Paper Presents Framework to Assess Trustworthiness in Research

A newly published paper in the Proceedings of the National Academy of Sciences (PNAS) introduces a systems-level framework for evaluating the trustworthiness of scientific findings across disciplines. Authored by a multidisciplinary group, the paper draws on expertise in metascience, research integrity, and science communication. All six authors are members of the National Academies’ Strategic Council for Research Excellence, Integrity, and Trust, though the paper itself is not an official National Academies publication.

The authors propose a framework with seven components of trustworthy research: being accountable, evaluable, evaluated, well-formulated, controlling bias, reducing error, and being well-calibrated, meaning that claims match the evidence. These elements operate at three levels—the research itself, the researchers who conduct and assess it, and the organizations that support it. Each level includes examples of possible indicators, encouraging a shared language for assessing research quality.

Rather than tying credibility to journal prestige or citation counts, the authors argue that trustworthiness results from systems and behaviors that encourage critique, scrutiny, and cumulative knowledge building. Trustworthiness is therefore presented as a process characteristic, not a fixed outcome or single metric. The framework’s intent is to help researchers, institutions, journals, and funders evaluate quality through systematic, transparent practices rather than reputation-driven measures.

The paper highlights a crucial distinction between trustworthiness and correctness. According to the authors, trustworthy research is not necessarily free from error, but is generated and reviewed in ways that make errors detectable and correctable over time. This openness to correction supports the self-improving nature of science. In contrast, when research lacks transparency or evaluative rigor, errors persist, slowing the collective progress of knowledge.

To facilitate improvement, the framework aims to align principles for both researchers and organizations. It encourages environments that make evaluation routine, promote accountability, and link research claims closely to the strength of their evidence. The authors also note that better indicators can improve how journalists, policymakers, and the public understand what makes research credible.

They conclude that current assessment practices often over-rely on indirect signals of prestige, which can mask important variations in quality. The proposed framework seeks instead to make trustworthiness observable through specific actions and institutional supports, acknowledging that developing valid and scalable indicators remains an essential next step.