Scientific progress depends on findings being discussed and eventually validated by the wider research community. It is vital that the findings of a study can be reproduced independently by others with a high degree of reliability when the study is performed again by other scientists. In research, this is called ‘reproducibility’, and it depends on specific study data and parameters being implemented by those looking to achieve the same results. This way, findings can be verified and real progress made across scientific disciplines.
In this article, we will explore the concept of reproducibility and why it matters, resolve the ‘reproducibility vs replicability’ confusion, and suggest how preprints support reproducibility in research.
What do we mean by ‘reproducibility’?
To understand what we mean by reproducibility, let’s look at an established definition of the term. The National Academies of Sciences, Engineering, and Medicine (NASEM) defines reproducibility in research as
‘obtaining consistent results using the same input data; computational steps, methods, and code; and conditions of analysis.’
A universal way of reproducing study results is necessary because of the complexity of study parameters. Indeed, study results are often produced by complex computational processes using large datasets. This means, in many studies, that a typical Methods section of a research paper cannot possibly cover all the information required to reproduce the results. Therefore, additional information related to data, code, models, and computational analysis is needed.
To make this information available to others, researchers can do several things. These include pre-registering study designs, sharing raw data and code, transparently documenting methodologies, and using persistent DOIs.
Reproducibility vs replicability
The ‘reproducibility vs replicability’ confusion is understandable. Both terms sound interchangeable, and when it comes down to it, they both describe distinct yet related concepts.
As mentioned, reproducibility is obtaining consistent results using standardised and transparent methods, data, code, and computational processes. In contrast, replicability is obtaining consistent results across studies addressing the same scientific issue, but with each study producing its own data.
Different yet interrelated, both reproducibility and replicability are important for maintaining scientific integrity.
Why reproducibility matters
Scientific studies aren’t conducted just for scientists to gain knowledge and develop their expertise in limited fields. Scientists may closely analyse certain elements and phenomena in the material world, but they are almost always looking towards society at large, considering how their findings can improve the wellbeing of people.
Indeed, scientific discovery is a key foundation of modern society. It helps us learn about the world and thus what can be done to solve issues relevant to our times. When we support its flourishing, we are putting our faith in the ability of science and the advancement of knowledge to change our society for the better.
But to enable science to function effectively, there has to be openness between researchers, journals, publishers, and institutions. Ensuring that open data and open methods are available to others is a more efficient way to validate and thus accelerate research progress and policy adoption.
For these reasons, the findings of studies take on increased importance because of their potential to directly impact the public. As a result, reproducibility becomes integral to building and maintaining trust between people and science.
Reproducibility issues in science
Despite its necessity, reproducibility has been (and still is) an issue for science.
A study by Nature reported that more than 70% of researchers had failed to reproduce other study findings, and more than 50% had failed to even reproduce their own results. These findings point to larger issues around study designs, reporting transparency, and publication bias.
Notably, however, these large issues point to one fundamental concern in science: transparency.
How preprints support reproducibility in research
Preprints are a type of manuscript that supports transparency, reuse, and early feedback in research dissemination. They are early versions of scholarly articles made freely available online on dedicated preprint servers. Preprints.org actively supports that these openly accessible research manuscripts are further supported by transparency- and reproducibility-adjacent practices.
Data accessibility and sharing
For data accessibility and sharing, Preprints.org supports:
- Open access: All uploaded preprints are posted under a Creative Commons (CC BY 4.0) license, making research and data freely accessible and reusable.
- Supplementary files: Authors are encouraged to upload supplementary documents alongside their manuscript. Here, authors can clearly document methodologies in full detail (and any study limitations).
- FAIR principles: Preprints.org aligns with global data governance frameworks such as FAIR (Findability, Accessibility, Interoperability, and Reusability), recommending authors to upload data to a recognized data repository before posting a preprint (such as one from Re3data).
- Versioning: Each version of a preprint is assigned a unique DOI, creating a transparent, trackable history of research evolution, enabling authors to easily update preprints with new data.
Whilst preprints are an enabler of reproducibility, they are certainly not a guarantee of it. Limitations such as variable methods and reporting quality as well as risk of citation of withdrawn or superseded versions still exist.
Despite the limitations, it’s the open framework that preprints actively fosters that’s important to focus on, rather than the possible limitations of individual works. The open-research agenda may need to be improved, but the principles are solid. It’s only the foundations that need supporting further.
Publishing with Preprints.org
Are you a researcher who wants to ensure their study findings are transparent and reproducible? At Preprints.org, we empower researchers to freely and instantly share their work with a global audience, helping you gain early feedback, boost visibility, and accelerate discovery. Join over 420,000 researchers advancing open science on our accessible, multidisciplinary platform. Ready to submit? Upload your preprint today and make your work quickly discoverable.
Just exploring? Browse over 120,000 preprints across disciplines and stay ahead of the latest research.
