A brief EPSRC update

Last summer I wrote a post about EPSRC’s plans to direct their funding towards certain areas and not others, and in particular on its effect on mathematicians, the most dramatic of which was to restrict their fellowships, which had previously been available throughout mathematics, to statistics and applied probability. The strongest argument I could see in favour of EPSRC’s position was that they were reviewing the various subareas of mathematics before deciding which should be grown, which maintained and which shrunk, and that so far only statistics and applied probability had been reviewed (with a decision that it should be grown).

It was of course a bizarre decision to remove fellowships entirely from areas that have not yet been reviewed — the obvious thing to do would surely have been to maintain the status quo in an area until the review was complete — but they promised that the reviews would be completed in November, so at least one could hope that this decision would represent no more than a brief hiatus.

The November deadline came and went, and EPSRC eventually explained that because of the protests by many scientists and mathematicians, they wanted more time in order to make sure that they got the decisions right. I think they finally finished in late March, but it’s taken me until now to comment on the results.

For mathematics, there is good news and bad news, with, it seems to me, the latter cancelling out the former. The good news is that the conclusion of the review seems to be that mathematics should be kept roughly as it is: statistics and applied probability are to be grown, mathematical physics reduced (it would be interesting to have an explanation of that decision — when I describe all this as good news I mean merely that it could have been a lot worse), and all other areas maintained. The chart that shows this is here.

The bad news, at least if I understand EPSRC’s website properly (which tends to be quite a challenge), is that these decisions don’t seem to be reflected in the availability of fellowships. If you look at their page about areas in which fellowships are available and scroll down to the mathematical sciences, you’ll find that postdoctoral fellowships are still not available except in statistics and applied probability. For early career fellowships, they have added “intradisciplinary research” (which I imagine will in a lot of cases mean normal research where applicants greatly exaggerate the connections between different areas) and “New connections between mathematical sciences and information communication technologies”. For established career fellowships it’s statistics and applied probability or intradisciplinary research.

So if, for example, you are working in number theory and your work happens not to be intradisciplinary, then there are no EPSRC fellowships available. Given that some excellent research in number theory is purely number theoretic, it’s not quite clear to me in what sense number theory is being “maintained”.

In general, the whole policy of directing research in this way is misguided. It is true that mathematicians particularly value research that cuts across subject boundaries: there are many examples where ideas from one area have been brought into another area and hugely benefited it. But this intradisciplinarity (if that is the right word) is surely a means to an end rather than an end in itself. Well, perhaps it’s a little bit of both, since establishing unexpected connections is certainly one way of advancing mathematics. It would be legitimate, I think, to say that intradisciplinary research is particularly encouraged, but to make it the only form of research that is supported is going much too far. And of course, if you’re unlucky enough to be a postdoc, then you can be as intradisciplinary as you like and you won’t get a fellowship unless you’re doing statistics or applied probability.

So it looks as though the “bizarre decision” I referred to earlier was not that bizarre after all, and that EPSRC more or less knew back then that they were not going to be offering fellowships except in a couple of areas. I don’t know whether their decisions have reached their final form. If an EPSRC representative is reading this and can offer any clarification, then I’d be very grateful: this is an occasion where I’d be particularly delighted to have misunderstood what is going on.

18 Responses to “A brief EPSRC update”

  1. mokshay Says:

    Generally “intra” refers to within, while “inter” refers to in between. So the real meaning of “intradisciplinary” is within a discipline, not straddling disciplines, but I am not sure if this is what EPSRC means by the word.

    • gowers Says:

      My interpretation of the word (as used by EPSRC, that is — I agree with your etymological remarks) is that the research has to cut across different areas of mathematics, but doesn’t have to connect mathematics to other areas of science and engineering. That is, I think “discipline” means (in this case) “mathematics” and not something like “topology”. If that really is what they mean, then a more accurate word might be “intersubdisciplinary”.

      Support for this interpretation can be found in the section Intradisciplinary Research on this EPSRC page. It contains for example the sentence, “We wish to promote connectivity across the mathematical sciences by supporting fellows with novel and potentially transformative research ideas that bring together techniques from different areas of the mathematical sciences.”

  2. Andy J Says:

    You can click on the circles to get a very brief summary of the ESPRC’s intentions for that area. For mathematical physics, it’s

    “However, over the past ten years capacity in Mathematical Physics has grown such that continuing support at current levels risks undermining support for capacity in other important areas. As such, support for this area overall will be reduced to ease the pressure on other areas.”

