“Open research” is an umbrella term for a set of practices that make the inputs, process, and outputs of a study accessible to others rather than kept closed by default. It is broader than any single practice on its own: a project can publish open access without sharing its data, or share its data without preregistering its hypotheses. Open research is the general commitment that runs across all of these — and this guide covers what the term includes and what doing it actually looks like at each stage of a project. For CASRAI’s coverage of specific pieces of this umbrella — the UK’s sector-wide data policy, a named metascience initiative, self-archiving mechanics, discovery tools, and licensing choice — see the linked guides and terms throughout.
What “Open Research” Means
Open research is not one standard with a single certifying body. It is a cluster of related commitments — making data, code, methods, manuscripts, and review reports available for others to access, scrutinize, and reuse — that together aim to make the research record more verifiable and more useful to people who were not part of the original team. UNESCO’s Recommendation on Open Science, adopted by its General Conference in November 2021 with the support of all 193 member states, is the closest thing to an internationally agreed definition: it frames the practice around open access to publications, open research data aligned with the FAIR principles, open source software and hardware, open evaluation (including open peer review), and wider engagement with society and other knowledge systems. The UK’s Concordat on Open Research Data (2016) operationalizes one slice of this — data specifically — as a set of ten principles that UK funders, universities, and researchers commit to.
In practice, most researchers encounter “open research” as a funder or publisher expectation rather than a single checklist: a grant may require a data management plan, a journal may require a data availability statement, and a lab may adopt preregistration voluntarily. Practicing open research well means understanding which of these commitments actually applies to a given project and following through on each one deliberately, not treating “open” as an all-or-nothing label.
The Core Components of Open Research
Most definitions of open research, including UNESCO’s, converge on the same handful of practices. None of them is mandatory in every context, and a project can legitimately be open in some respects and restricted in others — for example, sharing code and materials openly while keeping sensitive data under controlled access.
Open Data
Making the data underlying a study available for others to access, and ideally to reuse, typically by depositing it in a recognized repository with a persistent identifier and machine-readable metadata. The FAIR Data Principles (Findable, Accessible, Interoperable, Reusable) are the widely cited standard for what makes shared data actually usable, as distinct from merely available — FAIR data can be under controlled access and still meet the principles, provided the access conditions are documented and machine-readable. Not all data can or should be fully open: human-subjects data, commercially sensitive data, and data governed by Indigenous data sovereignty frameworks routinely require restricted or controlled access instead, and a credible open-data commitment accounts for that rather than defaulting to full openness regardless of context.
Open Access Publishing
Making the resulting manuscript freely readable, rather than restricted behind a subscription paywall. This can happen through gold open access (publishing in a fully open-access journal or paying to make an individual article open), or green open access (depositing an accepted manuscript in a repository, typically after an embargo period). CASRAI’s Open Access Publishing guide covers the models and mandates in depth, and the Self-Archiving for Green Open Access guide covers the specific mechanics of the deposit route.
Open Code and Methods
Sharing the analysis code, software, lab protocols, or materials used to produce a result, not just the narrative description of the method in the paper. This is what makes a result computationally reproducible: someone else can rerun the actual analysis, not just read a summary of what was done. Frameworks like the ACM’s Artifact Review and Badging scheme formally evaluate whether shared code and materials are available, functional, and reusable, not just present.
Preregistration
Publicly registering a study’s hypotheses, design, and planned analysis before data collection (or before analysis, for existing data) — timestamping the plan so that later readers can distinguish confirmatory tests specified in advance from exploratory findings arrived at afterward. Preregistration is a distinct commitment from open data or open access; a project can preregister without making its eventual data open, and vice versa. CASRAI’s Preregistration of a Study Protocol guide covers how to do this and how it relates to the stronger, peer-reviewed Registered Report format.
Open Peer Review
Making some part of the peer review process visible — this can mean publishing reviewer reports alongside the article (with or without reviewer names attached), publishing reviews on preprints before journal submission, or simply disclosing that review occurred and who conducted it. Practice varies substantially by publisher and journal: some require it, most still make it optional or don’t offer it, and “open peer review” is used inconsistently enough across publishers that it’s worth checking a specific journal’s policy rather than assuming what the label covers.
How to Practice Open Research at Each Stage of a Project
Open research is easiest to apply consistently when it’s planned in at the start of a project rather than retrofitted after the fact. The steps below follow a typical project timeline.
1. Planning
- Check funder and institutional requirements first — many funders now require a data management plan as a condition of an award, and some require or strongly encourage preregistration for specific study types (see CASRAI’s Data Management Plan (DMP) entry).
- Decide what will be open and what will need restricted or controlled access, and document the reasoning — a blanket “everything open” or “everything closed” default is rarely right, and funders and reviewers increasingly expect a case-by-case justification, the same principle the UK Concordat builds in explicitly.
- If the design supports it, write and timestamp a preregistration before data collection begins.
- Choose a repository and a license in advance rather than at submission time — see CASRAI’s Open Licensing for Research guide for how license choice affects what others can actually do with shared data, code, or text.
2. Conducting the Research
- Keep code, scripts, and protocols in a version-controlled, well-documented state as you go, rather than cleaning them up only at the end — this is what makes eventual sharing feasible without a large separate effort.
