Faulty scientific articles are not just a nuisance in journals. They also misdirect policies, waste money from taxpayers, and put lives at risk. A broken system of corrections is the reason why errors in high-quality research continue to persist. Take a look at our recent experience.
The year 2025 is March Communications Earth & Environment Published A paper claims that oil palm certification leads to land expansion and lowers production. The study misread the satellite data, interpreting temporary decreases in production during replanting to be a loss of land. After correction, data shows no decrease in efficiency.
It is not supported by the paper that certifies increase land demand. Our request for retraction, despite this, was denied. We were then asked to provide a text rebuttal, which is still being reviewed nearly a full year after we submitted it.
Land-cover classifications compared. Our reanalysis shows that the original study, (bottom), misidentified palm oil production as having destroyed bare (purple) land.
A 2023 is another example. Nature Paper estimating the deforestation caused by rubber plantations. Study’s errors in sampling overstated the footprint of rubber deforestation. After almost two years, our correction was finally published behind a paid-for paywall. By then the study’s flawed findings had already been cited by 98 people and used to shape multiple policy reports.
The papers were peer reviewed in two leading journals. This shows that top systems can produce errors just as easily as they do insights.
What makes errors so difficult to correct?
The “correction machine” in academia is failing. Retractions or errata are rarely given priority by journals, and those who point out errors often receive no encouragement.
Academics are more interested in novelty than accuracy. New papers are more important than careful revisions. Post-publication critique counts for little. Admitting error risks reputation. You may lose your reputation if you point out other people’s mistakes.
It is not surprising that even flagged errors can accumulate in this environment. The retractions of errors are slow and rare. One Nature After nearly 4,500 citations, the paper was withdrawn 22 years after it first appeared.
Delays have costs. Faulty data has led to bad clinical decisions in medicine. For example, the now-retracted Lancet 2020 hydroxychloroquine trial that temporarily halted COVID-19 global trials.
Satellite-based estimates of deforestation in conservation are often very different, which confuses policymakers and undermines their trust. Multiple studies produced very diverse pictures of the forest losses, which led to conflicting claims and unclear priorities.
Click here to read more
Scientific publishing is shaped by evolutionary pressures, both for the better and the worse.
What brought us here
Profit and pressure are the main reasons for this.
- Academic publishing commercialised
Others have also highlighted the problem of a system in which scientists are often paid by the public to conduct research and review articles for free. Then, their institutions must pay an exorbitant fee to gain access, while private firms pocket the profits.
Robert Maxwell is responsible for this dysfunction. In the 1960s, he turned Pergamon Press in to a perpetual financing machine. Maxwell created a business model which commodified prestige in academia and the vanity of researchers. This led to the current commercial giants.
Maxwell’s model remains successful. Springer Nature’s profit margins were around 28 percent on revenues of nearly EUR2 billion ($2.3 billion). Elsevier, Wiley and other publishers have even higher margins.
Wiley is one of the world’s largest publishing companies with high profit margins.
shutterstock
Academic publishing has become one of the most profitable industries in terms of output per input. However, the profitability is based on unpaid labour (peer review alone takes up more than 100 millions hours per year) and restricted access for the public to the results.
Click here to read more
The academic publishing industry is worth billions of dollars. Science is not always benefited by it
- Inequality of access to peer review
The gatekeeper for scientific integrity, peer review, has been unable to keep up with the demand. As nations increase their output of scientific research, the number of submissions increases. AI tools also help to produce more credible submissions. Nature Its own recent description of a “peer review crisis”.
In 2023, the number of retractions surpassed 10,000 and continues to increase. It is possible that this does not indicate a self-correction system in action, but rather deteriorating quality control.
Paywalls and APCs (open access charges) prevent many researchers from catching flaws. Nature’s APCs have now reached EUR10,690 ($12,690). This cost effectively prevents many low-income nations from being able to publish, correct or access published works.
Publishers can earn double income: researchers pay for publishing. The readers pay for reading.
In theory, science is a self-correcting system. A system that prioritizes profit and status will only correct when necessary – slowly.
Reform is needed
It is not about being correct, but finding out where science has gone wrong and fixing it. The systemic reform of science must redefine prompt correction not as an indicator of failure, but rather a mark of integrity.
Rapid, collaborative scrutiny is already possible with open correction platforms, data sharing, and AI assisted review tools. The only thing missing is the incentive and courage to change the norm.
Publishers can pay for open corrections if they profit from errors that are hidden behind a paywall. Our corrections can be counted by institutions, funders and other organizations that can track our papers.
Open Access Publishing ensures everyone can at least access the correct science.
shutterstock
Journals must make their corrections visible and citeable. They should also expand the “Diamond Open Access”. More scrutiny is possible with wider access, and fixes are made faster.
Funders and institutions should support post-publication validation, while researchers should prefer publishers who value accuracy over hysteria.
Encourage your university to become a member of cOAlitionS to promote faster, fairer corrections. Retraction Watch can also help readers.
What’s needed is the willingness to make use of these tools. Silence is not the answer to mistakes. The strength of science is not that it never makes a mistake, but how openly and effectively it can correct itself.
