How Open is Research Data in the Engineering and Natural Sciences? A Norwegian Case Study Using Data Availability Statements

Abstract

Policies that mandate the sharing of research data connected to academic publications have been in effect for the past years across research fields. Several funding agencies, research-performing organizations, and publishers now require that the data supporting an academic article are made available to the highest possible extent. Following this development, most journals ask that authors include a statement describing the availability of the data used in their study. However, there is limited information on the extent to which the data used in articles within the Engineering and Natural Sciences are actually made available, and how.

This paper addresses this gap in our knowledge by analyzing the data availability statements from a sample of relevant articles, with at least one Norwegian institution contributing to the article. Our focus on Engineering and Natural Sciences is due to the critical importance of these fields in helping solve the complex problems of today’s society, for example how to achieve successful transitions to sustainable practices across industries. The focus on articles with at least one contributing institution from Norway is due to the country’s increasing contribution to publications in this area.

We are also interested in exploring the differences across subfields within the Engineering and Natural Sciences with respect to research data sharing (e.g. Biosciences and ICT), following the classification provided by the Norwegian Publication Indicator.

To examine these issues, we use a sample of 600 academic articles published in 2024, obtained from the Scopus database. In the sample, we kept the distribution of articles per subfield similar to that of the population (N=10.770 articles), which is the number of articles published in 2024 in the Engineering and Natural Sciences with at least one Norwegian institution contributing to the study. We evaluate that our sample is a good representation of this population.

We managed to retain a sufficient number of articles per subfield, except for the subfield “Electrical Engineering” which corresponded to a small percentage of the total number of articles. Thus, we combined this subfield with “Energy”. This was appropriate because these two fields have an overlap in terms of research focus (e.g. power systems and grids). Then, we selected ten journals from each subfield from which we randomly selected the needed number of articles, observing the distribution of articles per subfield.

The next step is to analyze the data availability statements of the articles in our sample (mode of data sharing and degree of data openness), and to record the metadata provided by Scopus in our analysis table (e.g. title and abstract). Then, we will analyze whether the degree of data openness differs across subfields in the Engineering and Natural Sciences. We expect to find systematic differences across these subfields and that domain-specific traditions have an important role in this regard.