Discussion piece by the CDBU Steering Group
Results from the research excellence framework (REF) were publicly announced on 18th December, followed by a spate of triumphalist messages from University PR departments. Deeper analysis followed, both in the pages of the Times Higher Education, and in the media and on blogs.
CDBU has from the outset expressed concern about the REF, much of it consistent with the criticism that has been expressed elsewhere. In particular, we note:
Inefficiency: As Derek Sayer has noted, the REF has absorbed a great deal of time and money that might have been spent better elsewhere. The precise cost has yet to be reported, but it is likely to be greater than the £60m official figure, and that is not taking into account the cost in terms of the time of academic staff. Universities have taken on new staff to do the laborious work of compiling data and writing impact statements, but this has diverted funds from front-line academia and increased administrative bloat.
Questionable validity: Derek Sayer has cogently argued the case that the peer review element of REF is open to bias from subjective, idiosyncratic and inexpert opinions. It is also unaccountable in the sense that ratings made of individual outputs are destroyed. One can see why this is done: otherwise HEFCE could be inundated with requests for information and appeals. But if the raw data is not available, then this does not inspire confidence in the process, especially when there are widespread accusations of games-playing and grade inflation.
Concentration of funding in a few institutions: We are told that the goal is to award quality-related funding, but as currently implemented, this leads inevitably to a process whereby the rich get richer and the poor get poorer, with the bulk of funds concentrated in a few institutions. We suspect that the intention of including ‘impact’ in the REF was to reduce the disparity between the Golden Triangle (Oxford, Cambridge and London) and other institutions which might be doing excellent applied work, but if anything the opposite has happened. We do not yet know what the funding formula will be, but if it is, as widely predicted, heavily biased in favour of 4* research, we could move to a situation where only the large institutions will survive to be research-active. There has been no discussion of whether such an outcome is desirable.
Shifting the balance of funding across disciplines: A recent article in the Times Higher Education noted another issue: the tendency for those in the Sciences to obtain higher scores on the REF than those in the Humanities. Quotes from HEFCE officials in the article offered no reassurance to those who were concerned this could mean a cut in funding for humanities. Such a move, if accompanied by changes to student funding to advantage those in STEM subjects, could dramatically reduce the strength of Humanities in the UK.
Unaccountable flexibility in the funding formula: There are many different ways of achieving ratings: For instance, whether or not the ratings include ‘intensity’ (number of returnable staff who were entered), can dramatically alter rank orderings. Or we could look at a suggestion by Graeme Wise that a ‘bang for your buck’ metric that assessed outputs in relation to grant income would be most appropriate. Even more radical is a suggestion by Dermot Lynott, that we should be giving the most rewards to those whose outputs were impressive in relation to their scores on environment. Needless to say, a very different profile of winners and losers emerged from such an analysis. It will ultimately be a political decision as to how to translate the REF scores into funding. We have to ask whether it worth going through this entire long-winded exercise if, by simply changing the funding formula, one can make a dramatic difference to an institution’s funding to achieve a politically expedient outcome.
Damage inflicted on careers and morale: The criteria for entering staff for the REF could appear quite cavalier; for instance, the requirement for a numerical ratio between number of staff entered and number of case studies meant that some departments with few case studies were unable to enter all plausible staff. Derek Sayer has described instances of decisions to enter staff being made on what appeared to be flimsy evidence based on ad hoc internal evaluations. Yet being identified as ‘non-REFable’ is not only damaging to morale, but could have real impacts on prospects for promotion and job security.
Focus on competition rather than collaboration: The REF exercise creates rank orderings, and everyone is desperately trying to nudge ahead of the others. In fact, there are so many different ways of doing the ranking, that almost everyone can be satisfied that they are ‘among the top’ on some index or other. Those who crowed loudest about their success tried to temper this by arguing that they were celebrating a broader ‘British’ success, but this seems perverse. Why should concentrating ever more of the excellent research in an ever small number of institutions be regarded as a national success story? It is, of course, widely believed that competition is a force for good, stimulating people to do better than they otherwise might. However, many in academia take the view that they don’t need to be incentivised by competition to work hard: they are in the job for the love of it, and would like their efforts to be appreciated for what they are, not because they help push the institution up a league table. Competition can also damage relationships between different departments within a University, if there are disparities in REF performance that lead to bickering about who is more valuable.
