Double Checking The Scientific Process

Dec 11, 2020

As the FDA considers the approval of a COVID-19 vaccine this week, figures around the U.S., including here in Wyoming, are calling into question the vaccine and the approval process it's going through.

Professors from the University of Wyoming and CU Boulder said science isn't perfect. However, it is full of checks and balances that are in place to stop errors from going forward.

UW molecular biology professor David Fay said scientists are self-policing and they hold each other accountable.

"You can depend on scientists to call each other out simply because we enjoy being right," he said. "Especially if someone bigger than us is wrong."

Fay said there are examples of research that went wrong or were faked, but he said they usually get caught.

"There's a famous case in the stem cell field that happened in the last ten years where a scientist was making up stuff," he said. "Right away, I went to a meeting in the wake of this claim and people were talking about that and no one at the meeting believed it. Well within two months it had been retracted."

UW astronomy professor Danny Dale said academic journals help filter out bad science. He said experts review each research paper that is submitted—without knowing the author's identity.

"You have to do an iterative process with the referee and the journal where they give you feedback and you have to do your best to address their concerns," Dale said.

He said some research isn't accepted into journals. Even if a paper is accepted, that doesn't mean the critiques end there.

Dale said he published a paper that measured the brightness of galaxies in the nearby universe. That made up a data set that he wanted other people to use for future research. But when a colleague went to use the data, they found an error.

"When you're observing from the surface of the earth you have to look through the atmosphere," Dale said. "And I provided the brightnesses of these galaxies, obviously they're beyond the atmosphere, and I corrected for the effects of the atmosphere, but I did it twice unfortunately."

Dale fixed the data in his paper, but he said it's a great example of how the scientific community checks itself. His correction is not alone: there are whistleblowing websites, like PubPeer, that regularly check the results in published research papers.

"Science isn't saying trust any individual scientist or trust any individual paper or result in that paper because we're fallible and some people are actually dishonest, right, it's a small percentage but they happen. But you can trust the process."

Molecular biologist Fay said sometimes it's not even the data that's wrong, it's what scientists do with it. He said another group of scientists found a new explanation for some of his results.

"They cited tons of our data. Right, they just went through every one of our papers and they flipped the interpretation," said Fay. "They kept the data the same, but they kind of flipped everything and it was like looking at the negative or something of a photograph. But suddenly everything made sense."

Fay said he appreciates what that group did. He said science is capitalistic, and the good stuff eventually makes it to the top.

"Science isn't saying trust any individual scientist or trust any individual paper or result in that paper because we're fallible and some people are actually dishonest, right, it's a small percentage but they happen," he said. "But you can trust the process."

One topic that frequently comes up in the COVID-19 discussion is how health officials have changed their recommendation since last spring. Tin Tin Su, a molecular, cellular and developmental biology professor at CU Boulder, said that's standard procedure.

"To me that's not so much a mistake as people not being aware of the scientific process, which is that you have to change your conclusions as you get more and more data," she said.

Su said that means keeping up with new results.

"Educate yourself to what the latest is, because [the] CDC can say something very different next week," she said. "It doesn't mean they were lying before."

When it comes to a COVID-19 vaccine, Su said the gap between a company's safety approval and the FDA's approval has a reason: reproducibility.

"What the public doesn't seem to realize, and it's because it's not communicated to them, [is] that they're not sitting idle," she said. "They're [the FDA] not saying 'oh wait let me check my calendar and I have an hour in the afternoon on the 10th and that's sort of why we're meeting'. No, they're going through the data."

Su said validating scientific results by redoing experiments and reexamining data is essential, and that's exactly what the FDA is doing with the vaccine data.

Mistakes in science are caught at multiple stages: by the original researcher, by others in their lab, by the reviewers of their paper, and by the scientific community after publication. It's not wrong to take a critical view of scientific results, but that's also what scientists themselves do, over and over again. They said that's the whole point of the scientific process.