Publish and perish
The Undoing of a Star Scientist The defenders of traditional journals - esteemed publications such as Science and Nature - argue that online publications are not able to provide such rigorous screening. The widely publicized case of a scientist sending numerous - and fudged - articles to these journals undermines that claim. True, the journal editors argue, reviewers cannot be expected to spot every flaw.
But reviewers should be able to pick up on identical data submitted for separate results, unrealsitically precise data, or data that violates the laws of physics. Shouldn't they? The thing is, these articles wouldn't have lasted ten minutes on the web before someone spotted the anomalies - and saved scientists (and readers) two years of wasted work. [OLDaily]
 From the article:
Schön substituted data in his published papers, supplying fake graphs that he told investigators "looked better" than the real graphs.
He also used the same graph in a dozen papers on different experiments. And his data were often far too precise, far beyond reasonable statistical probability.
Also:
Bell Labs isn't the only scientific institution to be smacked with charges of fraud recently. The Lawrence Berkeley National Laboratory announced in June that it had disciplined a researcher for fabricating the results of a published experiment.
The experimenters claimed to have discovered the world's heaviest atom, element 118. These are cases of high-publicity research where the work was given much attention. But how many unreliable articles like this, but with less extraordinary claims, pass through? How many people are out there trying hard to build on card castles?
Let me quote a scientist from this piece in New Scientist:
"There is nothing more important for a laboratory than scientific integrity. Only with such integrity will the public, which funds our work, have confidence in us."
Amen to that. And who could imagine a better integrity-enforcing, trust-generating self-correction mechanism than a thoroughly open process? Especially now that we have had the means to implement it for a long while...
 10:32:00 AM
|
