This week’s Economist delves into the increasingly murky world of scientific publications, with disturbing news.
For example 70% of key scientific results at the foundation of crucial platforms of understanding in cancer research cannot be replicated. One problem is the system of check and balances; the peer review process that is suppose to filter out poor work and audit the good.
Behind this is a system in trouble. The peer review process and the professionalism of scientists has been undermined by the relentless need to publish to “get on” in their careers and the sheer number of scientists numbering in the millions.
In essence this is a system failure – a process that was not scalable, without incentives for the system to self-correct as it grew.
A number have been mentioned. Open up the process to full transparency; publish the data for all to see and scrutinize. This is unlikely to drive fraud down but the threat of being discovered might have some effect. Also, provide self-checking statements to filter out obviously devious practice. And software that self-monitors the process of drafting and analyzing the data .
Ultimately the incentives to cheat or be incompetent need to be tackled. The Arnold Foundation aims to help scientists up their game…