  3. Alan Weir Says:

    I agree entirely with your sentiments re the emphasis on interdisciplinary (if that is what the EPSRC mean) or ‘cross-disciplinaryh’research. I say this as a non-mathematician with interests in maths. What is depressing is the apparent absence in the funding councils of rational grounds for the heavy emphasis on interdisciplinarity to be found across all disciplines. It’s as if there suddenly was a fad for research conducted on the third floor of buildings, the research councils laying heavy emphasis on 3rd floor research, and we all ended up fighting to be housed on 3rd floors, renaming floors bizarrely, double counting floors by counting some as mezzanine, constructing bizarre new buildings and generally wasting valuable amounts of academic time and money.

    If you were to ask supporters of the cult of interdisciplinarity why this is not a similarly foolish fad they sometimes point out that lots of valuable work is interdisciplinary. But then lots of good research will be done on 3rd floors. The only half sensible defence I have seen is the claim that interdisciplinary research used to be a Cinderella area and found it hard in the past to attract funds, so the current emphasis is a sort of positive discrimination: like striving to direct funds towards female researchers. But is there any evidence to support the view that interdisciplinary research was (and still is?) at a disadvantage, compared to the overwhelming evidence of discrimination against female academics. (Some bureaucrats also seem to want to destroy ‘dead white male’ disciplines and replace them with new ones dreamed up on the back of envelopes over lunch in wine bars in Milton Keynes but I think that is probably not a main cause of the present cult.)

    I should add I have nothing against public funds being used, perhaps even used primarily, for applied research. But it should be directed at particular well-defined tasks: if a goverment agency (or a private company) want a particular problem solved or addressed- find ways of reducing carbon emissions, drugs to cope with new flu viruses etc- then put the job out to tender and let university departments and research groups compete for it. But what is so damaging to academic standards and ultimately academic freedom is mixing pure research and ‘impact’. What is likely to happen is that mediocre research is rewarded and mediocre time-servers rise to the top of disciplines. And a little knowledge of the history of science shows how foolish it is to imagine such people will spot where the really useful parts of pure research are likely to lie. Imagine Bombelli had been beholden to a 16th Century EPSRC to justify any time off from his work as an engineer for frippery like writing algebra books: “fifteen years we have waited to see if these complex numbers with their bizarre imaginary components will bear practical fruit, and nothing at all: do you expect us to wait another 300 years?!)

  4. state of foundational research | Peter Krautzberger Says:

    […] I find Tim Gowers discussing that last year’s scare is now reality — EPSRC has essentially cancelled pure math […]

  5. anon Says:

    Just to highlight the part that resonated most with me: “For early career fellowships, they have added “intradisciplinary research” (which I imagine will in a lot of cases mean normal research where applicants greatly exaggerate the connections between different areas)“. Exactly.

    It is one thing to prioritize work that is of interest, and of use, to the broadest possible segment of the mathematical community, or to prioritize work that fosters connections between large and constantly interacting groups of people, instead of work that takes place in relative isolation. But the consequence of these plans is not likely to be an outpouring of genuinely new interaction. It is likely to be an emphasis on the buzzword “intradisciplinary” in grant proposals, with the goal being simply to get the money so that one can go back to business as usual.

    To the extent that one sees this outcome as negative, it should be noted that it is just as common in statistics and applied probability as it is in other fields, and just as common among people who work in a number of different subfields as it is among people who work only in one area. Academics are academics and funding agencies are funding agencies. Every trend has its day, and everybody does his or her best to try to keep up. It is a pity that the current trend seems to arbitrarily favor some subfields over others.

  6. anon Says:

    Suddenly, for inexplicable reason, subdisciplines of math start breaking up even further. Real analysts suddenly split apart into “Cauchy Sequence Real Analysts” and “Dedekind Cut Real Analysts”. This, of course, accompanies a massive upsurge in intradisciplinary research. Yay!

  7. Disillusioned fledgling academic Says:

    Research is an uncertain process, and everyone knows that if you are investing in uncertainty then you should hedge your bets, and certainly not put all your eggs in one basket.

    “They” at epsrc don’t seem to know that. What if they accidentally identify the “wrong” area for support? What if they identify the “wrong” research leaders?

    I know at least one science Head of Department who has a marketing book on his bookshelf. In an ideal world marketing should not be needed to convince funders of science, they are professional investors in charge of lots and lots of taxpayers money.

