Neal Patel/Flickr

This study just revealed why the peer-review process is in so much trouble

Only 1 in 5 scientists are doing the work.

BEC CREW
24 NOV 2016
 

In recent years, scientists have been warning us about a reproducibility crisis in science, which has seen many seminal papers - particularly in psychology - failing to hold up when an independent team tries to reproduce the results.

Peer-review also appears to be losing its lustre, with studies finding little evidence that it actually even works, and now a new investigation has revealed that just 20 percent of working scientists are responsible for as much as 94 percent of the work. With a staggering 63.4 million hours dedicated to peer-review in 2015, something’s gotta give.

 

"[W]e showed that a small portion of the scientific community is carrying a disproportionate load of the peer review," researchers from the Paris Descartes University in France conclude

"This inequality may be the root of a potentially unmanageable burden. These 'peer-review heroes' may be overworked, with risk of downgraded peer-review standards."

The team used a mathematical model to reveal for the first time how the overall demand for peer-review compares to the amount of scientists available to perform it. 

They wanted to figure out if the scientific community could handle the fact that journals have been proliferating at a rate of 3.5 percent every year - meaning the amount of time scientists need to dedicate to peer-review is only going to increase in the foreseeable future.

They looked at all the papers listed in MEDLINE - a database of all published biomedical and life sciences papers from around the world - between 1990 and 2015. 

They estimated how many papers each year would have undergone multiple rounds of review, and roughly how many hours this would take to perform, and compared this to the amount of researchers on a global scale who were eligible for peer-review work.

 

They assumed that the average proportion of rejected papers was 25 percent, and 90 percent of the peer-reviewed submissions went through a second round of peer review. They calculated that reviewers typically spend 4 to 5 hours reviewing a paper.

Looking back through the records, they found that in 1990, a total of 372,589 peer-reviewed papers were published. Over the next decade, that number increased to 475,670 in 2000 - a pretty modest increase, all things considered.

But fast-foward to 2010, and that figure almost doubled, with 867,882 papers published around the world. By 2015, scientists were delivering 1,134,686 peer-reviewed papers within just 12 months.

If you expected the results to reveal that there are far too many papers needing peer-review than scientists can manage, here’s the thing - the reality is the exact opposite. 

The researchers found that in 2015, despite all the reviews adding up to roughly 63.4 million hours of review work, we actually have more than enough scientists to meet the demand - it’s just that most of them aren’t chipping in.

"The study concluded about 1.8 million researchers were keeping the global peer review system afloat, with up to 4.6 million more playing no role," John Ross reports for The Australian. 

"On that basis, the academic community could ­manage up to 249 percent more reviews, depending on the ­modelling scenario."

The team found that in 2015, just 20 percent of scientists eligible for peer-reviewing were responsible for between 69 and 94 percent of the reviews, which means they’re not only overburdened - they’re overburdened for no reason.

"In 2015 ... around 6.4 million authors were listed on life-science research papers - giving journal editors a vast over-supply of people to ask," Brigitte Osterath reports for Nature.

"Even if editors only asked specific authors to be reviewers - for example, those in prestigious first and last positions in the author list - they still had 2.1 million candidates."

So the bad news is that the vast majority of scientists aren’t pulling their weight. In fact, the team estimated that 70 percent dedicated 1 percent or less of their research work-time to peer review last year, while 5 percent dedicated 13 percent or more of it. 

But the good news is that even though the entire peer-review process is being propped up by a few 'peer-review heroes', they say this isn’t as unsustainable as it sounds (just kinda crappy). 

"Our results support that the system is sustainable in terms of volume, but emphasises a considerable imbalance in the distribution of the peer-review effort across the scientific community," they report.

So what’s going on here? 

While the researchers didn’t go so far as to investigate why so many researchers are publishing more papers than they’re reviewing - although the pressure to publish could be a big contributor - it could be that researchers are losing faith in the peer-review process itself.

As Richard Smith, the former editor of the British Medical Journal (BMJ), told Julia Belluz at Vox, "We have little or no evidence that peer review 'works', but we have lots of evidence of its downside."

Some critics even go so far as to call the peer-review process itself "unscientific", because with the lack of research on how peer-reviewing is actually done, and a scarcity of available data on how it’s been carried out across the globe, we’re putting a lot of faith in a system we’re not even sure is working properly.

"We need rigorous studies to tell us the pros and cons of these [peer review] approaches today. Until then, any advertised advantages of new arrangements are unsupported assertions," Drummond Rennie from the University of California, San Francisco, and deputy editor of the The Journal of the American Medical Association, writes in Nature.

So what now? 

The French team says that while there's a big imbalance right now in terms of workload, and a lot of criticism about how the process is actually carried out, it doesn't mean we should scrap it altogether.

But we do need to peer-review the peer-review process.

"An evidence-based approach to study peer-review, combining computer modelling, experimental studies, and sharing of data from journals and publishers, should be encouraged," they conclude in the journal PLoS ONE. 

"Improvements in peer-review will come in response to evidence."

More From ScienceAlert

Pay what you want for this White Hat Hacker 2017 Bundle

Become an ethical hacker this holidays. 

2 days ago
The total mass of Earth's 'Technosphere' is 30 trillion tonnes
2 days ago