- Record provenance: what produced each dataset or result, using what inputs and what software versions, so the eventual documentation doesn’t depend on memory reconstructed after the fact.
3. Analysis and Write-Up
- Distinguish confirmatory analyses specified in a preregistration from exploratory analyses added afterward, and label them as such in the manuscript — this is the actual point of preregistering, not the registration itself.
- Prepare a data availability statement describing where data and code live, under what access conditions, and how to obtain them.
- Deposit data and code in a recognized repository with a persistent identifier before or at submission, rather than promising to do so “on request.”
4. Publishing
- Choose a journal or platform whose open access route (gold, green, or a hybrid model) fits the funder mandate and budget in play — see CASRAI’s guide on Article Processing Charges (APCs) for Open Access if a fee applies.
- If the venue offers open peer review, decide in advance whether to opt in and understand what it commits you to (named reviews, published reports, or both).
- Consider posting a preprint to make findings public ahead of, or independent of, formal peer review — CASRAI’s guide to choosing a preprint server covers the criteria for picking one.
5. After Publication
- Make sure deposited data and code stay findable — register them with persistent identifiers and correct metadata so they surface through discovery tools; CASRAI’s guide to metadata search engines covers how tools like OpenAlex and CORE index open outputs.
- If the underlying repository or platform changes, keep persistent identifiers resolving correctly rather than letting links rot.
Open Research vs. Open Science vs. Open Access
These three terms overlap heavily and are often used interchangeably in casual usage, which causes real confusion:
- Open access is the narrowest of the three: it refers specifically to free-to-read (and sometimes free-to-reuse) publications. It says nothing on its own about data, code, or preregistration.
- Open research is the broader practice-level umbrella used in this guide: the set of things a researcher or team actually does — sharing data, code, preregistering, publishing openly, and so on — across a specific project.
- Open science is typically used as the broadest, most policy-level term, encompassing open research practices plus infrastructure, culture, and societal engagement — it’s the frame UNESCO’s Recommendation and initiatives like the Center for Open Science’s Open Science Framework (OSF) operate under. In everyday use on university and funder websites, “open science” and “open research” are frequently used as synonyms; where a distinction is drawn, “open research” more often refers to the practices a given project follows, and “open science” to the movement and infrastructure around them.
CASRAI’s SCORE (Systematizing Confidence in Open Research and Evidence) entry covers a specific DARPA-funded initiative that used replication and prediction markets to assess the credibility of social- and behavioral-science claims — a concrete example of open-research infrastructure applied to the separate problem of evaluating existing findings, rather than a general definition of the term.
Common Obstacles, and How to Manage Them
- “Open” is treated as all-or-nothing. It isn’t. A project can share code and materials openly while keeping human-subjects data under a controlled-access agreement, or preregister a confirmatory analysis while treating a related exploratory analysis as such rather than pretending it was planned. Partial, well-documented openness is more credible than an overstated blanket claim.
- Sharing gets left until submission. Cleaning up code and data for release is far cheaper when it happens continuously during the project than as a rushed step before a deadline. Planning it in from the start (see the Planning section above) is the practical fix.
- Restrictions aren’t documented. Funders and reviewers increasingly expect a stated, case-by-case reason for any restriction on openness, not just a default assumption that data or code will stay closed. Writing the justification down in the data management plan or availability statement is itself part of practicing open research, not a separate compliance task.
- Terminology gets treated as interchangeable when it matters. A funder mandate that says “open access” is not the same commitment as one that says “open data,” and neither implies preregistration. Reading what a specific mandate actually requires, rather than assuming “open research” covers all of it automatically, avoids compliance gaps.
Frequently Asked Questions
What does “open research” mean, in one sentence?
It means making the data, code, methods, and outputs of a research project accessible to others rather than closed by default, across some combination of open data, open access publishing, open code and methods, preregistration, and open peer review.
Is open research the same as open science?
The terms overlap and are often used interchangeably. Where a distinction is drawn, “open research” usually refers to the practices a specific project follows, while “open science” refers more broadly to the movement, infrastructure, and policy environment (such as UNESCO’s 2021 Recommendation) that those practices sit within.
Do I have to do all five components to count as practicing open research?
No. Open research is not an all-or-nothing certification. A project might share data and publish open access without preregistering, or preregister without an open-peer-review journal being available in its field. What matters is being deliberate and transparent about which commitments apply and following through on them, not maximizing every category regardless of fit.
Are funders now requiring open research practices?
Increasingly, yes, though requirements vary by funder and are usually specific rather than a blanket “be open” mandate — for example, a data management plan requirement, an open access mandate for publications within a set timeframe, or a data deposit requirement tied to a specific policy such as the UK’s Concordat on Open Research Data. Always check the specific funder and grant terms rather than assuming a general open-research expectation covers what a particular mandate requires.
What’s the difference between preregistration and a Registered Report?
Preregistration is a researcher unilaterally timestamping a study plan before data collection. A Registered Report is a stronger, journal-mediated version of the same idea, where the study protocol is peer-reviewed and provisionally accepted before data collection begins. See CASRAI’s Preregistration of a Study Protocol guide for the full distinction.