    I have an unlikely fantasy that one day I will become a world famous scientist, and utilise the power that it gives me to eradicate the foolishness that governs UK research funding. More realistically I may just quit.

  8. Peter Cameron Says:

    It is true that some of the best mathematics is “intradisciplinary” (I find it difficult to bring myself to write that), in the sense that it connects different parts of mathematics. But I strongly suspect that a relatively high proportion of this happens by serendipity, higher than the proportion which lies firmly within the boundaries of one subject. Would you appoint someone to a research fellowship in serendipity?

  9. Yiftach Says:

    The EPSRC (and most other grant funding bodies) created an extremely inefficient system. The amount of the time and energy invested in grant applications by applicants and referees is huge while the number of projects actually funded is very small. This is insane! Moreover, because universities receive large percentage of the grant there is always pressure to apply to grants and to apply to large grants. In addition, as Tim mentions above, the people controlling the EPSRC are directing the money not by quality but by arbitrary labels.

    As long as we as a community are willing to accept it, nothing will change. The only way to change things is to act together, similarly to the boycott on Elsevier. We need to stop sitting in the EPSRC committee, stop refereeing applications and stop submitting them; and we need to make a lot of noise about it. This is has abetter chance to work than the Elsevier boycott as the EPSRC is a public body.

    • Alex Bartel Says:

      Dear Yiftach, it seems to me that in this case, the situation is much more complex than in the Elsevier case. While in the latter, you can persuade a young postdoc not to submit papers to Elsevier journals (although it is already much harder to persuade a young mathematician to decline refereeing requests from the editors who are senior colleagues that one knows from conferences, from their work, etc.), there is no way you can persuade a young postdoc not to apply for funding for a job from a funding agency, “only” because you don’t agree with the way that money is being allocated. Most young postdocs simply cannot afford to be this picky about what jobs to apply for.

      Similarly, by refusing to referee applications, you often contribute to the application landing on a desk of a lesser expert, and thus, in the short run, to worse funding decisions being made.

      I think it makes sense to assume that EPSRC is acting in good faith, in the sense that they have no personal interest, other than to optimally invest tax money into research (if this assumption is dropped, then it really becomes a completely different story). It seems to me that it would make more sense to try to rationally persuade EPSRC of the misguidedness of their policies based on this assumption, rather than to fight them in the same way we fight a publisher, whose only aim is to channel as much of that tax money into certain private pockets as possible.

    • Yiftach Says:

      Alex, it is a bit hard to state everything you think in a comment so I apologize if I was not clear and in any case my intention is to start a debate rather than to express a final view. Also, one should be careful not to take the Elsevier analogy too far away. I agree that whatever we do we should try and minimize the damage to young mathematicians and I did not mean for young mathematicians to boycott the EPSRC. I also have no problem with fellowships for phd students and postdocs in general, although, I think they should be according to the quality of the candidate and not the area.

      My main objection is to the big grants given to people with permanent jobs. I cannot see the point of them in mathematics. This money should be given to support phd students and postdocs directly rather than through someone’s project (or maybe even better through the departments). The success rate is very low so the time and energy of applicants who didn’t get it is completely wasted. Most of us need very little to do research, just a bit of money for travel and for visitors. The EPSRC insists on funding mathematics based on a model that fits sciences where there are big laboratories.

      Now, I am not sure the EPSRC is acting in good fate, because I am always suspicious of big organization. I can definitely see why it is in their best interest to direct money though a complicated system rather than a simple one. Also, they are controlled by politicians and politicians have their own interests. But even if they do act in good faith, there is no reason to believe they have a good idea about what they are doing. My impression is that they do not seem to listen to the mathematical community, they push agenda based on buzz words and they make the wrong decisions. I also have the impression that the attempts to engage with the EPSRC in a meaningful discussions failed. Now, not being insider I might be wrong, so I’ll be happy to hear more views. I think nothing will change without a lot of pressure. Big organizations do not tend to listen to reason, but they might listen to public pressure.

    • Disillusioned fledgling academic Says:

      As much as I agree with many of the complaints about RCUK, I think that it is important for criticism to be constructive. Its reasonable for RCUK to want to make strategic investments in certain areas, and it is also reasonable to rely on peer review to some extent. The main problem is the fact that noone ever acknowledges that any such decision process will inevitably make mistakes, and that some safety mechanisms need to be in place to prevent positive reinforcement of bad decisions. Repeat investment in the same academics must be analysed carefully, and no areas should be totally ringfenced or cutoff.

