Decimal point errors, switching the numerator and denominator in fractions, or even just plain typos—scientists aren’t immune to these mistakes when publishing their research. Oklahoma scientist Jonathan Wren is trying to fix that.
When Wren and Constantin Georgescu, researchers at the Oklahoma Medical Research Foundation, conducted a study published in the journal Bioinformatics last year, they found that the math in many medical research papers didn’t add up.
The researchers extracted 486,318 statistical results from the abstracts of 196,278 published papers on the database MEDLINE. When they recalculated those results using the data in the papers, they found many discrepancies between the published results of those studies and the recalculated values.
“If you give humans tasks, we make mistakes. It’s just that simple. We’re error-prone creatures. It just happens,” Wren told KGOU in a phone interview.
Some of the errors were due to researchers exaggerating their results to appear more significant. Others, like misplaced decimal points, were easily identified and categorized. Wren said he couldn’t know the cause of all of the discrepancies, but he could help researchers find them more easily.
Wren and his colleagues at the Oklahoma Medical Research Foundation are creating software that will search for these mistakes and more. The program isn’t finished yet, but Wren wants to put it online and charge institutions and universities a subscription fee so their scientists can use it for free.
“You don’t want there to be errors in the literature, and you want to be able to find them and correct them. You want to be able to know if there’s problems with the study,” he said.
Wren says the software will help catch errors of commission — information that is incorrect — as well as errors of omission — information that is left out, like a catalog number for an antibody used in a study that could help someone reproduce the results of the research. However, the software won’t be able to catch errors in logic or interpret images.
“That’s going to remain the realm of humans for a very long time to come,” Wren said.
Wren says the increasing number and length of research papers is placing a heavier burden on the process of peer reviewing, in which academics screen research for publication in journals.
“When someone asks us to review a paper for a journal, we do it free, we do it on our own time and we do it because we’re interested in helping out,” he said.
Wren hopes his project will ease that burden and increase the quality of the peer review process.
“We want ways of being able to judge the studies that come out and identify the problems in them,” he said. “It’s better to find these problems up front than after the fact.”
As a community-supported news organization, KGOU relies on contributions from readers and listeners to fulfill its mission of public service to Oklahoma and beyond. Donate online, or by contacting our Membership department.