Stand up for science! We saw these rallies across the United States on March 7th to protest cuts in National Institutes of Health (NIH) funding. But should we blindly believe the science? No — and how do I know?
I was a scientist in academia for over 20 years. I earned a Ph.D. from Rutgers University and worked at the Brigham and Women’s Hospital and Massachusetts Institute of Technology.
My dreams of what it was like to be a scientist were shattered early on. I had envisioned a world with me in the basement with the likes of Marie Curie making exciting discoveries. Instead, I encountered a world where career advancement trumped a pursuit of the truth, a world where “publish or perish” ruled the day. No paper, no money for your lab. Period. And the motherlode for money is the NIH.
The system invites bad science and cheating. I could write endlessly about the cheating I observed. One common way is to simply toss out mice that didn’t “cooperate.” Namely, you had a hypothesis, did the experiment, but didn’t get a statistical difference. But, you think, if I toss out the data from two mice in Group A and one in Group B, I’d have a difference. Yay! Let’s publish!
You might be asking, “Why not publish the negative results — you had a good hypothesis, but it didn’t work out?” Because — prepare for the insanity –you cannot publish “negative data.” Bad for you (no paper) and bad for Dr. Smith across the pond who is currently wasting time and money doing the exact same experiment (yes, you weren’t the only one with that hypothesis).
Then there’s the interaction that wasn’t — published in a prestigious journal. The “discovery?” Molecules A and B bound together! It was a big discovery in cancer research. What the journal didn’t know was A and B never bound directly. Instead, they both bound to Molecule C, giving the appearance that A and B bound directly. This was easily proven by dissolving C. The kicker? The researcher did this experiment before submitting the paper, but mum’s the word to the journal. And wouldn’t you know — that paper resulted in a big grant from the NIH.
And then there’s the “smart” gene. A gene in mouse brains was altered to see if it affected intelligence. Success! It made them smarter! The study got so much attention that David Letterman included it in his act. But then…
Oops. No one could reproduce the results, and a close look at the raw data by a clever new researcher revealed the truth — the famous paper was bogus.
Not everyone cheats, of course, but cheating is common. I doubt many people start out with intention of being dishonest. The problem is the whole “publish or perish” system and the dependence on NIH funding. It needs to change.
As does the questionable models. Take mice, for example. C57BL/6 are the most commonly used mice in research ranging from cancer to neuroscience. These mice have been inbred (brother-sister mating) for decades in order to eliminate genetic variability. The result was “fixed” double dominant or double recessive genes, as well “fixed” mutations.
Rama
Jesus Christ is King
Comments are closed.