Saturday, June 30, 2012

tenure blog may 2012 1

Retracted papers, imprisoned scientists; the dismal future of funded science


A recent NY Times article by Carl Zimmer refers to a 2011 study in Nature that, in the preceding decade,  indicated retractions had increased about 10-fold in absolute terms and 7-fold per unit published; 
Why is this happening? There are a variety of reasons. First of all, as U Vermont’s Eric Peohlman – who was imprisoned for a year for lying on federal grant applications and fabricating data – points out, the pressure to keep labs going can be enormous. It is not only for oneself as PI that one works; there now is very problematic over-production of PhD’s in what has become a pyramid scheme. Currently, 15% of PhD’s can expect to be on tenure track within 6 years versus 50% in 1973.
As previously signaled here, much of the low-hanging fruit in science has already been picked. A century ago, Einstein and Dirac could gain unprecedented insight into nature with what is relatively simple math. Now there are too many researchers chasing too few facts. The new  MO has rather portentously been called "big data' with statisticians to be employed in the alchemical transmutation from sows' ears.
Fang and Casadevall did an analysis of retracted papers, concurring with the burgeoning retraction problem, and remarkably indicating that the New England Journal of Medicine had the biggest rate. Peohlman’s work also was biomedical. It is fair to say that people are dying because of a worldwide flawed administration of science. 
The idea of the research university reached Ireland about 2000, just as its limitations, particularly pedagogical,  were becoming evident elsewhere. Stupidity is only one part of it; the extravagant state funding of science affords many opportunities for scams like Medialab. Likewise, the idea of launching native industries on foot of work in our academies has been dispensed with; it is now understood that the idea is to do research beneficial to multinationals on state university campuses. 
So what can be done? This writer’s choice has been to avoid the whole rat race and publish open source peer-reviewed material, much of which has ended up on this website. Many subjects (like Biology) function in such an egregious absence of theory that there is space to write good fundamental papers. However, that is only an interim solution.
A first step is insisting that all data and software used in research should be included in the journal submissions. Next, anonymous review must be dispensed with; in fact, it would be simplicity itself simply to use the open science model simply to post papers, run the programs, and solicit critique. Ratings of reviewers could be performed with one of the many good ranking algorithms that exist. At some point, the paper could be deemed part of the corpus of that journal.
However, there are bigger issues here. The first is that, whether we like it or not, much of university education is going to be done over the web through sites like coursera.org (which now has 30 or so courses). Harvard and MIT, getting the hint, are about to collaborate on a giant alternative;

  http://www.edxonline.org/about.html.

 It is difficult to see how states can continue to justify to their citizens the setting-aside of land grants for activities that can be done at home. In many ways, this represents a lucky escape; the mantra “autonomous statutory responsibilities “, used so often in our parliament to justify inaction in the face of what was criminal activity by university management, indicates the grim future planned by the Irish state . 
There are several research themes that should never have been funded by the state. My favourite such is software localization; the existence of a national center for this, which replaced a successful degree program, is egregious – this should be paid for by Microsoft and other companies that benefit from it. In fact, beyond a few micro-specializations, there is no reason for Ireland to fund science at all, as it is simply duplicating work done better elsewhere. Countries like the USA might consider whether it is appropriate for private corporations to do drug development and testing. Here the USA’s massive resources might help.
In a decade’s time, all the above will look self-evident
Seán Ó Nualláin Ph D Stanford 28u Bealtaine  2012

PS 3 Meitheamh 2012

It's worth noting for a variety of reasons  that the QS university rankings are absolute nonsense, in particular placing good old Cambridge ahead of such upstarts as Stanford and Berkeley;

 http://www.newstatesman.com/blogs/david-blanchflower/2011/09/world-university-faculty