Slashdot Mirror
Bell Labs fires Hendrik Schon for Data Falsification
Raiford writes "Bell Labs has fired physicist Hendrik Schon for falsifying scientific data. Schon was thought to be a likely candidate for the Nobel prize based on the promise his reported research findings had for the advancement of molecular scale computing. In a Reuters report the dismissal was described as the only conclusive case of scientific misconduct ever identified in the history of the prestigious laboratory."
In addition, fraud or forgery is comsidered far more serious in the mathematical sciences than, say, business, religion, or politics. Fraud is actually not common, and it is considered the most dispicable of dishonorable acts when it is done (as opposed to politics, in which it is expected)
Computer Science is no more about computers than astronomy is about telescopes. --E. W. Dijkstra
I heard a story about a biological researcher who went to some lengths to forge his results. When confronted with the accusation he produced his raw data and even autoradiograms. It eventually came out that he had decided where he wanted his results, and had then used an iodine isotope to create the bands on his blot.
Due to the extreme competition that exists in most research these days, forged results are only going to become more and more common.
It is the hunch that usually leads scientists to study the phenomena/theories in question in the first place. The hard part is devising an experiment to prove/disprove what you're looking for without too many intervening factors that can get in the way. In fact, sometimes just coming up with the experiment itself is worthy of a Nobel Prize.
But scientists should NEVER EVER fake data, no matter HOW STRONGLY they believe they are right. If they're that sure, then they can publish all the theoretical articles they want. But NEVER publish fraudulent data as true. Science is about truth, truth is about absolute, not about hunches. That's why scientists do (or should, if they don't shy away from it) report estimated uncertainties for all experimentally-determined values and data points. If scientists didn't adhere to these lofty expectations, one wouldn't be able to believe any of the journals, which would be a major setback for all fields of science. If you had inherent mistrust of scientists, then science would become just like politics.
I dont know what he was working on, but I would like to give the guy the benifit of the doubt until I can read the report and experimental data.
Sorry, this guy WAS given the benefit of the doubt for many years. His results were irreproducible, which as you know, is one of the main characteristics of science. Everything must be reproducible. He claimed to grow Aluminum Oxide films that could withstand far greater electric fields before breaking down than anyone else on the planet, which is odd considering people mimicked his exact sputtering/growth techniques. For years nobody could reproduce any of his experiments. Much of the discord boiled down to a specific sputtering chamber Schon had back in Germany, where he claimed he was able to grow his thin films. Eventually Schon tried to regrow some films again in this chamber, and said he was unable to repeat his earlier work.
I worked in a physics lab this past summer where nearly every day at lunchtime the professor (Dr. Michael Tinkham, who's rather reknowned in superconductivity circles) would hold up a copy of Physics Today with a picture of Schon and warn us of the consequences of abandoning truth in favor of increased publications.
What Prof. Tinkham pointed out to us is that Schon became something of a minor deity in the realm of experimental physics, getting significant publications, usually quite often in the top physics journals such as Nature, Science, Physical Review, etc. The problem was that he soon had a reputation of greatness to maintain, so he may have gotten a little clumsy regarding data acquisition and analysis, in favor of keeping his astonishing rate of publications steady.
Eventually, things caught up to him. I'm not sure how much of his questionable work was little details that slipped though his fingers, how much was semi-conscious oversight, and how much was flat-out fabrication and fraud. But after he was caught then all his work became suspect.
The worst thing he did was re-use a dataset entirely, claiming it was a plot of something else, and left the exact same noise spurs and other anomalies.
Usually it's rare to find such blatant scientific fraud, but there was another recent fraud.
At least he's not moving the Lab's money into offshore shell companies to show earnings.
Sure, and at least he's also not killing people. But in the realm of science, what he's done is destroy the credibility that scientists strive for, and even NEED to be respected for. It's great that he's been caught, and hopefully it'll be a lesson to any up-and-coming experimentalists that no matter how much you believe in your theories, you have a committment to truth.
Maybe there should be some kind of hippocratic oath for scientists, that would be cool.
make world, not war
Like many physicists, I have spent a good chunk
of this morning reading the Beasley report on
this case. There is nothing about this guy or
his data that is not "troublesome", i.e. fake.
When your read that virtually every paper he
published is the result of scientific misconduct
it gets very hard to feel bad for the guy.
Instead I feel bad he is ruined at 32, not at 28.
Scientists at rival laboratories, however, had difficulty reproducing the results of Schon's work, thwarting a checks-and-balances process integral to the scientific method.
Um, no. If the "checks-and-balances process" were being thwarted, then it would have been circumvented or avoided somehow. This is an example of the process working as it's supposed to. You don't need a checks-and-balances system if everyone in the field is always going to be a good boy at all times. What happened here is that someone wasn't, and the scientific process caught him at it.
I would love it if these wire services would assign beats to reporters by taking into consideration what subjects they actually understand. They should also be fluent in the language in which they are writing, and display some comprehension of the words they're using.
And the brethren went away edified.
Its odd that they make a big thing out of finding the forgery though. What does that buy them? Why not say "Ouch!" fire him and move on?
It buys them them the sort of respect in the scientific community that being open about bugs and security flaws buys Debian or OpenBSD in the hacker community. Quietly sweeping this under the carpet would create among scientists the sort of sentiment MS and others recieve from hackers and admins when those companies hide or ignore security holes.
"...his work led to speculation by some peers that he could one day be nominated for a Nobel Prize, a high honor."
Glad they specified that. Otherwise, I might have thought they were referring to some other Nobel prize, like maybe the Gertrude P. Nobel Prize in Experimental Cosmetology.