The Growing Emphasis on Metascience

Moving into 2026, I have been reflecting on the current state of the world, scientific advances, and a general increase in metascience interest. The motivations discussed in my original post announcing the motivation for Metascience Hub have only grown throughout 2025: AI's impact on science and society, decrease in public trust, and the slow pace of institutional adaptation.

The perspective that I hadn't had at that point in time, however, was in the discussions held at the Swissnex Quantum Summit, the GESDA Summit, and Trust Valley Day in 2025. These discussions underscored the rapid pace at which technology is developing: shining a light on errors, missteps, and negligence in the global roll-out of AI, and urging policy makers, innovators, and researchers to think more critically as we move towards other breakthroughs and challenges.

Metascience answers this call by helping us to distinguish best practices in frontier fields, and find ways to maintain vigilance and adaptability in the face of rapidly-developing technological and intellectual landscapes. Don't take my word for it, though. Here are some examples of what is happening globally:

• In the UK, a Metascience Unit has been established in the Department for Science, Innovation, and Technology, with a £10M starting budget and £5M call for proposals.

• The Metascience Allianace was launched.

• Canada has built on recent momentum in metaresearch by opening a call for a Meta-research Community Connection Grant.

• Other efforts to roll out strengthened metascience initiatives in Japan are underway.

These efforts are all geared towards adapting science culture and policy in the face of 21st century challenges. Improvements in science funding, experimental design, the science-society interface, and the institutional framework of science are all necessary for bringing the scientific method into the information age and beyond with integrity.

Funding

Funding bodies are often under pressure as the financial gate-keepers of academic research. Recent global trends in research budget cuts have only served to increase the responsibility and pressure put on national and international funding schemes. This could be one reasons that organizations such as the Swiss National Science Foundation, National Science Foundation in the US, United Kingdom Research and Innovation, and others have become more involved with metascience and research on research efforts.

Collecting more data on the way research is being carried out introduces a new set of key performance indicators (KPIs) for funding bodies to track, ensuring that resources are being used efficiently and for high-impact means. This can hopefully maintain progress in the face of diminishing research budgets, while assuring the quality of the funded research.

Experimental Design

In referencing experimental design, I am referencing everything from the way of posing research questions, to technical methodology, to statistical analysis. Optimizing these aspects of the scientific method have long been a goal of metascience, and there is reason to believe that this will maintain an evolving effort as technology shifts our understanding of the world. The Duhem-Quine thesis specifically discusses the bias embedded in experimental design when empirical tests are based on assumptions and background beliefs of a given researcher, known as auxiliary hypotheses. Technological advances can bring about new considerations for auxiliary hypotheses, and therefore new challenges to modern researchers.

As AI begins to play an even larger role in researchers' day-to-day lives, digital literacy and the ability to assess the quality of data, results, and research methods will become more critical skills. With the uptick in AI's role in scientific fraud by generating data and articles, we run the risk that a given researcher may be more exposed to fabricated data in their literature review steps, and therefore studies may contain auxiliary hypotheses based in fabricated results.

The rapid development of science and technology leads to other challenges. The First Law of Thermodynamics, formally stated in 1850, has been subject to review in 2023, concerning systems that are not in equilibrium. New methods, mathematical approaches, and questions could have an impact on the way we think of the natural world, and therefore on scientific results collected with outdated sets of auxiliary hypotheses.

Technological developments, therefore, have a significant impact on the conception of research questions and experimental design. Understanding the impact of auxiliary hypotheses, increasing scientific literacy to account for these considerations, and quantifying best practices for experimental design in a modern context are key goals of research on research.

Science-Society Interface

Taking science out of the lab and communicating results to other scientists, policy makers, and the general public is a critical step in translating research into meaningful applications.

Before the onset of LLMs in scientific writing, in 2017, P. Plavén-Sigray, G. J. Matheson, B. C. Schiffler, and W. H. Thompson remarked on the decreasing readability of scientific texts in life sciences. After quantifying readability with two different metrics across over 700,000 abstracts published between 1881 and 2015, the authors found a three-fold decrease in the Flesch Reading Ease (FRE) scores over time. From an average FRE of 30-40 in the 1880s to an average of 10 in the 2010s, putting 22% of all scientific articles beyond a college graduate reading level.

Metascientists observing this trend are proposing many solutions to increase accessibility of scientific literature. Varying approaches to resolving issues of science communication, such as social media outreach, lay summaries, and potential readability metrics are being discussed as approaches that academic and publishing institutions can implement.

Institutional Frameworks

Globally, academic institutional frameworks are wrought with bias and dysfunction, but the good news is that the large majority of institutions are aware of this and taking decisive action to mitigate the impacts of bias.

In Switzerland, for example, the Swiss Academy for Arts and Sciences has supported surveying efforts to identify causes for the "leaky pipeline" in STEM, and the Swiss National Science Foundation (SNSF) has supported studies to investigate how gender biases play a role in the grant review processes over the periods of 2009-2016, and 2016-2023. Gender bias is not the only bias that exists in the peer-review process. For grant assessment, biases against high-risk projects, certain scientific fields, or interdisciplinary projects are also present in the decision-making panels. These efforts seek to understand the role biases play in shaping the research ecosystem.

The work doesn't stop at just understanding, however. Reformed CV structure for applicants and other bias mitigation approaches have been initiated to address the issues uncovered in their investigative research. In 2021, the SNSF proposed a lottery approach for awarding grants to proposals assessed as good, but not excellent. This approach directly acknowledges the factor of chance in proposal assessment and leverages it to counter biases of particular panels of evaluators.

Other institutions outside of Switzerland are also moving in this direction. The introduction of new measures and willingness to adapt to a shifting political, societal, and scientific landscape is promising and these efforts can only be further supported by metascientific studies on the effectiveness of new approaches.

Take-Aways

There are many issues in the scientific method and research frameworks, which are only being exacerbated by rapid technological development, decreased accessibility to accurate information, and increasing polarization. Many countries around the world are developing stronger metascience programs because metascience can provide the tools necessary to more completely understand the nature of these challenges, quantify their impacts, and track developments after implementing solutions.