Research Data Governance. The Need for a System of Cross-organisational Responsibility for the Researcher's Data Domain

The paper addresses a central problem in research data management (RDM): institutional governance and support structures are poorly aligned with researchers’ practical needs across the research data life cycle. As a result, responsibilities and accountabilities are fragmented, leading to recurring obstacles. Three archetypal examples illustrate these dead ends: (1) publishing impeded by missing or unclear ethical approvals when multiple organizations are involved; (2) lack of sustainable data hosting during active project phases due to absent or poorly communicated responsibility models for service-supported data maintenance; and (3) impeded data transfer and reuse due to incompatible provider regulations across the life cycle. The guiding questions concern which governance systems affect research data at different life cycle stages and how decision-making and accountability should be distributed between individual researchers and other stakeholders. The study argues that current institutional policies typically assign broad accountability to researchers while leaving organizational support ambiguous, creating a gap that a research-centric governance structure must fill.

The essay situates research data governance within established data governance literature and RDM standards. It adopts definitions emphasizing decision rights and accountability (Data Governance Institute, 2024) and distinguishes governance (defining decision domains and loci of accountability) from management (implementing decisions) per Khatri and Brown (2010). Governance is framed as a company- or organization-wide framework for assigning decision-related rights and duties (Otto, 2011), typically instantiated through governance bodies (Ebel, 2021) and rule-level instruments such as policies, standards, and guidelines. Within academia, research data policies and guidelines for good scientific practice (e.g., DFG Code of Conduct) and cross-organizational standards like FAIR (Wilkinson et al., 2016; Barker et al., 2022; Chue Hong et al., 2021; Strasser, 2015) provide high-level compliance goals but generally lack explicit decision-rights and accountability mechanisms throughout the research data life cycle. The research data life cycle is treated as a process model helpful for implementation (Cox and Tam, 2018) but not inherently governance-aware. Studies note that data governance is more mature in computational domains (Benfeldt Nielsen, 2017) and that researchers often identify with multiple communities across the data lifecycle (Kouper, Raymond, and Giroux, 2020). Frameworks distinguishing intra- and inter-organizational scopes and data scopes (Abraham, Schneider, and Vom Brocke, 2019; Lis and Otto, 2020) underscore the need for cross-organizational, research-domain-aware governance, particularly since research data differs significantly from standardized administrative data (Jim and Chang, 2018).

This is a conceptual and argumentative essay grounded in the author’s several years of professional experience consulting on research data management in the digital humanities, cultural studies, and social sciences. The paper distills recurring obstacles into three illustrative examples (publishing/ethics, hosting/maintenance, and transfer/regulatory incompatibilities) and synthesizes insights from data governance frameworks and RDM standards to articulate requirements for a research-centric governance model. No formal empirical study is conducted; rather, the approach combines reflective practice with a structured engagement with existing governance literature and policy instruments to propose a workable organizational design.

- Institutional policies typically assign overarching accountability for the research data life cycle to researchers or PIs while leaving institutional support structures vague, creating responsibility gaps at critical stages (ethics approval, hosting/maintenance, and data transfer). - The research data life cycle can be viewed as a data domain requiring governance; however, conventional life cycle models and data management plans lack explicit decision-rights and accountability specifications. - Research data differs from administrative data in variability, interdisciplinarity, cross-organizational flows, and value across stages, making organization-wide governance insufficient without domain- and discipline-specific integration. - Existing governance in libraries and computing centers tends to focus on specific stages (e.g., repositories for publication; compute hubs for creation/analysis) and is not well integrated across the life cycle, leading to compliance and interoperability challenges. - A research-centric, cross-organizational governance system is needed to bridge institutional policies and discipline-specific requirements and to explicitly define decision-making and accountability across life cycle stages. - A polycentric governance model anchored at the faculty or department level, with an RDM committee representing disciplines and key stakeholders, can act as a broker linking intra- and inter-organizational governance (e.g., ethics boards, IT governance, libraries) and support case management. - Such a committee can clarify ethical approval responsibilities (publishing), co-develop hosting and maintenance strategies with IT (active data phases), and provide expertise on data transfer and regulatory alignment, building sustainable, research-specific governance knowledge over time. - Locating governance near researcher communities also facilitates integration with curricula and development of data literacies linked to governance.

The proposed research data governance model directly addresses the identified dead ends by making explicit who decides what, when, and under what conditions across the research data life cycle. By positioning governance close to researcher communities (faculties/departments) and adopting a polycentric structure, the model aligns decision rights with domain knowledge and practical workflows while connecting to organization-wide and external infrastructures. This improves coordination among ethics committees, IT services, and libraries, thereby reducing delays in ethics approvals, clarifying hosting and maintenance responsibilities during active research, and harmonizing data transfer regulations and identity management across services. The approach enhances compliance with FAIR and funder requirements by translating high-level policies into actionable, accountable processes embedded in researchers’ contexts. Ultimately, it supports data quality, reproducibility, and sustainability, while preserving research autonomy and accommodating disciplinary heterogeneity.

The paper advances a research-centric, cross-organizational governance proposal that reframes the research data life cycle as a governed domain requiring explicit decision rights, accountability, and coordinated support. It contributes (1) a diagnosis of responsibility gaps caused by misaligned institutional governance and the realities of research workflows; (2) an argument for polycentric governance anchored at faculty/department level; and (3) a practical design idea for an RDM committee to broker between intra- and inter-organizational stakeholders, build governance knowledge, and support case-specific needs. Future work should pilot such governance structures across diverse disciplines and institutions, develop standard operating procedures and accountability matrices for life cycle stages, evaluate impacts on compliance and FAIRness, and explore integration with curricula to cultivate governance-aware data literacies. Further research might also assess scalability, resource implications, and interoperability with national and international research infrastructures.

The work is a conceptual essay without formal empirical evaluation, relying on the author’s consulting experience and selected illustrative cases. Its perspective is grounded primarily in the digital humanities, social sciences, and cultural studies, and in the context of European (notably German) academic governance landscapes; generalizability to other disciplines or regions may be limited without adaptation. The proposed structures (e.g., faculty-level RDM committee) are not tested for feasibility, resource requirements, or long-term sustainability, and the paper does not provide quantitative evidence of effectiveness.