Open research: what is it, really?

What is open research? This is a question I keep hearing, from both those coming across the term for the first time and those who might be considered experts. The question might be stimulated by Research Excellence Framework 2029 developments, UK Reproducibility Network (UKRN) activities or any number of UK research culture initiatives. The question often prompts debates that won’t come to a definitive answer but do generate a lot of agreement.

So, I have taken it upon myself to digest years’ worth of pondering into an article that might help those who need a primer or a fresh perspective.

In my opinion, “open research” is a wastebasket term. In my academic discipline of taxonomy, fish would be considered a wastebasket group, so I don’t mean this in a pejorative sense. I mean that it’s a collection of “things” that are helpful for us to think about together but don’t necessarily have a natural or perfect definition. So, open research is a collection of best practices that enable us to meet principles of openness, transparency, replicability and accountability.

You might also come across the synonym “open science”. Open science is Unesco’s preferred term, but in the UK, “science” is a loaded word, which is typically exclusive of arts and humanities. We use “open research” to be clear that we are being as inclusive as Unesco and other international organisations intend to be.

• Open access: why and how to do it
• Researchers should harness the explorer’s spirit to find their career path
• Conferences, TED talks and social media: the ultimate tools for academic profile building

Examples of open research practices might include: pre-registrations (to ensure that findings are reported honestly without a posteriori bias); early sharing of findings and data, as preprints or staged micropublications (for example, NIHR threaded publication model); citizen science; open source software; open hardware; open access publishing or archiving; open and “findable, accessible, interoperable and reproducible” (FAIR) records of non-journal article outputs, including FAIR data; and open or collegiate peer review.

Depending on your sensibilities, you might place some of these practices under the banner of reproducibility, research integrity or research culture as well, or instead. Similarly, declarations of competing interests, contributor role definitions for the authorships (for example, CRediT), and open educational resources might be included in open research practices or might be considered adjacent to them.

All these options contribute to my sense that “open research” is a wastebasket term. What is clear is that open research is done throughout the research lifecycle, and I would refrain from saying it’s from “cradle to grave” because if open research is done properly, the materials can be replicated, reused and built upon on an ongoing basis.

It’s at this point that many people reading this might think: “Hang on, I’ve been doing some of this for years.” Yes, you probably have to some extent. What I would like current research culture initiatives to achieve is for researchers to start optimising existing practices so the outputs remain reproducible 20 years in the future and beyond. For this, we need to convince our institutions to invest in curation of data and digital preservation, and to identify approaches or systems that support best practice.

Some would go further and say that tools and data supporting open research should be non-proprietary and that institutions should invest only in open infrastructures that cannot be bought and sold and have community governance (according to, for example, POSI principles). Some might also extend this to sources for meta-research (research on research) and institutional decision-making. In part, the lack of community governance, transparency, rigour and “measuring what matters” (see Inorms’ “Rethinking the Rankings”) was what drove Utrecht University to withdraw from the Times Higher Education World University Rankings. Meanwhile, other institutions are signing declarations to say they are more than their rank, and the Leiden Ranking has transitioned to using only open data in the interest of transparency.

It’s inevitable that despite my opinion indicating that a definition of open research is subjective and which practices are included is personal, the problems created by transparency and reproducibility in research are pervasive, and so many organisations have attempted to define and redefine open research. We shouldn’t allow these definitions and their inclusiveness or exclusiveness to become barriers to the best practice they are intended to promote. Individuals should do what works for them, enhances their research and helps build trust in that research, according to discipline-accepted standards and their audiences.

For me, it’s important that all research, whether it’s exploratory or explanatory, heuristic or analytic, follows these core principles and practices at the very least:

1. Researchers should open up their projects are early as possible. This can foster trust, and potentially leads to collaborations and gains in priority over work or an idea, protecting intellectual property (IP) in plain sight. It often seems counter-intuitive, but it isn’t at all. By opening your work up to scrutiny and collaboration, you take priority over those ideas in the public domain. Also, if corporate espionage is suspected but the material is hidden on a hard drive, it would be difficult to prove that IP had been accessed or stolen.
2. It’s unlikely that a research paper or review article is truly the research output, but given the effort that has gone into creating open access platforms over the past 20 years, making articles open access is relatively painless and should be practised.
3. Consider what the research output actually is and think about how that can be made more reproducible. For example, through open source licences for software or removing barriers to the reproduction of hardware and making it clear what users can do through high-quality publicly available and machine-readable metadata.
4. Data must be FAIR and “open as possible and closed as necessary” or the research is not trustworthy.
5. Publish results that weren’t expected or those that might be considered negative or a failure. This will not only advance research as much as the “success stories” but will prevent resources being wasted in the future. Scholarly journals in a particular discipline might not accept these outputs, so if that is the case, there is an opportunity for communities to come together to agree appropriate existing platforms or to develop them. This is potentially a function that scholarly societies could fulfil as a value proposition for membership.

These approaches will work to establish trust in research, grow an individual’s profile and foster collaborations and future opportunities, to name a few benefits. Your work may benefit from being built upon long into the future, meaning that you will leave a richer legacy, and the public will benefit from faster development of technology, policy and practice.

After reading this article, please just take five minutes to think about what you could do that you’re not doing now or how you could optimise what you are doing in line with open research principles, and how you could be a better advocate for open research practices.

For the purpose of open access, a Creative Commons attribution (CC BY) licence has been applied to the submitted version of this manuscript.

Steven U. Vidovic is head of open research and publication practice at the University of Southampton.