Perverse incentives that damage research: These may play out differently in different subject areas, but overall, many academics feel that they are not able to do the research they want in the way they would like. In science, there are intense pressures to publish in high-impact journals and bring in grant income. Some institutions are notorious for threatening redundancy to scientific staff who do not meet some agreed quota of research income, creating incentives to do ever more expensive research (see for instance cases at Kings College London, Imperial College, and Warwick University Medical School). In humanities, the pressure to produce a steady stream of research articles and monographs has led to the sense of an enforced move to over-specialisation, with academics increasingly incapable of explaining or demonstrating the broader significance of their work. Younger generations of academics find that their direction of research is being wholly driven by ‘REF-ability’, and that journal publications automatically trump those in volumes of collected essays, even when the latter may be more important for the field.
Perverse incentives on hiring practices: This is another consequence of the intensely competitive culture that is induced by the REF. We have, particularly around the time of the REF, a market in research ‘super-stars’, who can attract impressive transfer fees. People from other institutions who are employed at only 20 per cent part-time suddenly appear on the books, boosting the institution’s return on funding and outputs.
Devaluation of non-research activity: Academics whose positions require them to teach and do research have felt pressured to focus principally on research, and teaching has consequently been devalued. It is sometimes suggested that the Impact agenda of the REF also encourages academics to spend time on public engagement, but in fact it has the opposite effect. Public engagement does not count as ‘impact’ for REF purposes: to demonstrate REF impact, one must provide concrete documentation of how a specific piece of research has influenced non-academic users, such as policy-makers, health professionals, museums, etc.
How have we come to this?
Given that many of these points were made in the run-up to the announcement of REF results, we have to ask how it is that we find ourselves trapped in such an undesirable system. It is noteworthy that the REF is popular with many vice-chancellors and administrative teams. It makes it easier to manage staff, with objective criteria for hiring and firing, and provides league tables to measure progress by. For those already attached to the vision of a university as big business, it seems the natural next step to have objective rules for defining winners and losers so that one can directly measure an individual’s likely ability to bring money into the university without having to make difficult judgements about the intrinsic quality of their work.
It is a moot point whether the collapse of the circle of winners to Oxbridge and London was the result of deliberate planning, or an unintended consequence of how the system operates. Be this as it may, one concern is that this could provide further pressure for the British university system to reconfigure itself so that it can compete with the American private elite. There would be increasing reluctance to use general taxation to concentrate even more educational resources within close proximity to London, and instead we could see a shift to private funding, increasing the extent to which access to the best institutions is distributed according to wealth rather than ability. In a few short years we could see the destruction of a genuinely national system of higher education, publicly funded because it is designed to serve everyone, to a privately funded system that is world-class for the few who can afford to access it, but a disaster for the country as a whole. As in the US, educational opportunity would be concentrated overwhelmingly in places where it can be accessed primarily by the wealthy, privileged and well-connected.
At the time of the announcement of REF results, there was a sense that anyone who criticised the celebrations was either a bad loser, or – if they came from an institution who did well – a traitor for not celebrating British success. At CDBU we are proud of UK Universities and their research reputation, but our loyalty is to our discipline, our profession, our vocation and our sense of their place in the wider scheme of things: we fear that the assessment process embodied in REF will in the longer term damage these.
It is, of course, all very well to criticise and paint visions of a dystopian future. If we wish to replace the current system, we must look at alternative ways forward. At a debate about the REF, organised by Sage Publishers on 8th December, David Willetts MP, who until July 2014 was Minister of State for Universities and Science, took the view that some of those who disapproved of REF were just dinosaurs who wanted to go back to the 1970s, when funds were allocated to institutions by a group of the great and the good making judgements over dinner at the Athenaeum. We would dispute that this is the only alternative to the current system. But Willetts was right on another count: he pointed out that the current REF system was not imposed by government, nor by HEFCE. Indeed, they had been actively pursuing the idea of using a simpler metrics-based system for the REF, but it was resoundingly rejected by the academic community. Government, according to Willetts, would listen to any reasonable proposal for a new system. Clearly, it is now up to the academics themselves to propose a viable alternative.
