In the biomedical sciences (where the original author is speaking from), there is a dichotomy: yes, most researchers don't believe that a single publication is

In the biomedical sciences (where the original author is speaking from), there is a dichotomy: yes, most researchers don't believe that a single publication is the final word on a topic and that in theory anything could be retracted or modified, but in practice that's how the system works. Publications--both how many you have and where they are--are the prime metric for promotions, tenure, and overall success. Furthermore, when applying for funding your publication record is taken into account, as well as your promotions, tenure, etc. It is currency in the life of an investigator. Not to mention that research is all about who comes in first; the incentive to validate published work is by and large not there. I recall a Nature article on data reproducibility: only 11% of the TOP, PEER-REVIEWED PAPERS were able to be reproduced[1]. [1] http://www.nature.com/nature/journal/v483/n7391/full/483531a.html http://www.nature.com/nature/journal/v483/n7391/full/483531a...