      I would like to see an analysis of the statistical variation in referee reports on the same proposal, or of variation in panel member assessments. I would also like to see the percentage of referee reports that are returned in the various categories of “overall assessment” (outsanding, significant etc). I suspect that such statistics would be quite informative with regards to the error in the assessment process.

      Having said that the sooner the “inter”, “intra” disciplinary buzzwords are eliminated the better – I happen to work in an interdisciplinary field, and most colleagues of mine agree that these buzzwords are damaging.

    • Disillusioned fledgling academic Says:

      One more thing – Yiftach I agree with your point about big grants in mathematical sciences to people with permanent jobs. But I think that the problem is much much bigger. The trend towards grant income being used as a metric to assess an academic’s performance is very damaging for areas such as mathematics. A competent mathematician is not as valuable to a university as a comparably (or even less) competent experimental scientist, because the latter can justifiably apply for large grants with equally large overheads. But to change that would require a change in philosophy at a much larger scale than just EPSRC.

    • Yiftach Says:

      Dear Disillusioned, There are two different points: one is in what areas to invest. In theory the EPSRC has the “right” to make such decisions, but they seem not to listen to the feedback from the mathematical community and their decisions make no sense. The other point is even worst, at least in some sense, grants became a lottery there is very small chance of success on the one hand and on the other hand they are important both financially and for the REF so that universities put pressure on academics to apply. This means we invest lots of time and energy in gambling.

      I agree that some of the problem is to do with the REF and with how universities are managed. However, the only way to change how all universities are managed is by changing their incentives. This can happen by changing the way the EPSRC run things. The other important change should be to minimize the weight of grants (and the impact) in the REF, but that is a different battle.

    • Disillusioned fledgling academic Says:

      Dear Yiftach – I totally agree with you that the whole grant system makes a lot of capable people waste their time on a process that everyone agrees can be a lottery. It would indeed be much better if they could streamline the application process, perhaps not asking for all the information at the first round, or perhaps not following a one size fits all policy for the application form. I don’t know what the best method is, but certainly the current process is overly bureaucratic.

      I also agree that it is unfortunate that a system that has many ingredients of a lottery is relied upon for so much more – a 5 year fellowship guarantees a permanent job if you play your cards right, and a successful grant record helps you secure more grants – both through increased output and the fact some calls are only open to those with enough previous grant income. This positive feedback will only reinforce any bad decisions made.

      I know of many well deserved awards made by EPSRC, as well as many very very bad decisions. Unfortunately the donwsides of the bad decisions seem to more than outweigh the good ones.The problem is that there is no quantitative handle on when they mess up. What actual value do EPSRC add as an investment body? How much do they scrutinize the impartiality and judgment of their referees and panels? What is the standard deviation in referee assessments on a single proposal? What percentage of referees consistently grade proposals badly or well? How much do they scrutinize the advice of the advisors that tell them which areas should be shrunk or expanded? Which advisors led them to that decision? How do they choose their advisors and referees?

      It seems quite hard to obtain all of this information and many decisions appear to be made behind closed doors. As a professional investment body, EPSRC (and RCUK in general) should be taking very good care of the taxpayers money that they invest and should expect rigorous scrutiny from the academic community.

    • Yiftach Says:

      Dear Disillusioned, I don’t think the main problem is that the EPSRC is giving the grants to the wrong people. It is probably an impossible mission to judge research proposals in math, so I don’t believe we should expect accurate results. The main problem is that even if the EPSRC would have done a perfect job in giving grants to the best people it would still cause (in math) more damage than good. I am not sure what the success rate for grants is, but let’s says it is 20%. Then still 80% of the applications were a complete waste of lots of time and energy (of the applicants and of the referees). Worst, even if we look at grants that were funded, if they are interesting, it is very likely that people would have worked on these problems anyhow. So even within the winning projects there is a lot of waste.

      What I believe should be done is assigning 1-2 fellowships to each department and the rest will be given directly to the best candidates to take it wherever they like. Then some money should be assigned for travel and meeting with a very simple application process. Projects should be funded only in very special circumstances.

  10. EPSRC update update « Gowers's Weblog Says:

    […] brief post is to update further a recent post that was itself an update on the situation with EPSRC. The good news is that EPSRC postdoctoral fellowships in mathematics are now available for […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: