number of places where research is funded. And the “best” research is that which impresses the appropriate REF panel – research outputs (not just articles!) submitted by institutions as their “best” within the subject area (Unit of Assessment) in question. The panels read and analyse all of the research submitted to them – some also use citation metrics to inform their decision, which is fascinatingly controversial as arguments rage as to the degree a high number of citations indicate the “best” research. Panels will also examine the research environment to get a feel for the broader research culture within an institution.

We are in the middle of the REF2021 cycle right now, so what has been announced today (following the previous consultation) will be a relief to the sector. Continuity has been maintained in most cases, and there are few surprises in the “initial decisions”. The inclusion of an Open Research section in the environment template, with space to provide details of open access publication and data sharing, it a particular highlight. But the areas where decision making are still ongoing are key – I summarise each below.

Any portability in a storm

In earlier competitions, research outputs were fully “portable” – when a researcher took on a new job in a new institution everything they had written and publish travelled with them, and could be submitted to REF by their new home institution. From this, a thousand metaphors bloomed – allegations flew that some institutions were appointing “star” researchers on (very well paid) fractional contracts to add power to their REF submission.

The Stern review attempted to put an end to this aspect of game playing, recommending that outputs should be non-portable – remaining at the institution the researcher was at when they were created for the purposes of REF. Lord Stern argued that “Disincentivising short-term and narrowly-motivated movement across the sector, whilst still incentivising long-term investment in people will benefit UK research and should also encourage greater collaboration across the system.”

There are, frankly, good arguments on both sides – but the main issue is not the basis of the decision itself but the timing of it. We are midway through the REF2021 cycle (2013-2020) – any decision made now would unpick institutional planning and preparation in progress. So the issue David Sweeney dealt with on the HEFCE blog earlier this summer was very much one of implementation. And, like the debate on Brexit, the answer appears to be a transitional period.

During said period (the 2021 cycle), the proposal is that research outputs will be counted twice – once at the originating institution, and once at the institution that currently employs the academic in question. Though this avoids the complexity of a mid-cycle cut off (the other proposal on the table) it does this by sacrificing precision.

REFability for all!

The other major change Stern recommended was to widen the pool of researchers at a given institution that should have their research submitted. For REF 2021, the proposal was that all “research active” staff in the institution should be submitted. As you would expect, this created confusion as to how to identify “research active” staff: contractual status is not reliable, institutional declarations would re-introduce the “tactical submission” element, and issues ranging from independent researchers to interdisciplinarity cause further headaches. David Sweeney – my candidate for hardest working wonk of Summer 2017 – has also been grappling with this, setting out his thinking in July on Wonkhe. The consultation itself suggested the use of HESA staff data, which is provided annually by institutions. The latest proposals are in Annex A.

Submission to REF is a big deal to academics – whether or not work is “REF-able” can have serious impact on a career. It builds into the often-unhelpful mythos that has been constructed around REF – that certain types of output, topics of research, even (ridiculously) chosen journals for publication are favoured by the panel. In practice, just about any research is potentially REF-able – datasets, monographs, articles in the press – and any academic is REF-able. The new proposals make this clear, and effectively remove the institutional selection stage that so often dominates faculty politics.

The REF consultation also proposed the decoupling of researchers from outputs – although one output per FTE was expected, this did not mean one output from each submitted researcher.

Entering REF2021

But which institutions can enter REF? In an expanding sector, the proposal is that previously eligible institutions – plus those who have since gained university title, whether or not they receive funding – may enter REF2021. Eligibility in future rounds will be consulted on as a part of the eagerly awaited consultation on the wider regulatory framework. To quote the circular:

Among providers in England, we welcome submissions to the 2021 exercise from previously eligible, HEFCE-fundable institutions, including connected institutions. We also welcome submissions from those with university title, whether or not eligible for funding. We expect all providers submitting to the REF to have research degree provision. This includes institutions with validation, sub-contractual or franchise arrangements with other higher education institutions. Institutions wishing to participate in the exercise will also be required to meet the submission requirements that will be set out in the detailed guidance on submissions.

If all this feels like a bit much, remember that REF’s rules have changed over time and the current proposals represent an iteration rather than a wholesale transformation. Institutions will, as they have always done, try to play the game within whatever constraints provided by the rule-setters. Ultimately, with so much prestige (and cash) riding on the exercise, playing the game as well as possible is essential – but for the game to work the rules need to be clear.

3 thoughts on “REF2021: much still to be decided”

  1. It’s too soon to say that the proposals “effectively remove the institutional selection stage” because in the next paragraph you say that “The REF consultation also proposed the decoupling of researchers from outputs” so it’s no longer a question of “were you included?” but instead potentially on of “how many of your outputs were included?”.

  2. David Kernohan says:

    Hi Simon – thanks for your comment! Was writing at speed so possibly could be clearer – what I was trying to get at is that the idea of an individual research-active academic being “REF-able” (or not) has gone, and that this idea has been proposed to relieve the career pressure on a “non-eligible” academic.

    Of course, institutional politics are such that another way to apply this demarcation has been quickly identified… Sadly, I’ve also heard tell of people’s contract designations being changed.

  3. Dan says:

    The staff submission problem is not just about the negative effect of whether somebody who is not research active will be classified as such. It is also about the opposite: some institutions will now want to submit non-research active staff as research active to improve their submission.

Leave a Reply

Comments 3 See all

  1. Dan View
  2. David Kernohan View
  3. simonjohnthompson View