The Early-Bird Papers: Who Really Uses Preprints? The Global Divide in Open Science.

Before scientific papers get published in a journal, they undergo peer review: Independent experts check the science and can require improvements. This is a long process that can last over a year. The manuscript, which has not yet gone through peer review, is called the “preprint.”

Recently, publishing such a preprint on the internet seems to have become quite popular, but it isn’t a new practice. The US and the Soviet Union had already used preprint paper copies in the 1960s to spread information quicker. In the 1990s, arXiv made it easy for physicists and mathematicians to share early versions of their work online. Later, other fields created their own preprint servers.

But wait—isn’t that a bad thing? This is non-peer-reviewed science, so it is bad quality …or is it?

Some research suggests preprints aren’t that different from final papers. They’re still early versions, but they come with perks: anyone can read them for free, they often get lots of attention, and scientists can list them when applying for grants.

To understand how preprinting is evolving across disciplines and different places in the world, we can look at—a preprint. Narmin Rzayeva, Stephen Pinfield and Ludo Waltman from the Netherlands and the UK are examining this. Their study is not yet published in a peer-reviewed journal, but the preprint manuscript can already be read.

The authors wanted to understand four big questions: how preprinting has changed over time, how different subjects (like physics, medicine, or psychology) use it, how different parts of the world use it, and how subjects and places interact, for example, which regions lead in which subjects.

Here is what the authors did:

To explore this, the authors first checked what earlier research had found, then turned to huge databases like Crossref that store information about millions of scientific papers and preprints. These databases don’t hold the full studies but their “metadata,” such as titles, authors, dates, and links. Using this information, the researchers looked at every paper published from 1991 to 2023 and tried to match each one to a preprint…like pairing socks in a giant laundry basket. They used special IDs, titles, and other details to link the pairs, finding 2.2 million matches, and then checked a sample by hand to make sure the links were correct. Next, they sorted the papers into disciplines (like math, biology, or engineering) and into countries based on where the authors worked. If a paper had authors from several countries, it counted once for each country involved. They then grouped countries into larger regions, such as North America, Western Europe, or Eastern Asia to compare how these regions differ. Finally, they looked at simple stats like percentages and trends over time, such as what share of papers in each field or region had a preprint in a given year.

So, what did they find?

Generally, preprinting is not common practice. Out of 104.6 million peer-reviewed articles only 2.2 million were linked to a preprint in their database. Interestingly, 1.9 million preprints had no link to a peer-reviewed article. Indeed, in their review of previous studies, the authors describe that automatic linking on preprint servers has some technical issues and often fails.

Even if the percentage of peer-reviewed articles with a preprint seems low, it is currently at an all-time high. This becomes clear when we look at different disciplines and what percentage of their peer-reviewed articles are linked with a preprint.

Preprints have been rising across all fields. Physical Sciences and Technology were early adopters, and they still lead; about 70% of all preprints in the study came from math and physics. Other fields began increasing their preprints around 2016 as more servers appeared. The sharp rise in Life and Medical Sciences in 2019 was driven by COVID-19, when scientists needed to share results faster than peer review would allow. In humanities (subjects that study human culture, ideas, and history), however, preprinting is still quite rare.

Preprinting isn’t equally common everywhere. In Europe, about 10% of peer-reviewed papers now have a preprint, but in Africa it’s only around 5%. Asia is also low, though East Asia does a bit better than the rest. These gaps matter because places that preprint more may get more attention… and more funding.

The researchers combined their data on disciplines and regions for a clearer picture. This highlights how big the differences are – in 2021/22, almost 50% of all Physics articles from Northern, Western and Southern Europe had a preprint. Curiously, while Africa’s numbers are generally low, they surpass Northern and Western Europe in Medical and Health science.

So, what can this study tell us?

This study uses huge datasets and careful checks to show that preprinting is rising across the world, though at very different speeds in different fields and regions. But the authors also admit their numbers aren’t perfect—they may have missed preprints (especially in economics), counted some non-peer-reviewed pieces as peer-reviewed (especially in the humanities), and lacked full links across databases. The study also doesn’t dig deeply into why these patterns appear: why do poorer regions preprint less, even though preprints can boost visibility and funding, and why does Africa outperform Europe in medical and health sciences? One author is also the president of ASAPbio, an organization that promotes preprinting, which is worth noting, though the methods are transparent and the study received no funding. And of course, this is itself a preprint, so the findings are still preliminary.

Still, it offers a clear message: if we learn to interpret preprints well, they can speed up science, widen access, and help researchers everywhere, showing where we need to act if we want these benefits to be shared more fairly.