Earlier this month Times Higher Education carried an article about the REF, suggesting that institutions should be required to submit all relevant staff to the REF, rather than selecting a subset - presumably the best researchers - in order to get a high score when the rankings are published in 2014. The online article included some readers' comments, some agreeing, others describing the article's view as complete nonsense. So there's evidently an issue here that perhaps needs a bit more elaboration, which is what I shall try to do now.
The article I'm referring to above was actually published on January 5th, but this has been such a busy month - with a visit to Kuwait (see previous post), one to Galway (see next post), and lots of work on a multi-partner research funding application to the EU - that I'm only rather belatedly finding time to comment. Life as a retired academic is proving quite strange these days, and sometimes I do wonder how I found time, not so long ago, to do a 'proper' job! Luckily, most of what I'm doing is extremely interesting and/or fun.
Now, back to the REF.
Current guidelines do allow institutions, in any given subject area (unit of assessment, or UoA), to submit whichever academics they wish to the REF. The normal expectation is that these will be academics able to offer four research publications over the assessment period, but some people can be submitted with fewer than this, e.g. new academics, and women whose research has been interrupted by maternity leave, etc.
However, institutions know that a low rating is worth virtually nothing in financial terms, as the funding councils only provide a research stream of funding for subject areas achieving an average rating of two stars or above, and it is likely the threshold will soon rise to three stars. Hence there is a clear incentive to submit only the strongest research performers, to make sure the unit gets some research funding. There is potentially a trade off here, of course, as submitting fewer academics reduces the volume measure used to calculate the research grant; but submitting lots of people and getting a low score provides a big volume measure but a low quality measure, and that could easily end up securing a zero research contribution if average quality falls below the threshold.
Submitting a subset of academics to the REF, therefore, is the way to get a high quality rating and hence secure some research funding from the funding council. It is also likely to score well in reputational terms, as institutions with a high research rating - even if based on a subset of academic staff - can be expected to advertise the fact widely. And it will no doubt help both staff recruitment and student recruitment.
The downside, though, is that the resulting scores for research are supposedly not so great in terms of the accountability of an institution to the funding public. However, personally, I don't find this line of argument at all compelling, not least because we have plenty of channels through which the accountability function is already pretty well served, so I can't imagine that many people will worry over much if research is being measured in an allegedly 'distorting' manner that only takes account of some academic staff rather than all of them.