Download Lectures on Biostatistics (1971).
Corrected and searchable version of Google books edition

Download review of Lectures on Biostatistics (THES, 1973).

Latest Tweets

Jump to follow-up

Today the front page story in Times Higher Education concerned a letter, Modest revolt to save research from red tape a letter signed by 20 people (including me).

My reason for signing the letter was that I am interested in how to get good science, and I am concerned that the government, and many vice-chancellors, are getting it wrong.  Read also the wonderful essay. The Mismeasurement of Science, by Peter Lawrence (who is one of the signatories of the letter).

The letter had an accompanying article by Zoe Corbyn, Scientists call for a revolt against grant rule they claim will end blue-skies research, “Letter blames research councils’ policies for fall in number of UK Nobel laureates, reports Zoë Corbyn”.

It also had a leading article by Ann Mroz, Leader: Short-term outlook, no blue skies, with the subtitele “By forcing academics to do R&D for industry, ministers stifle the type of curiosity-driven work that delivers the biggest bangs”.

The “revolt” proposed in the letter is very modest indeed. We propose merely that those who are asked to referee grants will confine themselves to the part of the application that they are able to judge. In most cases that will not include the two pages of arm-waving guesswork about how the results might save the nation twenty years hence.

Everyone loves the great discoveries, but great discoveries are very rare  If you want them, you must hire the best people you can, let them get on with it, and accept failure. Lots and lots of failure.

It’s a pity, but nobody knows any better way to do it.  But at least please don’t try to impose worse ways.

The writers point to the fact that since 1980, roughly the time when bureaucracy started to run rampant,

” . . . the changes have resulted in almost a tenfold decrease in the rate at which researchers at UK universities win Nobel prizes.”

OK, it’s true that we have no way to be sure that the undoubted increase in top-down control has actually been the cause of the smaller number of Nobel prizes, but there is no doubt that it has wasted a great deal of time that could have been spent more profitably.  It’s intriguing that 1980 is also the time that Francis Wheen identified as being roughly the start of what I like to call the age of endarkenment.

“But what is the point of having a second-to-none academic sector if its commitment to innovation is not matched by commerce and business? British industry’s relatively low investment in science and technology has long been a serious problem that our proxies and governments have consistently ignored. Academics are a much easier target.

While understanding that our proxies are often in a difficult situation, they must become more courageous in dealing with Government or they won’t have an enterprise worth protecting. In recent years, they have acquiesced in subjecting academics to withering barrages of control, and researchers’ lives have become bureaucratic nightmares.

The latest turn of the screw inflicts yet another distracting burden, namely that of requiring prospective researchers to write “a two-page impact plan in addition to the case for support”.

We the undersigned suggest that it is time for a modest revolt.  We would urge that reviewers for grant applications decline invitations to take these additional pages into consideration and confine their assessments to matters in which they are demonstrably competent.

Indeed, in research worthy of the name, we are not aware of anyone who would be competent at foretelling specific future benefits and therefore in complying with the request in any meaningful manner.”

I cannot speak for the other signatories, but I expect they, like I, have nothing against industry, and nothing against applied research. The problem. of course, is is that the really big developments are never obvious at the time they are made, and still less at the time you are writing a grant to do the work.

The government, and far less excusably, the Research Councils, often appear to be populated by the sort of people who would have told Michael Faraday to stop wasting his time with those wires and coils, and get to work on something really useful like designing better leather washers for steam pumps.

Certainly there can be no more inappropriate time than this to be urging universities to act more like businesses.  The sight of a bunch of very highly paid bankers who proved incapable of distinguishing good investments from worthless ones is pitiful.  That was their job and they failed to do it on a grand scale,   The fact that the bastions of conservatism, who have spent most of their lives telling us how wicked it is to bail out failing businesses are now taking many thousands of pounds from me, inspires feelings that are perhaps better left unwritten.

It is all part of a much more general move to trust nobody to do a job properly, and, worse, to then construct a hierarchy of regulators that consists largely of people who have themselves failed at the job they are supposed to be regulating.  They then lay down the law about how to do the job to those who have succeeded.

Sadly, this process is evident not only in government, and not only in Research Councils, but also within universities themselves.  We have the managerial culture that disempowers anyone who does real research and teaching. We have vice-chancellors who support degrees in various sorts of quackery which they can’t possibly believe to be true, like Professor Petts of Westminster (where Amethysts emit high Yin energy), and Michael Pittilo of Robert Gordon’s University Aberdeen. We have vice chancellors who defend corrupt practices by the pharmaceutical industry, as in Sheffield. We have lunatic assessment targets, as at Imperial Medical School. And of course we have some curious “Morals in High Places“. And we have the vice-chancellors’ Trades Union, UUK, that does nothing to stop these things.

