Open Access as a Foundation for Reproducibility

Research reproducibility remains a concern in the scholarly community. A recent investigation, conducted over seven years, showed that researchers could reproduce the results of only half of the studies tested.

In September 2026, Horizon Europe launched the project REWIRE (Reproducibility Enablement for Widespread Implementation, Reuse and Evidence in the European Research Area). The initiative brings together organisations across Europe to encourage and support researchers, institutions, funders and publishers to adopt reproducible research practices.

While just one recent example of efforts to improve reproducibility, initiatives such as these further underscore the ongoing concern and efforts to address this issue.

Open Access (OA) publishing lays the groundwork for reproducibility, since research that cannot be widely accessed cannot be widely checked. Building on this foundation, broader open science practices such as open code, open data, and open methods extend the reach of OA by giving researchers the materials they need to verify and reproduce findings.

In this article, we discuss research reproducibility, introduce OA as a strong foundation for reproducible research, and provide guidance regarding open science practices authors can implement to ensure their research is transparent and reliable.

What is reproducibility?

Broadly, reproducible research is research that can be independently verified. This entails another scholar following the exact same steps as the original study and obtaining the exact same result.

Therefore, a crucial requirement for producing reproducible research is ensuring other scholars have unrestricted access to the detailed methods, data, code, and analysis conditions that were present in the original study.

Differences between reproducibility, replicability, and repeatability

At a glance, reproducing, replicating, and repeating can seem like overlapping terms. However, scholars  make important distinctions between them.

Jon Claerbout and Martin Karrenbach, scholars in the field of computational science, outlined the distinctions between reproducibility, replicability, and repeatability:

  • Reproducible research is when other scholars use the same dataset as the original study, with the same analysis, to achieve the same result.
  • Replicable research occurs when scholars conduct new analyses using a description of the dataset and consistently produce the same findings.
  • Repeatable research is when the original scholars use the initial dataset and analysis to repeatedly achieve the same result.

However, such definitions have not been universally accepted.

For example, the Association for Computing Machinery (ACM) provided an alternative definition whereby the meanings of reproducibility and replicability, as suggested by Claerbout and Karrenbach, were swapped. Scholars have argued this has resulted in a mixed consensus of what these terms actually mean.

For the purpose of this article, we understand reproducibility in line with Claerbout and Karrenbach, and more recently The National Academies of Sciences, Engineering, and Medicine (NASEM), which define reproducibility as:

Obtaining consistent results using the same input data; computational steps, methods, and code; and conditions of analysis.

Why is reproducibility important?

Scientific discovery has always built on what has come before. So, while scholars encounter increasingly complex questions, much of our scientific understanding rests on discoveries made decades and even centuries ago.

Therefore, reproducible research is crucial for providing essential verification that the findings of a study are accurate and, as a consequence, trustworthy.

Reproducible research also shows that the original results did not occur due to bias or chance, which in turn underpins the reliability of the conclusions.

Given that research forms the basis of our understanding of the world around us and has direct implications on advancements across health, technology, and public policy, being able to verify findings is essential for maintaining trust and reducing future research waste.

Challenges in maintaining reproducibility

Reproducibility rests on the ability to access, understand, and follow the exact processes of an initial study. However, the growing frequency in which scholars are unable to reproduce others’ research has contributed to rising concerns that the scholarly community is facing a reproducibility crisis.

A previous survey found that more than 70% of researchers found themselves unable to repeat other researchers’ work, demonstrating how undetailed, missing, or inaccessible methods can limit the extent to which existing research can be used as a foundation for scientific progress.

Incomplete methods sections

Methods sections have always been fundamental for explaining how a study’s data were collected and how its findings are supported by evidence.

However, across subscription-based publishing, methods sections have become increasingly short and less detailed due to restrictions on the number of words or pages that articles can have. Exceeding these limits can incur charges, diminishing the incentive for researchers to detail their methods in their entirety.

Furthermore, complex and computational studies often rely on vast datasets, but article length constraints can prevent authors from publishing all the data needed for their work to be reproducible.

Inaccessible data, code, or materials

To ensure research reproducibility, scholars must understand more than just the steps taken in the research process; they also need to know exactly what was included.

Raw data, complex datasets, computer code, research materials and more are needed to provide scholars with all the tools to reproduce a study.

Without these core elements, reproducibility is vastly hindered and independently verifying results can be challenging.

A lack of support

Importantly, reproducibility isn’t solely technical. In a study of researchers across the EU and UK, a lack of reproducibility was argued as less of an individual choice but as a result of insufficient support to deliver it.

Equally, this sentiment is mirrored across other studies that highlight ‘limited infrastructure, insufficient training’ and ‘technical challenges’ as factors hindering researchers from adopting reproducible practices.

This means that even if scholars endeavour to make their research reproducible, it might not be achievable without established support and clear guidance.

