Tuesday, 29 May 2012

"Death of Science"

At it's launch earlier in the month Science for the Future proclaimed the "Death of UK Science" and had a coffin with a floral wreath with the word "Science" delivered, along with a petition, to 10 Downing Street. Over the subsequent two weeks several people have complained that this launch was "divisive", "misguided", "risible hyperbole". However, the EPSRC's own data show that all is not as rosy as they would have everyone believe. 

To put this in context let's review the argument that EPSRC use which explains it's current "demand management" policy on the back of falling success rates. This was discussed in a Nature News Blog http://blogs.nature.com/news/2011/09/amid_falling_success_rates_one.html

The first graph in the one EPSRC want you to see as it shows increasing success rates compared to other research councils. 
(graph 1)
The second graph shows the true picture, which EPSRC fail to address and that is the fall in applications is happening faster than the fall in funding.
(graph 2)
In order to determine whether the accusation of "risible hyperbole" is valid let's look at the data for physical sciences. This data is taken from the EPSRC's own website, and I thank Prof. Chris Hayes for providing this analysis.

2004-2005 Proposals Funded Proposals Submitted Funded Value £M Success
Chem 184 526 40.5 35%
Phys 126 332 37.8 38%
Mat 210 677 47.9 31%

520 1500 126.2 35%










2005-2006 Proposals Funded Proposals Submitted Funded Value £M Success
Chem 174 600 49.2 29%
Phys 113 290 37.8 39%
Mat 171 814 47.1 21%

458 1544 134.1 30%





2006-2007 Proposals Funded Proposals Submitted Funded Value £M Success
Chem 193 603 49.2 32%
Phys 125 329 46.6 38%
Mat 156 624 53.2 25%

474 1497 149 32%





2007-2008 Proposals Funded Proposals Submitted Funded Value £M Success
Chem 149 552 42.4 27%
Phys 108 338 33.7 32%
Mat 201 543 58.3 37%

458 1431 134.4 32%





2008-2009 Proposals Funded Proposals Submitted Funded Value £M Success
Phys Sci 230 1045 100.4 22%





2009-2010 Proposals Funded Proposals Submitted Funded Value £M Success
Phys Sci 217 804 88 27%





2010-2011 Proposals Funded Proposals Submitted Funded Value £M Success
Phys Sci 151 503 85.8 30%
















Proposals Funded Proposals Submitted Funded Grant Value £M
2004-2005 520 1500 126.2
2005-2006 458 1544 134.1
2006-2007 474 1497 149
2007-2008 458 1431 134.4
2008-2009 230 1045 100.4
2009-2010 217 804 88
2010-2011 151 503 85.8

A number of conclusions can be drawn: 

1) In physical sciences the pre-demand management success was for the most part above 30%.

2) Total EPSRC funded activity has decreased. In 2004-5; 520 proposals were funded, but this dropped to 151 by 2010-11, when the total number submitted was only 503! The introduction of FEC does not account for this fall as the 2005-8 period, when FEC was working through the system, is pretty stable.

3) The biggest change happened in 2008-9, when a single physical sciences panel (which operates as 3 tensioned panels) was introduced.

4) Since demand management was introduced in 2008-9, the number of grant applications had fallen sharply. To get some idea how catastrophic the situation is 1150 chemists, 1700 physicist and 500 material scientists were returned in the 2008 RAE, which makes 3350 scientists who could apply for EPSRC funding. Only 151 proposals were actually funded! What is very worrying is that we know scientists are being excluded from applying to EPSRC due to blacklisting and now shaping capability restrictions on areas that will be considered for funding.

5) It is important to note that the money spent by EPSRC in physical sciences is being eroded over time to £85.5M, but the decline in money is nowhere near as great as the decline in proposals being submitted.

This can be visually represented as graph 3.

 (graph 3)

Just so that we are under no illusions, the EPSRC's own data supports the warning of Science for the Future at its launch this month. The "Death of UK Science" (or the physical and mathematical parts of it) will occur in under 2 years if the current policies are maintained!

I would therefore suggest that those who labelled Science for the Future as peddling "risible hyperbole" are the "misguided ones" and are themselves guilty of the scientific equivalent of fiddling while Rome burns.


8 comments:

  1. Surely EPSRC would say that the above figure indicates the dramatic success of demand management. Far fewer people apply for a not-too-much-smaller total pot. Result! Fewer, larger grants, to far fewer superstars; that was the original idea, wasn't it? It means the RCs can spend less time on administering grants, and far more time on the more intellectually interesting process of deciding on funding priorities, as a way of continuing to justify their existence.

    ReplyDelete
  2. And that is is exactly the point! If anything is risible it is current EPSRC policy.

    ReplyDelete
  3. I'm sorry but what exactly is your argument here? Getting blacklisted by EPSRC is pretty difficult: if you submit 1 proposal every 8 months it is by definition impossible to get blacklisted. So is the drop in applications entirely caused by fear? Your calculation that 3350 scientists submit 500 proposals a year is interesting but has little bearing on EPSRC's policies. EPSRC is not perfect but the drop in applications might well be a good thing.

    ReplyDelete
  4. @ Klaas, that is the point. Applications have dropped drastically since demand management was introduced with no increase in success rate (for physical sciences). I would suggest that this decrease is down to 4 factors: 1) as you suggest, fear of being blacklisted. Some institutes may see this in a negative light and hence people would prefer not to submit proposals rather than risk the chance of blacklisting; 2) The decision by EPSRC to ban applications in certain areas (fellowships) and discourage applications in other areas due to shaping capability (surface science, synthesis); 3) the reduction in the overall funding; 4) the ban on resubmissions.

    I cannot see how reducing the number of proposals being submitted can be viewed as a good thing. I just recieved my EPSRC college membership report for the last 12 months. It says that average number of review requests per college member is 2 per year, yet only 63% of the college were actually sent proposals to review. This hardly seems like a system under pressure (as EPSRC would have us believe).

    ReplyDelete
  5. There are about 3700 college members, at 2 requests per year, that's 14 requests per submitted proposal. If a typical proposal has 3 referee reports then that implies 11 non-respondents? That doesn't seem to add up. Anyway, why is the funded grant value only ~£100M. I thought that EPSRC's total spend was about £1G.

    ReplyDelete
  6. @Anonymous For the refereeing data EPSRC also say that the average number of review requests for a non-college memner is 1 per year. Of college members who are sent a proposal to review 28% either do not respond, send an excuse or submit an unusable review. It is also worth noting that only 298 college members sat on one of 96 panels over the last 12 months. This just seem to show that the peer review process is not in crisis as EPSRC say it is.

    The figure of £85.5M is th emoney allocated through the phuscial sciences panel. It does not include money allocated by other means.......(see previous blog post)

    ReplyDelete
  7. One major fallacy is that it is better to fund 'supergroups' in big chunks than many smaller groups. Sure fire way to decrease the diversity of new knowledge, etc. etc. Certainly a less efficient way to spend the money too (students get poorer training in larger groups, generally speaking, in my opinion (may be based on first hand experience)).

    ReplyDelete
  8. @dange82 That is certainly one of the problems with EPSRC policy at the moment. EPSRC has yet to produce any evidence to support their belief that large cohort trained PhDs are 1) better trained, or 2) do better science. They have also failed to demonstrate that their "developing leaders" policy, where money is given to selected individuals without peer review delivers better, "more impactful" science.

    ReplyDelete