It really is time that the Research Councils, and university vice-chancellors resisted a bit more strongly government pressure to do things in a way that can only harm the economy (not to mention the universities) in the long run.

They have nothing to lose but their knighthoods.


One of the signatories of the letter mentioned this beautiful quotation..

“Every important discovery is by definition unpredictable. If it were predictable, it would not be important.”

Freeman Dyson (speech when receiving the Britannica Award for excellence in the dissemination of learning for the benefit of mankind, part reported in the Independent, March 12th 1990)

He added

“I intend forthwith to copy and paste this into the section of any grant application which asks for impact etc. No more and no less.”

There has already been some criticism of our letter. It has come, though, not from scientists, but from “science policy professionals”, a new breed of hangers-on who mostly seem to be too far from the coal-front to know how science works.

Testing hypotheses is written by the head the Research Councils UK Strategy Unit. He chooses to concentrate on the secondary matter of Nobel prizes, rather than answer the real questions. Fortunately Philip Moriarty has tried to keep him to the point, though without any very satisfactory reply.

The Prometheus blog from the USA has an article by Ryan Meyer that has given rise to some comments and questions. I have asked him to explain to me what he means by the “social context” of work on single ion channels.

Print Friendly, PDF & Email

12 Responses to How to get good science: again

  • Dr Aust says:

    Mike Taggart’s article in Physiology News last Summer, which you might remember, covers somewhat similar ground, David.

    It also generated a bit of follow-up correspondence here.

  • Chris says:

    Another article by Stephen Quake covering similar ground appeared recently in the New York Times.

  • phayes says:

    ” . . . the changes have resulted in almost a tenfold decrease in the rate at which researchers at UK universities win Nobel prizes.”

    Well never mind. There are no exclusive rights in Nobel prizes. Acquiring lots more patents would be far more productive and lucrative. As an American academic (IIRC) once said: “we are creatures of the colleges and universities within which we work, whose very essence is about the creation and dissemination of intellectual property”. 😛

  • phayes says:

    (Science (and other knowledge) is about intellectual property, not ideas, these days – in case you anti-Bush, anti-Blair, Bolshevik commie pinkos hadn’t noticed) 😀

  • David,

    Thanks for this excellent post. You, and the readers of DC’s Improbable Science, may be interested to know that Steven Hill, Head of RCUK’s Strategy Unit has responded to the letter/articles in the THE via his blog, Testing Hypotheses .

    Hill implies in his blog post that it is only “a small number of senior researchers” who are concerned with the imposition of economic impact criteria. He seems to have forgotten that RCUK carried out a “consultation” in mid-2007, asking all UK universities for feedback on plans to impose economic impact criteria in peer review. The feedback was overwhelmingly negative. To get a flavour of the feedback, Nottingham’s, Oxford’s, Cambridge’s, and Glasgow’s responses to the consultation are described in a response to Hill’s blog post here . All university responses are, however, available at the RCUK website . They make for very interesting reading.


  • James Ladyman says:

    You are quite right to make the connection between the managerialism that infects universities and the ‘impact’ agenda. Researchers who are the lifeblood of higher education are continually required to be accountable via systems and procedures designed by individuals who may do little by way of teaching and research themselves, and who have no more and usually less expertise than the rest of us. How many of us have had the experience of being given stern instructions to increase our numbers of postgraduate students and grant income by managers whose own record in those respects is utterly dismal? There is no accountability of the managers to the academics who do the teaching and research, and now many of the former are getting off on the tough decisions they are going to have to make in the current economic climate.

  • Amanda Goodall says:

    I believe that many of the issues raised here are about leadership. In my longitudinal work I show that research universities that are led by top scholars improve their later research performance. On average, the best universities in the world — that win the highest number of Nobels — are led by outstanding scholars. Why? I interview 26 heads in UK and US research universities, and in short most beleieve that scholars create the best conditions for other top scholars. Top researchers tend to prioritize scholarship. Unfortunately, in the same way I would argue that managers often create the best conditions for other managers. Like-selects-like. In my forthcoming publication (Princeton U Press) I argue that top scholars should lead research universities. Also, and possibly even more important, the best researchers should lead those institutions that fund research.

