Episode #2: Richard Thomas (Imperial) on an EPSRC Mathematics Panel 30/31 January 2012
I just finished a 2-day EPSRC panel, the first I think since shaping capability took hold. Here are some observations.
(I will sloppily and incorrectly abbreviate ‘statistics and probability’ to ‘stats’ throughout.)
Some of the views I ascribe to the panel might just be my own: inevitably I ended up speaking too much, and other panel members were more reserved — something I aspire to but never quite manage.
First, the good:
The panel of 10 were all superb, sensible, and no one was territorial. The 4 from stats largely seemed to hold the same views as the rest of us: that (a) the new rules devalued fellowships that people from stats had won in open competition in the past, and that (b) of the stats fellowship applicants we were considering, a few were excellent, many were quite good, but many of those would not have been funded in full competition.
One advantage of the new rules, that I’d not previously been aware of, was that there was a feeling that that people in applied maths and stats (particularly the latter) generally mature slower than pure mathematicians who are expected to be independent straight after their PhD and proven (world leading or otherwise) within a few years of postdoc. Removing pure mathematicians from fellowship consideration therefore helps stats fellows, though of course there are much better ways of achieving the same effect.
We had a day of responsive mode proposals and first grants (in all areas of maths) to evaluate, then a day of stats fellowships and a programme grant.
I was impressed that EPSRC seemed genuinely intellectually honest and keen to ‘tension’ the stats fellows against the non-stats responsive mode proposals, and understood that fellowships ought to be a higher standard but were now lower. They were not going to fund things significantly below the level of applications from the old regime.
Encouragingly, Philippa Hemmings afterwards gave me the impression that everything would be fed back to the Strategic Advisory Team, for instance to consider opening up other areas of maths to fellowship applications: particularly postdoctoral fellowships, where there were almost no applications in stats.
If, say, postdocs were indeed opened up wider, Philippa made it clear it could not be to the whole of maths because she has to fit it inside shaping capability. Even if they don’t know what shape they’re after, they’re committed to shaping anyway. This of course would be divisive; potentially pitting pure vs applied etc.
And I fear the application dates would be March and September, meaning anyone good enough to be an international candidate could not participate due to the foreign (especially US) job timetable. EPSRC’s bureaucratic convenience has to come before what’s good for science.
The loss of real star applications to the panel was clear. Even more worrying to me was that I now see that EPSRC’s promise of funding only excellence, however well-intentioned, will inevitably be effectively meaningless, as we all feared. Over time the standards of old applications will be forgotten, and weaker applications will appear excellent because referees want to be nice (they gave top scores of 5s and 6s while hinting work was not exactly internationally leading) and because they’re not being asked to compare internationally (hence lots of ‘this is the best person in his field in the country’ comments).
On the subject of postdoc fellowships, I got the impression from talking to people that EPSRC were under pressure from BIS etc to differentiate their offerings from those of RS, ERC, etc. Hence perhaps their blatant abuse of statistics in comparing their fellowships to Newton fellowships (1 a year in maths, not the 50 they claimed), ERC starter grants (young postdocs not eligible; mostly given to people with permanent jobs) etc etc.
I also got the impression that while EPSRC had been keen to bulldoze through the worst excesses of shaping capability, justifying things with dishonest data after the fact, they are now hoping to enter a phase of winning people back, making adjustments based on what they see at panels, and making the system work as well as possible (within the shaping straitjacket). That impression may have been given to keep us from causing more trouble, of course.