At the same meeting, David Sweeney, Director of HEFCE responsible for Education and Knowledge Exchange, implied that critics of REF just wanted to be handed public money without any accountability. He emphasised that the government and taxpayer put money into university research and had a right to know about the outcomes from the investment they had made. We totally agree. But we take issue with those who, like Mark Leach, Director and Editor-in-Chief of Wonke, think that the REF, for all its limitations, provides a good solution. Our position is that, for all the reasons given above, the REF is a seriously flawed system for deciding on disbursement of research funds, which in the long run will do the UK University system more harm than good.
What alternatives are there?
1. One possibility that has been discussed is to remove the QR component of funding altogether, and give all funds to the research councils. The problem with this solution is that it would mean the research councils would have to grow in size enormously, the load on reviewers, already seen as unsustainable by many would increase yet further, and pressures on academics to bring in research grants would be even more intense. It has also been argued that it would further increase disparities between the Golden Triangle (Oxbridge and London) and the rest, and would disfavour non-STEM disciplines.
2. HEFCE is has been looking at various publication-based metrics that might substitute for the REF, and plans to do some empirical studies comparing metric-based evaluation with REF results. However, metrics have been vigorously opposed by many in the academic community, especially in the humanities, where there is much worse agreement with expert opinion than in the sciences. We do at least now have hard data from the REF that can be considered when evaluating how metrics would perform, but there’s a real risk that introduction of any metric will further distort incentives, so that the measure becomes the goal.
3. At the Sage meeting, Derek Sayer put forward another interesting alternative, which was that funding should be based just on the ‘research environment’ component of the REF, which focuses more on inputs than outputs.
4. An even simpler option that would retain the dual support system but remove quality-related funding would involve disbursing funds purely on the basis of the number of active researchers in a department. This could be criticised for leading to a ‘prairie farming’ model whereby departments would band together to create enormous conglomerates that would benefit from economies of scale. One could, however, put a limit on the size of unit entered. At the Sage meeting, David Sweeney expressed himself as strongly opposed to this solution, even though in the last round it gave a funding result that was highly correlated with actual funding outcomes from RAE, in both science and humanities. He is right in noting that, despite the high correlation, there would nevertheless undoubtedly be winners and losers who would suffer substantial gains and losses in real terms, relative to the RAE result, but the question is whether this would involve unfairness. It’s hard to say, given that we have no gold standard. You could of course further argue that you have to have some measure of quality to incentivise people to do better, as well as a need to guard against freeloaders, like Laurie Taylor’s Dr Piercemuller. Yet, as we have noted, for most academics, exhortations from managers to ‘do better’ don’t achieve much and may indeed be counterproductive. We need to be accountable for the public money spent on university research, but subjecting every apple in the barrel to an exhaustive x-ray examination may not be the best way to identify the rotten ones.
We do not have a single solution, but we think that academics must take control of this process and not leave it in the hands of HEFCE and the government. This article about Dutch Universities Netherlands has strong parallels with the UK situation: the author argues academics have been too passive in accepting an emphasis on competition, and the use of evaluation systems as means of control. There is unlikely to be an ideal solution and we may have to live with the ‘least bad’ option. But let us consider all options in terms of how far they are likely to exacerbate or resolve the problems outlined above, or we may find ourselves saddled with something even worse than REF2014.
We hope this article opens up the discussion on this topic. Please do add your comments. We are moderating comments to exclude spam, but non-anonymous, on-topic comments will be published unless they contravene the usual rules.
Finally, if you agree with the broad concerns expressed here, please do consider joining the CDBU to help us campaign more effectively for change.