Thursday, November 13, 2008

Peer review problems

As the following letter by Marc Hendrickx in today's Sydney Morning Herald points out, it ain't all plain sailing in the world of academia!

"The peer-review system for publication in science journals is inherently flawed and in desperate need of an overhaul. Emeritus Professor Brian Milton's comments (Letters, November 13) in support of the system reveal a naivety that appears common among senior academics raised at a time when academic performance, pay and position were not so closely tied to publishing papers.
The pressure to churn out papers has put exceptional demand on the peer-review system and resulted in many low-quality papers slipping through - often by watering down or exaggerating conclusions to pander to the whims of editors or high-profile reviewers.
The chances of publication are boosted by adding the names of well-credentialled supervisors or senior academics who typically contribute nothing or very little. This would be considered fraudulent in most other fields of publication.
In my experience, blinding reviewers to the names of authors and vice versa is practised by only a few journals. Reviewers are not paid for their time and there is little recognition of their work. Their opinions are often disregarded by editors keen to fill journals and meet publication deadlines.
The failures of the system were highlighted by the paper by the epidemiologist John Ioannidis, presumably peer-reviewed, cited by Michael Duffy ("Truly inconvenient truths about climate change being ignored", November 8-9). It suggests most published findings are proved false within five years of publication. If the system is as good as Emeritus Professor Milton suggests, how can this rubbish, produced by "pure-minded" academics supported by "independent" government funds, get a tick?
The best tests of science remain the long-term reproducibility of results and strength of theories against falsification. The current theory of dangerous climate change driven by anthropogenic carbon dioxide emissions has not yet passed those tests. Getting published these days is no guarantee of quality."

No comments: