A relational approach towards data sharing in a digital society

• In Event: Global Knowledge Networks: Exploring Social Capital and Knowledge Dynamics Across Borders

Proposal

We have remarkably little systematic empirical or theoretically supported understanding about the underlying structures, mechanisms, and drivers leading to successful data sharing: its extent; the relationships that advance scientific quality; how and why data sharing facilitates the transformation of scientific discoveries into innovation (Wu et al. 2022: 1f) as well as its impact on education and learning in a digital society––questions that remained unanswered with the available theoretical and methodological toolbox in the study of science. Traditionally separate scientific communities focus either on small-scale, in-depth case studies or large-scale “big data” research (Fortunato et al. 2018; Wang & Barabási 2021). Data sharing across disciplinary and organizational boundaries raises unresolved challenges and ethical concerns.

Data sharing has gradually gained more attention of researchers, policymakers, funding agencies, and research managers as they attempt to foster expansive sharing of data, and have articulated a clear and consistent vision of global Open Science (OS) (Bartling & Friesike 2014) as a driver for enabling a new paradigm of FAIR, data-driven research as well as accelerating innovation. It pushes the digital transformation forward and enables increasing access to educational resources worldwide. Governments consider OS as a disruptive approach to pursuing scientific efficiency, open innovation, and knowledge commercialization. Nevertheless, many researchers are extremely resistant to sharing their data, even when this is required by the funding agencies sponsoring that research. Forced data sharing contrasts with the demand for independence, autonomy, and integrity of scientists (Krücken 2020)––a phenomenon that has far reaching implications for researchers and the organization of research. Questions related to a fair and ethical treatment around the collection, processing, use, and sharing of data for individuals and research communities remain unanswered. However, data sharing is among the most important developments within global science, but a clear conceptualization and holistic approach towards data sharing is a research gap (Barlösius 2022).

Seeking explanations for varying incentives for and drivers of data sharing requires the analysis of cases that use different types of data and have contrasting disciplinary and organizational form-specific modes of data sharing, vary in their way to do science, and differ in their collaboration networks based on co-publications. I have pre-selected potential, contrasting cases from the Social Sciences and Humanities, Health and Medical Sciences, and STEM. To understand the impact of data sharing on scientific quality, the research problem will be investigated on multiple levels of analysis.

I explore the organizational field of fundamental research (DiMaggio & Powell 1983; Scott 2014) that enables data sharing among researchers within different disciplines and various collaborations between researchers in contrasting organizational forms. Combining institutional (Powell & DiMaggio 1991) and relational perspectives (Donati 2010; Powell & Dépelteau 2013) will generate new disciplinary and organizational insights into data sharing in established or emerging fields, including education. I argue that recent progress in neo-institutional theorizing holds important analytical potential for the study of data sharing dynamics, especially competition and collaboration within disciplinary, organizational, and global networks (Authors 2019, 2021, also Gazni et al. 2012; Powell & Oberg 2017). Relational sociology, as a nascent social-scientific paradigm, offers crucial theoretical and methodological tools to conduct such a study (Authors 2021).

My research applies an explanatory, sequential mixed-methods research. I will integrate a quantitative factorial survey (vignette study), qualitative, semi-structured interviews, and bibliometrics (publications, citations, patents, acknowledgements, new indicators for SSH), with methods borrowed from computer science (machine learning; deep learning).

Generating new research questions and methodological approaches on the basis of shared data leads to a transformation of research practices. I hypothesize that some disciplines and organizational forms of collaboration advance much slower and are less innovative than others. Those who do not collaborate will not advance their scientific quality, thus are systematically disadvantaged in making scientific discoveries leading to innovation. Refusing to share data reflects the norm that publications remain the only valued and rewarded factor in research evaluation.

I will pioneer a relational approach to data sharing within research processes to reveal a new paradigm of an open and collaborative research culture in education and beyond. This will motivate change in how scientists work together and raises awareness of and rewarding sharing data. and engagement with societal actors, encourage policymakers to reorganize and define new guidelines for interdisciplinary research collaborations and data sharing to solve complex scientific and societal problems.

Author

• Jennifer Dusdal, University of Luxembourg