Good science is good. Poorly designed science is bad.

[FarCenter]

37. C. Glenn Begley worked at Amgen, Inc., who could not confirm the preclinical studies

The problem is that preclinical research conducted by universities and medical research establishments, much of it government funded, could not be reproduced when corporation in the biotech or pharmaceutical industries attempted to base new therapeutic techniques and drugs on the research.

This indicates to me that --
- there is a large amount of government funded research that is useless, since no one finds out that it is irreproducible by attempting to make use of it,
- the existing mechanisms for evaluating publishable research via peer review are not weeding out bad science, and
- when results are found to be irreproducible they are in most case never retracted unless strong evidence of deliberate fraud is found.

Even when fraud is found, in many cases the principle investigators get away with only a slap on the wrist due to universities and institutions reluctance to have their names associated with it. And very rarely is the principle investigators tenure revoked.

The paper by C. Glenn Begley & Lee M. Ellis is "Drug development: Raise standards for preclinical cancer research" and is freely available from Nature at http://www.nature.com/nature/journal/v483/n7391/full/483531a.html