For keen observers of the science policy scene, the most eye-catching announcement in last week’s government response to Nurse 2.0 (or The Independent Review of the UK’s Research, Development and Innovation Organisational Landscape, to give its formal title) was the launch of a new metascience unit.
Headed by Ben Steyn —a rising star in the Department of Science Innovation and Technology’s (DSIT) ranks, who was part of the start-up team for the Advanced Research and Invention Agency (ARIA) and has a PhD in the philosophy of science and technology—the new unit will work across DSIT and UKRI, with an initial budget of £10m to conduct experiments, test and evaluate the effectiveness of new approaches to research and innovation funding.
This is a bold and encouraging move, which signals DSIT’s ambition to operate at the vanguard of a growing movement for more sophisticated and systematic uses of evidence in research policy and funding.
And while there is now a scattering of metascientific initiatives and centres in universities, think tanks and funding agencies across Europe and the United States—my Research on Research Institute (RoRI) among them—it makes the UK the first country in the world to give metascience a formal institutional berth in government. If it weren’t profoundly un-metascientific, one might say we’re now on track to becoming a “metascience superpower”.
The rise of metascience
Like most good ideas, the new unit has several parents, with Ben Johnson, ministerial policy adviser in DSIT, and Stian Westlake, Executive Chair of the Economic and Social Research Council, also playing an important role in its creation. Both have long track records of engagement and support for such work: Stian from his time as policy and research director at Nesta, ministerial adviser and chief executive of the Royal Statistical Society; and Ben from his stints at HEFCE and Research England, then as an adviser to the rotating cast of ministers holding the science brief over recent years. Ben was also the secretariat lead and a co-author of The Metric Tide, my 2015 review of research metrics, which included as one of its recommendations a call for more government investment in “the science of science and innovation policy”. So eight years on, it’s great to tick that one off the to-do list!
Metascience is sometimes talked about as meta-research, science of science, or research on research. But beyond fashionable buzzwords, is this anything more than old wine being shipped in shiny new bottles? Haven’t policymakers long drawn on different types of evidence and data to inform decision-making about research, as in other fields? Or as my University of Cork colleague, Des Fitzgerald, asked on social media this week:
someone convince me that ‘metascience’ is anything other than STS [science and technology studies] stripped of any kind of structural or critical analysis.
I have some sympathy with these points. The turn to metascience can sometimes be used in a rather breathless and unreflective way to assert the importance of certain questions and methodological approaches over others. Or it can be cast as a new or emerging discipline, which is, in my view, a straightforward category error.
Questions of definition
I see metascience not as a discipline, but as an orientation or mode of engaging with questions and challenges that all researchers, funders and research policymakers encounter from time to time in the systems, networks and institutions that we inhabit. These may centre on how we govern, deliver, evaluate or communicate research; how we can make research funding and investment more agile, efficient and effective; how we can expand the diversity of the people and places that contribute to, and benefit from, research; or how we can improve the integrity, rigour and reproducibility of research findings. Many of us choose at certain points in our career to devote time and energy to such questions, typically as a side project to our main work, turning the methods and tools that we’ve mastered in other contexts back on the research system itself.
In 2021 and again in 2023, RoRI joined forces with Brian Nosek and colleagues at the Center for Open Science and the AIMOS network in Australia, to hold a big international conference on metascience. Hopefully we’ll do the same again in 2025. Ahead of the 2021 meeting, we gave a lot of thought to the way we were defining and talking about metascience, and ended up with this explainer:
There is a long history of research about the scientific process, particularly with fields such as philosophy of science, sociology of science, and science-technology studies contributing unique insights about how science operates. There is also a growing cadre of researchers deploying modem methodologies and big data to investigate the scientific process. Together, these communities of researchers and stakeholders are the research and development pipeline for improving research practices.
We also ran a couple of thoughtful panels on defining metascience, which you can still revisit online.
Scaling up metascience
If one adopts this more inclusive definition, then from a government perspective, metascience is a distributed capability that can be found all over the R&D system and, with modest investment and coordination, harnessed as a powerful resource to support strategy and decision-making. I hope that this is the approach the new unit will adopt – seizing opportunities to connect and scale up existing efforts, and to mobilise these towards more systematic experimentation and analysis.
From a UKRI perspective, this also brings us back to some of the early ambitions for the integrated agency when it launched in 2018 to develop an evidence-informed “culture of evaluation”. For a number of reasons – many external to UKRI itself – this proved harder to realise during UKRI’s first phase than its architects anticipated.
Deployed creatively, the metascience funding line that ESRC will operate with the DSIT unit, could be a game changer. There’s now a growing roll-call of experiments that funders worldwide have carried out with their grant-giving and evaluation, including NERC, Wellcome and the British Academy in the UK. But as we argued recently in The Experimental Research Funder’s Handbook, too many of these are small-scale, short-term pilots without proper baselines against which to assess the effects of any intervention.
The reach that DSIT and UKRI combined have across the UK system creates the potential to be far more ambitious in the design, delivery and evaluation of experiments – particularly when combined with other methods and perspectives. Replying to queries on X (formerly twitter) last week, Stian Westlake was keen to emphasise this plurality, noting that:
this announcement shouldn’t be taken as a sign of RCT/experimentation purism… Experimentation is very much “part of a nutritious breakfast”, along with improving SR&I data…qualitative analysis, and the experience of researchers and of relevant organisations like learned societies.
When John Kingman stepped down two years ago as the inaugural Chair of UKRI’s board, he ended his valedictory speech with a lament that too much in science and research policy “tends to turn on gut feel of the individuals involved, than on hard evidence and analysis.” And he called on UKRI “to take a deliberate strategic decision to sponsor and promote more good research, analysis and evidence-gathering on “what works” in policy on science, R&D and innovation.”
For all the noise and distractions over “wars on woke in science” and other ephemera, we should credit Ottoline Leyser, Stian Westlake and others in DSIT and UKRI for ensuring that, when they eventually pass the leadership of UK research onto others, they can point to concrete progress in moving beyond gut feel and towards good evidence as the basis on which we determine policy and funding.
Let’s seize this metascientific moment to make UK research work better and smarter for everyone.
