Results from an alarming number of studies have been about as easily repeatable as losing your virginity. This raises some questions about experimental design, what passes for statistically significant results, and why so many published studies display all the finer qualities of dogshit.
The whole system was so deeply fucked long before some asshole attempted to bullshit his way to a Nobel Prize.
Today’s Moment of Science… The Jan Hendrik Schön Scandal.
Science exists in a continuous cycle of confidently knowing stuff followed by being a tad super fucking wrong. New ideas often come at the expense of older, less advanced ones that we’re comfortable with. You’d think it would help, but evidence isn’t always enough when people are met with a paradigm shift. The big bang, evolution, and wearing a fucking mask during a pandemic are ideas that still give people big runny dumps.
Conversely, sometimes we’re happily convinced by the thinnest of evidence when we want something tp be true so fucking badly. After all, a zippy new breed of transistors wouldn’t be a paradigm shift, no. It would just be clinging onto Moore’s Law for a few more minutes, right?
On the heels of wrapping up his PhD in physics from the University of Konstanz in Germany in 1997, Jan Hendrik Schön was hired by Bell Labs. His research focused on using organic molecules to make a new class of molecular transistors. There had been attempts, but those equaled in number to the failures.
Then Schön managed to make the molecules fucking dance. When a thin layer of organic dye molecules were hit with a current, they “behaved as a transistor.” The potential to work far faster than silicon based electronics was seemingly within grasp.
So, uh. Do we start with the irreproducible results, the absurd number of papers he was churning out, or the (fucking alleged or whatever) repurposing of data?
He was listed as an author on a new paper every eight days. And true, depending on the type of research, being listed as one of the authors on a paper doesn’t necessarily mean you were in charge. But from 1998 to 2002, he was an author on over 100 papers, listed as the lead author on the lion’s share of them.
The depth of the fuckery hurts. It’s typical for researchers to put out two or three papers a year. Maybe four if they’re annoyingly motivated. This was almost one a week.
Some colleagues were suspicious. Then, the anomalies started showing up. They spotted graphs used in more than one article. A curve in one study was constructed not from data, but from a mathematical function. When asked for his original data, he said he deleted it because they were low on memory.
The fucking audacity to claim Bell Labs was low on memory.
Nobody could reproduce his results. Hell, after people started pointing out the problems with his work, even Schön couldn’t make the molecules break into a jig again. Fancy that. There are probably kids who went to grad school to get a PhD carrying on this work. Know what they got instead? Screwed.
It hurts thinking how many of his papers were published in top notch publications like Science and Nature. An editor at Nature commented that “referees had questions over the interpretation of the data. But they were nearly unanimous on the importance of the findings.” I’m not sure how important findings can be when the data is utter shit.
Everyone wanted to publish the kid that was so certain to win a Nobel Prize. Then, nobody did.
Jan Hendrik Schön was fired from Bell Labs. He lost the job offer he had lined up at the Max Planck Institute. In a nearly unheard of move, his PhD was revoked.
His current whereabouts are unknown.
This has been your Moment of Science, always willing to show my work.
To get the MOS delivered to your inbox every weekday with an appropriate smattering of f-bombs for light subjects like academic fraud, head to patreon.com/scibabe.
Join the discussion!