  • Amanda Goodall. Thanks for the comment. I’d read with interest about your work

    I can certainly agree that managers-select-managers, so the evolution towards a managerial culture has a dangerous positive feedback built into it, Our recent regrading system rewarded managers according to the number of underlings they had, thus providing a direct incentive to inefficiency and proliferation.

    Can you tell us whether there has been a change over time in the proportion of top universities that are led by great scholars? It’s tempting to think it has declined, but I can certainly think of counter-examples.

    The same problem exists at lower managerial levels too. Anybody who wants a job like that is, almost by that fact, not suitable for it.

  • rmeyer says:

    Your quotation from Freeman Dyson is beside the point. Just because some benefits were unpredictable does not mean that we need to avoid actively seeking benefits. Some of the most important scientific breakthroughs were totally unexpected, but came from research on particular problems. Think vaccines, DNA, radar, aerodynamic flight…

    Trying to orient research such that it helps us deal with particular problems does not corrupt it, nor does it lessen the chances of a breakthrough.

  • Good heavens, nobody said that we should “avoid actively seeking benefits”. Nobody said that we should avoid working on useful applications. Nobody said that working an a social or medical problem corrupted science. These are so many straw men.

    Your assertion that concentrating on applied research does not “lessen the chances of a breakthrough” is sheer guesswork. Think of penicillin, lasers etc etc.

    Or, better still, think what would have happened if the policies you advocate had been applied to medicine in 1900. Progress would have been frozen and great harm would have been done to the progress of medicine.

    As far as I can tell from Google, your own interest is in policy regarding climate change (from a somewhat sceptical point of view). Perhaps you are unaware that in the case of much human disease, ignorance is so profound that there is no real idea of what causes the condition (back pain is not a bad example) so there isn’t even a hypothesis to be tested? I was recently told by an administrator at UCL, “first publish your Nature paper, then translate it to the clinic”. He has yet to answer my questions of how we should translate to the clinic the results in Lape et al. (2008).

    The fact of the matter is that translational research is fine as long as you have ideas to translate. The policies you seem to advocate would soon ensure that we didn’t.

  • Given that the inspiring quotation from Dyson failed to make an impact on rmeyer, how about the following?:

    Research is always about risk taking, no matter whether the risk involves failure to meet a certain set of expectations or failure to create truly new,
    significant knowledge or understanding of a problem. To be clear, if the outcome of the effort can be anticipated, it is highly questionable whether this effort should be called research.

    rmeyer, would you agree with this statement? It’s taken, by the way, from the report of the recent International Review of UK Materials Science commissioned by the Engineering and Physical Sciences Research Council (EPSRC). That report also contains the following advice from the panel:

    … the UK materials research activity is, in some cases, too focused on shorter-term economic outcomes. It is appropriate for some university research to be focused on near term economic benefit, but universities are tremendous sources of long-term innovation, and this role cannot be compromised.

    In several fields ranging from more traditional to very high-tech fields, the concern was expressed regarding the funding level for core knowledge creation.

    .. further concern is over the balance between short-term research that might be very relevant to industrial needs replacing longerterm, more speculative research.

    So, in addition to arrogantly ignoring extremely strong negative feedback from UK universities on the introduction of economic impact criteria into peer review (see here ), EPSRC also saw fit to disregard advice from an international panel it itself commissioned to review UK Materials Science.


  • Amanda Goodall says:

    David, thanks for your comment and question. You are asking whether there is any longitudinal evidence that better scholars, as leaders, improve the later performance of universities? My key quantitative study looks at 55 UK research universities’ performance in the RAE (paper: ‘Highly Cited Leaders and the Performance of Research Universities’ on my website). In it I show that those universities that appoint better researchers as leaders go on to perform better in the RAE. I have looked at the recent RAE results and have a similar finding, although it is not yet written up. With regards to whether there is a decline in top scholars becoming university leaders in the UK, that’s an interesting question. I can certainly check that in my data. My instinct would be to suggest that it is, because of the overwhelming interference of government these days, and concomitant bureaucracy. But I am about to do a small study related to this and will get back to you! My hypothesis actually came from having worked with two university leaders. To put the argument into another setting — it would be unheard of for a law firm or accountancy practice to appoint either a weak lawyer or accountant or non-lawyer/accountant as a leader. Quite why we have different standards in universities where expert knowledge is so central to our core business is remarkable.

Leave a Reply

Your email address will not be published. Required fields are marked *


This site uses Akismet to reduce spam. Learn how your comment data is processed.