Saturday, 19 October 2013

The Economist Examines the Self-Checking Processes of Science

In a leader, an article, a discussion and a poll, the Economist asks "Has Science Gone Wrong?"

The critiques are as follows:
  • The bias in journals and the pressure to publish novel results results in many results that do not stack up when replicated or really checked
  • Peer review is not doing its job in weeding these out
  • Replication of results is thankless, time-consuming and usually not chosen by funding agencies
  • Negative results are difficult to publish and hence this is not usually done
  • Estimations of the number of errors in published results is probably underestimated
  • Much data, methods and program code is effectively not openly shared
  • Scientists downplay rather than admit their mistakes
Although mostly concentrated on bio-medical research, the critiques have a lot of traction elsewhere in science.




A lot of interesting issues on which simulation could help understand -- given you admit when they don't work, of course ;-)

No comments:

Post a Comment