Anecdotally, during my (fairly short-lived) academic career, in which I did research with three different groups, 2/3 of them were engaging in fraudulent resear

Anecdotally, during my (fairly short-lived) academic career, in which I did research with three different groups, 2/3 of them were engaging in fraudulent research practices. Unfortunately the one solid researcher I worked for was in a field I wasn't all that interested in continuing in, and as a naive young person who believed in the myth of academic freedom and didn't really understand the funding issue, I jumped ship to another field, and found myself in a cesspool of data manipulation, inflated claims, and all manner of dishonest skullduggery. It all comes down to lab notebooks and data policies. If there is no system for archiving detailed records of experimental work, if data is recorded with pencils so it can later be erased and changed, if the PI isn't in the habit of regularly auditing the world of grad students and postdocs with an eye on rigor and reproduciblity, then you should turn around and walk out the door immediately. As to why this situation has arisen, I think the corporatization of American academics is at fault. If a biomedical researcher can float a false claim for a few years, they can spin their research off to a startup and then sell that startup to a big pharmaceutical conglomerate. If it fails to pan out in further clinical trials, well, that's life. Cooking the data to make it look attractive to an investor - in the almost completely unregulated academic environment - is a game that many bright-eyed eager beavers are currently playing. As supporting evidence, look at mathematical and astronomical research, the most fraud-free areas of academics. There's no money to be made in studying things like galactic collisions or exoplanets, the data is all in the public domain (eventually), and with mathematics, you can't really cook up fraudulent proofs that will stand the test of time.