On 1 July 2026, the UK Reproducibility Network (UKRN) published Working Paper 12: the STAR Study (Sustainable and TrAnsparent Research data) — a report on where open research data (ORD) practice actually stands across UK higher education, roughly a decade after the sector’s open-data policy push began in earnest. It is one of the more candid assessments UKRN has produced: the headline is not that ORD has failed, but that progress has stalled at what the report calls a “glass ceiling.”
What the STAR Study looked at
STAR is a qualitative study, not a metrics audit. UKRN and the research team gathered university staff — the people who actually run research data services, from repository managers to open-research leads to library and research-office staff — through interviews, focus groups, and workshops across a cross-section of UK institutions, to ask a more specific question than “how much open data exists”: what has changed operationally since institutions started building ORD infrastructure and policy around 2016, and what is still getting in the way. Findings were also presented and summarized in a companion paper, “Realising Open Data Principles In UK Research Institutions”, published in the International Journal of Digital Curation.
That decade-long lens matters for how the findings should be read. STAR isn’t asking whether UK universities have repositories, DMP requirements, or open-data policies — most now do. It’s asking whether those structures have translated into open data being a routine, well-supported part of how research gets done, or whether they remain scaffolding around an activity that is still exceptional rather than default.
Key findings
The report’s central finding is a gap between capability and practice. UK institutions have, over the past decade, built the infrastructure — repositories, ORD-specialist roles, institutional policies — that open data requires. But STAR finds that having the infrastructure has not made sharing data a routine part of the research process for most researchers. Several more specific patterns recur through the report:
- Compliance-driven, not values-driven, engagement. Where researchers do share data, STAR finds it’s more often prompted by a journal’s data-availability requirement or a funder mandate than by institutional policy or researchers’ own motivation to make data open. The report frames this as a structural weakness: compliance-driven behavior changes when the requirement changes, whereas values-driven behavior is more durable.
- Low researcher familiarity with the practical mechanics. Many researchers remain unfamiliar with the actual process of depositing, describing, and licensing a dataset — a gap that persists even at institutions with mature repository infrastructure and support staff in place.
- Monitoring has not kept pace with policy. Institutions can point to an ORD policy far more reliably than they can point to systematic data on whether it’s being followed. Without that monitoring layer, it’s difficult for institutions to know where their own gaps are, let alone close them.
- Disciplinary variation is still largely unaddressed. What counts as “the data” and what a reasonable sharing norm looks like differ substantially by field — and STAR finds that institutional ORD policy is still largely written in generic, discipline-agnostic terms that don’t reflect that variation well.
Barriers the report identifies
STAR is explicit that the stall isn’t attributable to researcher indifference alone. The barriers it identifies are largely structural and institutional:
- Capacity constraints in institutional repositories and the staff who support them, relative to demand.
- Workload recognition. Time spent preparing, describing, and curating a dataset for sharing is rarely counted, credited, or costed as part of a researcher’s workload in the way that publication output is.
- Unclear or insufficient funding for long-term data archiving and stewardship, particularly beyond a grant’s active period.
- Limited career-framework recognition for ORD contributions — open-data work doesn’t reliably count toward promotion or appraisal the way publications do.
- Support for external, discipline-specific repositories lags behind support for generalist institutional repositories, even though field-specific repositories are often the more natural home for a given dataset.
What the report recommends
STAR’s recommendations are aimed less at researchers than at the institutional and system-level actors who set incentives:
- Formal recognition of ORD contributions within academic career and promotion frameworks, comparable to how publication record is already recognized.
- Development of discipline-relevant data-sharing norms and guidance, rather than a single generic institutional policy applied uniformly across fields.
- Investment in the infrastructure needed to actually monitor ORD practice, not just to publish a policy — closing the visibility gap the report identifies between stated policy and observed behavior.
- Formal processes for institutions to evaluate the real value, cost, and impact of open data work, so that support and funding decisions are grounded in evidence rather than assumption.
- Stronger institutional support for external, discipline-specific repositories alongside generalist ones.
- A shift from a compliance-first framing of ORD toward one that engages researchers’ own disciplinary values — the report’s own language is that further progress depends on “clearer incentives, stronger alignment of values and practices, and coordinated action across funders, institutions and the wider research ecosystem.”
Why this matters for research administrators
STAR is squarely aimed at the audience that runs research-data services day to day — the same people it interviewed to produce the report. A few practical implications stand out:
- If your institution’s ORD monitoring is closer to “we have a policy” than “we have data on whether the policy is being followed,” STAR’s finding is that you are not unusual — but it also flags this as the most actionable near-term gap to close, since it’s a prerequisite for everything else.
- Workload and career-recognition gaps for data curation work are a recurring theme across UKRN’s broader Concordat on Open Research Data work, not unique to this report — institutions revisiting workload models or promotion criteria have a concrete, cited justification to point to.
- Discipline-specific guidance is flagged as underdeveloped almost everywhere STAR looked — a generic, one-size-fits-all DMP or repository-deposit policy is unlikely to satisfy the report’s own recommendations, even where it satisfies a funder mandate on paper.
Frequently asked questions
What does STAR stand for?
Sustainable and TrAnsparent Research data. It’s UKRN’s Working Paper 12, published 1 July 2026.
Who ran the STAR Study?
The UK Reproducibility Network, a UKRI-recognized network of researchers and institutions promoting research reproducibility and open research practice across the UK. UKRN also maintains the broader Concordat on Open Research Data initiative and the annual UKRN Open Research Indicators work.
Is this the same as UKRN’s open and transparent research practices survey?
No — that is a separate, earlier UKRN dataset and survey (published via Scientific Data in 2024) on researcher-reported open and transparent practices. STAR is a distinct, more recent qualitative study focused specifically on institutional ORD infrastructure and practice, drawing on interviews and workshops with university staff rather than a researcher survey.
Does STAR say UK universities have failed on open data?
No. The report’s own framing is more specific than “failure”: institutions have built real infrastructure and made real progress since around 2016, but that progress has reached a plateau — a “glass ceiling” — where further gains depend on addressing incentive structures, workload recognition, and monitoring gaps rather than on building more infrastructure alone.
Note: this article draws on UKRN’s own published summary of the STAR Study and the companion peer-reviewed paper in the International Journal of Digital Curation. Readers who need to cite specific figures (for example, exact numbers of institutions or participants involved) should consult UKRN Working Paper 12 directly, as pagination and precise counts are best confirmed against the primary report rather than secondary summaries.