Open Access as a foundation for reproducibility

Open Access (OA) makes entire scholarly articles available to view, download, and reuse at no cost to the reader. By removing paywalls, OA ensures anybody with access to the internet can engage with scientific literature.

This is greatly beneficial for reproducibility because the first step is understanding the methodology; OA enables this as it removes any barriers to accessing articles and engaging with their content.

The inherently boosted visibility of OA research also allows for wider public scrutiny, which supports the integrity of research. Open licences such as CC BY also allow other scholars to reuse articles, tables, or figures in their work, provided they credit the original study.

Finally, OA publishers such as MDPI typically operate only online. This removes page restrictions and encourages authors to be as detailed as possible in their methods.

Open science holds the solution

While OA provides crucial access to research articles, published manuscripts can seldom assure reproducibility if methods are not rigorous, data are not clearly presented, and analytical code are left obscured.

Open science involves making the entire research process, from initial observation to discovery, freely available for all to access. Open science includes multiple branches, such as open methods and open code, which when published or deposited as individual outputs alongside an OA publication work to bolster reproducibility.

Improving the reproducibility of your research

By following various open science practices, scholars can make an active effort to ensure their research is reproducible by others.

Below, we explore some of the key practices to achieve this.

Publish methods as open protocols

You can enhance the reproducibility of your research by sharing your detailed methodology as open protocols. These are detailed, step-by-step accounts of experimental procedures taken throughout the research process and operate as an individual, citable research output.

They offer a comprehensive account of procedural steps, along with a detailed list of materials and reagents used, amongst other details. Publishing your methodology ensures your experimental workflows can be accessed and replicated.

Furthermore, when given Digital Object Identifiers (DOIs), your processes become citable, which benefits you as it acknowledges how your own methods have influenced the work of another.

Learn more in our article ‘Advancing Transparency and Reproducibility with Open Methods’.

Standardise materials with RRIDs

The Resource Identification Initiative was designed to improve the traceability, reproducibility, and reuse of specific materials reported in the Life Sciences literature.

It allocates unique identifiers called Research Resource Identifiers (RRIDs) to resources such as antibodies, cell lines, and organisms and other tools such as software. These identifiers can be inserted into a manuscript, taking the format of ‘RRID’, followed by the unique identifier.

They provide clarity into exactly what was used within a study, which is useful for scholars looking to reproduce studies, as small differences in resources can lead to greatly different results.

RRIDs work to eliminate ambiguity around the specific materials used and improve reproducibility by directing readers to curated registries or research articles where they are also featured.

Share your code and not just the data

The code used within a study is a valuable research output and making it openly accessible for others to view and inspect has become an important part of open science. Previous studies have shown that sharing code leads to reproducibility rates rising by nearly 40% compared with those that only require data sharing.

MDPI encourages open code practices such as organising research code, tracking changes with version control platforms such as GitHub, and depositing in a repository along with an open licence and instructions on how best to use it.

Learn more about open code here.

Follow FAIR data practices

The FAIR data principles were established to provide guidance on how scholars can ensure their research data are findable, accessible, interoperable, and reusable.

These principles emphasise standardisation and machine-readability to cope with the increase in volume, complexity, and creation speed of data.

In practice, this entails sharing your research data in a repository, making sure it is described clearly, and directing scholars to where it can be found and how it can be reused.

Learn more in our article ‘Understanding and Practicing FAIR Data’.

Registered Reports

Registered reports are a publishing format in which a study’s hypothesis, methodology, and analysis plan are subject to peer review before the research begins and before any data are collected.

Registered reports aim to reduce publication bias. They discourage researchers from changing their hypothesis after commencing a study or discarding negative data to suit a theory. They allow and encourage negative results to be published in a structured and robust way.

This is important for ensuring reproducibility: when only positive results are published, certain data may be excluded from the output. Then, other researchers may fail to obtain the same result when repeating the study.

MDPI introduced registered reports as an article type in 2020, enabling researchers to receive early feedback on their study rationale and methodology before data collection, thereby promoting transparency, reproducibility, and accountability.

Read MDPI’s guidelines for authors submitting a registered report.

MDPI’s commitment to open science and reproducibility

MDPI continues to lead the global transition towards OA and understands the immense value of ensuring research is reproducible.

All articles across MDPI’s portfolio of more than 500 OA journals are published under a CC BY licence, meaning they are immediately available to read, download, and reuse with no permission needed.

MDPI is committed to supporting open scientific exchange and encourages authors to comply with best practices for sharing and archiving research data by ensuring their data adheres to FAIR principles.

Moreover, MDPI journals ask that scholars submit the full experimental details of their work. All articles include a Data Availability statement that explains how to access data associated with the research. This is important for ensuring the reproducibility of the study as well as enhancing the reuse potential of datasets for further analysis.

We’re dedicated to giving you all the information you need to understand Open Access. Our article All You Need to Know About Open Access covers a range of topics.