Article Lead Image

Scientist’s reputation may not crumble when research is scrutinized, study says

Can’t replicate your results? No problem, study says.


Cynthia McKelvey

Internet Culture

No one likes being proven wrong, especially not scientists. A new open-access study in PLOS One found that researchers think having their research proven wrong—producing irreproducible results—will harm their career to a far greater degree than it probably would in reality.

The study’s authors asked researchers to estimate how much it would hurt their career if their own results were proven wrong, versus a colleague’s. Researchers consistently rated the damage to their own careers as higher than if the same thing had happened to a colleague. According to the study’s authors, this likely reflects both a fear of being proven wrong while also an understanding that sometimes scientists are simply wrong.

Scientific studies need to be reproducible in order to determine their validity. And researchers aren’t always forthcoming with their data, perhaps out of fear that their results won’t be replicated. Yet more researchers are finding that a lot of studies can’t be replicated–casting doubt on ideas and theories that have been accepted as true.

Part of the problem is that some researchers, intentionally or not, are hacking their own data to make the results look sexier. This sort of data hacking is shockingly easy to do, which is why it’s often unintentional.

But the current study should give confidence to researchers.

“Admitting that you were wrong about a finding gives the impression that you are a humble person who is interested in the pursuit of knowledge,” study author Adam Fetterman told Retraction Watch. “Being wrong is a part of science and no one should blame anyone for being wrong. We are all learning as we go and no one is perfect, and very few researchers are sinister in their intentions.”

The study is limited, the researchers wrote. They recruited participants over Twitter, where there is an ongoing discussion about reproducibility problems in science. So the participants may already have a stake in improving research quality, the authors said.

Yet the researchers are confident that their research is, well, reproducible. Anecdotally, they’ve found that researchers who admit they’ve been proven wrong (and aren’t also proven of research fraud) are not penalized. Retraction Watch agreed.

“These findings align with what we’ve seen—and pushed for—in the past: Transparency pays,” Retraction Watch author Alison McCook wrote. “For instance, a 2013 paper showed that scientists who come forward to retract their own research do not face a so-called citation penalty, or a drop in citations following the event.”

Time will tell if this research can be replicated—hopefully it can.

Illustration by Jason Reed

The Daily Dot