Open Science

Open science is the umbrella term for a set of movements and practices aimed at making the entire research process — not just the final published article — openly accessible, transparent, and reusable. It is broader than open access: a paper can be free to read while the data, code, and methods behind it remain closed, which by the definition below is open access without being open science.

UNESCO’s four pillars of open science

The most widely cited normative reference is UNESCO’s Recommendation on Open Science, adopted by its General Conference in November 2021 — the first international instrument to set out a common definition, shared values, and areas of action for open science policy. It defines open science as “an inclusive construct combining various movements and practices aiming to make multilingual scientific knowledge openly available, accessible and reusable,” and organizes it into four pillars:

• Open scientific knowledge — scientific publications, research data, educational resources, and open-source software and hardware, made openly accessible and, for data specifically, aligned with the FAIR principles. On CASRAI: open access, open data, open code, open materials, and CASRAI’s step-by-step guide to making a dataset FAIR.
• Open science infrastructures — the non-commercial, interoperable, sustainably governed repositories, platforms, and services that openness actually depends on. On CASRAI: the Principles of Open Scholarly Infrastructure (POSI), the European Open Science Cloud (EOSC), the African Open Science Platform, and repository trust marks like CoreTrustSeal.
• Open engagement of societal actors — crowdsourcing, citizen and participatory science, and science communication that opens the process of knowledge creation and evaluation to people outside the traditional research community. On CASRAI: citizen science and the European Citizen Science Association (ECSA).
• Open dialogue with other knowledge systems — recognizing Indigenous, traditional, and local knowledge systems on their own terms, rather than folding them into Western open-data norms by default. On CASRAI: the CARE Principles for Indigenous Data Governance and the Global Indigenous Data Alliance (GIDA).

Open peer review and preregistration

Two further practices are usually discussed under the open-science umbrella even though UNESCO folds them into the pillars above rather than naming them separately. Open peer review — publishing reviewer identities, reports, or both alongside a paper — stands in contrast to double-anonymized peer review, where neither party’s identity is disclosed. Preregistration of a study’s hypotheses and analysis plan before data collection, and the Transparency and Openness Promotion (TOP) Guidelines that many journals now use to declare their openness requirements, are among the most concrete, checkable open-science practices a research office will encounter.

Why it matters: reproducibility and public accountability

Open science gained institutional momentum for two connected reasons. The first is a documented reproducibility problem: the Open Science Collaboration’s 2015 Estimating the Reproducibility of Psychological Science (published in Science) attempted to replicate 100 published psychology studies and found that while 97% of the original studies had reported statistically significant results, only 36% of the replications did — a gap widely cited as evidence that closed, undocumented research processes make error and selective reporting harder to catch. Frameworks such as TOP and the broader set of discipline-specific reproducibility frameworks (ARRIVE, CONSORT, PRISMA) exist largely as a direct response.

The second is public accountability for publicly funded research. Funders that pay for research increasingly require the outputs to be usable by others, not just by the funded team. In the United States, the White House Office of Science and Technology Policy’s August 2022 “Nelson Memo” directed federal research agencies to require immediate public access to funded publications and data, underpinning both the NIH Data Management and Sharing Policy (in effect since 25 January 2023) and NIH’s move to zero-embargo public access for accepted manuscripts. In Europe, funder coalition cOAlition S’s Plan S set comparable open-access requirements for participating funders — see CASRAI’s coverage of Plan S and the shift toward Diamond OA. Individual institutions and funders outside these two examples set their own open-science mandates, but the underlying rationale is consistent across them: research paid for by public or philanthropic funds should be checkable and reusable by the public that funded it, not only by the original team.

How this maps onto a research office’s day-to-day work

In practice, an institution’s open-science obligations arrive piecemeal rather than as a single named policy: a Data Management Plan at the proposal stage, a open access mandate at publication, a reproducible research practices checklist for computational work, and increasingly explicit software- and materials-sharing expectations for anything a paper’s conclusions depend on. CASRAI’s Research Data Management hub collects the data-specific side of this in more depth; this entry is the umbrella definition that ties the individual practices together.

References

• UNESCO, Recommendation on Open Science (adopted 23 November 2021).
• Open Science Collaboration, ‘Estimating the reproducibility of psychological science’, Science, 2015 (doi:10.1126/science.aac4716).
• White House Office of Science and Technology Policy, ‘Ensuring Free, Immediate, and Equitable Access to Federally Funded Research’ (the “Nelson Memo”), 25 August 2022.
• cOAlition S, Plan S.