Abstract

There is a growing wish for science to become as open as possible, with the aim to increase the impact, reproducibility, interdisciplinarity, and efficiency of research. Two important elements in this respect are open research data and FAIR data management (Wilkinson et al., 2016), which are increasingly becoming priorities on the agenda of funding agencies and governments.
Open research data refers to freely accessible information in the form of text, numbers, images, audio, etc., that has been used and/or produced in research endeavors; it includes the data underlying scientific publications, as well as their metadata (European Commission, 2019; Norwegian Ministry of Education and Research, 2018; OECD, 2015; The Research Council of Norway, 2017; UNESCO, 2012).
The FAIR principles stand for Findable, Accessible, Interoperable and Reusable (Wilkinson et al., 2016). In a nutshell, Findable refers to the presence of metadata and a unique and persistent identifier for a dataset; Accessible means that a dataset and its accompanying metadata are understandable to both humans and machines, and are deposited in a trustworthy repository; Interoperable refers e.g. to the use of accepted vocabularies for the data and metadata; and Reusable involves e.g. the use of standard licenses for a given dataset (LIBER, 2020). The main purpose of FAIR is to increase the quality of research data, and to facilitate their interpretation and potential re-use.
Despite the increased attention on the matter, there are still challenges to make research data as open and as FAIR as possible. We present and discuss some of these challenges inspired by a survey among academic staff at Nord university. An example of such challenges is that researchers need to adhere to GDPR, which imposes several restrictions on the handling of personal data. In particular, researchers in the social sciences, who often work with data involving human participants, experience a tension between open research data and GDPR. In this respect, many issues arise. For instance, how to anonymize research data containing personal information while keeping most of their value for future re-use. Another example refers to the presence of various traditions and practices across research fields, and the need to operationalize the FAIR principles accordingly. For instance, building upon existing research data in the social sciences may be challenging not only because of the variety in types of data (e.g., qualitative interviews, observational data, and data from surveys), but also due to the plurality of theoretical perspectives and philosophical views, where knowledge is advanced not necessarily through a process of accumulation, but through the illumination of different aspects of a phenomenon.
The topics of open research data and FAIR data management are complex, and discussing the challenges associated with them is important. Such a discussion can promote a reflection on how support services within research data management can assist researchers meet the increasing demands from funding agencies, governments, and publishers regarding open and FAIR research data.