Science is littered with zombie studies. Here’s how to stop their spread.

Many people think of science as complete and objective. But the truth is, science continues to evolve and is full of mistakes. Since 1980, more than 40,000 scientific publications have been retracted. They either contained errors, were based on outdated knowledge or were outright frauds.

Identifying these inaccuracies is how science is supposed to work. Finding and correcting publications — and keeping the scholarly record up to date — is part of the process. Yet these zombie publications continue to be cited and used, unwittingly, to support new arguments.

Why? Almost always it’s because nobody noticed they had been retracted.

Ensuring that scientists can build future research on solid foundations is essential, making it imperative to build a better system.

Just by citing a zombie publication, new research becomes infected: A single unreliable citation can threaten the reliability of the research that cites it, and that infection can cascade, spreading across hundreds of papers. A 2019 paper on childhood cancer, for example, cites 51 different retracted papers, making its research likely impossible to salvage.

For the scientific record to be a record of the best available knowledge, we need to take a knowledge maintenance perspective on the scholarly literature. We must keep a parallel record to reflect what we actually know. And we need to build on that knowledge, not on the errors and fraud.

Currently, it can be difficult to determine that a publication is retracted. And it’s even harder to be sure what is retracted — and what is labeled as retracted varies depending on where the information is coming from.

Even if science is self-correcting, the scientific record is not. Researchers are paid to publish — not to curate the literature and especially not to correct it. So, who will do this work?

For zombie-hunters, there is some good news: Tracking down citations to retracted publications has gotten easier. In September, Crossref and Retraction Watch partnered to open up the best data on what is retracted. Getting the database started was easier than finding a way to pay for it to be maintained. Yet the costs are minimal: $775,000 over five years to hand-curate unique information, such as why a publication was retracted.

Another big challenge is that researchers perceive retraction largely as a career risk. At best it is time-consuming: Correcting or retracting a publication can take two years, even when the authors agree. Though retraction can be beneficial, particularly when others discover retraction, the process can be painful.

But not promptly correcting the record can be career-altering: Consider Marc Tessier-Lavigne, who resigned from Stanford’s presidency the same day Science retracted two of his publications. Tessier-Lavigne didn’t persist in updating the literature, even when he knew his own publications had problems, and it cost him: for Tessier-Lavigne not retracting was a career risk.

Retraction itself is not the problem, but rather how it is handled. Another lesson is that science needs to prioritize lab working culture: We have to lionize failure. Slow science, living articles and reducing the pressure to publish are among the interventions that could help. We need a healthy, trustworthy ecosystem that rewards effort, not just results.

Early career researchers can find their careers boosted, not threatened, by self-retraction. And by being transparent about honest errors, they show their integrity. Nobel Prize winners retract papers, too. The correct response: Bravo!

Individuals and organizations that do the work of science must ensure that the work doesn’t end at publication. Sometimes, it is just the beginning.

Jodi Schneider is an associate professor of information sciences at the University of Illinois Urbana Champaign, NSF CAREER awardee and Public Voices Fellow of The OpEd Project. She leads the Alfred P. Sloan Foundation project “Reducing the Inadvertent Spread of Retracted Science.”