Saturday, 15 November 2008

Peer reviewed scientific publications

Letter in SMH.

The system of peer review in science has become corrupted, with rorts such as rampant cross-authoring (putting names of non-contributing colleagues on papers to build their CVs) and "coffee time" agreements to approve each other's works.

Almost all grants, funding, pay, promotions and accolades are as a result of numbers of papers published. A far more useful statistic would be the number of citations, a measure of the work's usefulness.

As stunning as it is that most published papers are found to be false within five years, two other statistics are even more blunt: 50 per cent of publications are never read by anyone, and 95 per cent of peer-reviewed science is never cited, other than by the authors themselves.

In other words, most peer-reviewed published science is useless rubbish.

Jon Jenkins, Bogangar


I've often suspected that the quality of peer-reviewed science publications suffer from a 'you scratch my back, I'll scratch yours' arrangement.

2 comments:

David Palmer said...

Hi Gordon,

Somewhere deep down in my pile of climate change papers I have an article done by an Australian several months ago on the the 50 or so researchers responsible for the IPCC findings on climate change.

Two points were made - the group was dominated by UK Hadley Centre scientists and Americans (but with 3 Australians as well) and few from anywhere else, and secondly they all had a record of peer reviewing one another's work.

Late addition: Here is a summary of the paper appearing in The Australian - http://www.theaustralian.news.com.au/story/0,25197,24315169-7583,00.html

Gordon Cheng said...

A lot of money is going to be spent by policy-makers chasing after this "consensus". It looks increasingly like the scientific equivalent of the sub-prime mortgage.