Most countries build research integrity frameworks the way they build fire codes: after something burns. Singapore did it backwards. It convened the world's researchers first, drafted four sentences in 2010, and then spent the next decade discovering how hard those sentences are to enforce. That inversion is the real story behind Singapore's national research integrity framework.
The country's most important export in research integrity is not a law. It is the Singapore Statement on Research Integrity, adopted at the 2nd World Conference on Research Integrity in July 2010. Four principles. No penalties. No central inspectorate. The document now circulates through university handbooks, funder policies, training modules, and codes of practice across dozens of countries. In Singapore itself, enforcement still lives in institutional offices rather than a single national tribunal.
A global statement with local gaps
The Singapore Statement rests on four principles: honesty in all aspects of research; accountability for conduct from proposal to publication; professional courtesy and fairness in working with others; and good stewardship of research on behalf of others. A version of each principle appears in policies at the National University of Singapore, Nanyang Technological University, the Agency for Science, Technology and Research (A*STAR), and Singapore's polytechnics. What differs is not the language. It is what happens after an allegation arrives.
Fabrication and falsification get defined in similar terms; plagiarism sits in its own administrative route. Investigation routes vary. A clinician working under the Human Biomedical Research Act falls into one regulatory stream. A lab scientist at A*STAR enters an institutional disciplinary procedure. A graduate student may face a university academic misconduct process. The result is a framework of shared principles and distributed enforcement, not the single adjudication body found in countries such as Denmark or the United States.
Here's the catch
The Singapore Statement is globally admired and locally optional. It is a standard, not a statute. Endorsing it costs nothing, and nobody loses funding for ignoring it. That is the catch: a document can shape norms without changing incentives.
Incentives are where research integrity usually collapses. Promotion committees still read journal lists before they read methods sections. That is not a Singapore-specific problem, but it is a Singapore-relevant one because the country's research evaluation system rewards output volume and international ranking. Journal indexation is not a quality guarantee, and treating it as one pushes researchers toward fast, clean, positive results rather than slow, messy, transparent ones.
Singapore's institutions have built research integrity offices and made responsible conduct training more common. What they have not built is a national public registry of misconduct findings, nor a uniform appeals route across universities and public research agencies. A finding at one institution may stay invisible to another. That matters when researchers move institutions, which they do often.
Photo by Arul Kumaran on Unsplash
What the principles mean in practice
The four principles are deliberately broad, so they need translation before they become usable policy. Here is how Singapore institutions typically operationalise them.
- Honesty. Accurate reporting of methods, datasets, laboratory records, and results; no fabrication, no falsification, no misrepresentation, and no undisclosed image manipulation in grant applications, submitted manuscripts, published papers, or peer reviews.
- Accountability. Researchers are responsible for work they supervise, including data ownership, record retention, author lists, and correction requests. Authorship disputes remain one of the most common integrity complaints because contribution thresholds are set at faculty level, not nationally.
- Professional courtesy and fairness. Fair treatment of collaborators, reviewers, students, and peers. This stretches into harassment and abuse of power inside laboratories, where early-career researchers are most exposed.
- Good stewardship. Responsible management of data, materials, funds, and human and animal subjects. The Human Biomedical Research Act adds statutory consent and review requirements for human biomedical research.
What the framework does not include are specific sanctions. It tells institutions what to value; it does not tell them what to do with a confirmed case. Sanctions run from retraction, correction, dismissal, grant termination, and referral to professional regulators, depending on the institution and the severity of the misconduct.
Human Biomedical Research Act and institutional routes
Singapore's strongest legal backstop is the Human Biomedical Research Act, passed in 2015 and phased in over subsequent years. It regulates consent, institutional review boards, tissue banking, and restricted research. It is not a research misconduct statute across all disciplines, but it creates enforceable duties for a significant slice of Singapore's research portfolio. A breach can carry fines and imprisonment for the most serious offences. The full statute is available on Singapore's official legislation portal.
For non-biomedical work, there is no equivalent act. Engineering, computing, social science, and humanities research sit under institutional policies rather than national legislation. That split is practical: biomedical research carries direct human-subject risks, while other fields carry different risks such as dual-use outputs, privacy breaches, or algorithmic harm. But it also means a national framework in name can feel like two frameworks in operation.
Training has become the most visible compliance tool. NUS and NTU require research integrity coursework for many incoming graduate researchers and postdoctoral staff. A*STAR runs its own training for scientific staff. The effect is real but uneven; a mandatory online module can tell someone what fabrication is without changing whether they report a supervisor who asks them to clean an image. The full text of the 2010 Statement remains available from the World Conferences on Research Integrity.
What changed once AI entered the room
Research integrity in Singapore now runs into generative AI on two fronts. First, AI makes fabrication faster: paragraphs, figures, datasets, and even fabricated reviewer identities can be manufactured at scale and submitted before a human reviewer notices. Second, AI detection gives institutions a new tool to catch anomalies in figures and manuscripts. AI tools flag doctored research data with increasing speed, but their use raises due-process questions when a flag becomes the basis for an investigation.
Cross-border collaboration adds another layer. Singapore's research workforce is highly international, and many projects involve co-authors in jurisdictions with different misconduct rules. A retraction initiated by a US funder or a Chinese university investigation can reach a Singapore co-author without a clear jurisdictional protocol. The Committee on Publication Ethics supplies guidance on retractions and authorship that many Singapore journals use as their default reference. The Singapore Statement remains useful here precisely because it was written for global collaboration, not for one country's courts. Its limit is also obvious: global documents do not decide who must investigate whom.
Photo by Guo Xin Goh on Unsplash
Reality grade: B minus
The hype says Singapore has a national research integrity framework. The reality is closer to a shared ethical vocabulary with a patchwork of institutional enforcement and one strong piece of biomedical legislation. The principles are clear. The training is widespread. The enforcement gap is where misconduct goes quiet.
Grade: B minus. Above average in Asia, where few countries have institutionalised research integrity training at this scale. But below the standard of a country that wants to be a global research hub. The missing pieces are a national misconduct register, a consistent appeals process, protection for whistleblowers, and public reporting of confirmed findings. Without those, the framework protects reputations more than it protects the next PhD student who says no.
The Singapore Statement was never meant to be a national enforcement code. It does what a statement can do. It names the behaviour. The rest is work for deans, research funders, laboratory heads, and the senior researchers who set group norms. A framework is a floor, not a culture. Culture is what happens when nobody is watching, and four principles only work if someone in the lab is willing to say, 'we don't do that here.'