Only one in 10 physical science researchers are sharing their data in a format that meets internationally agreed principles, a new study suggests.
In a paper??by IOP Publishing (IOPP), which analysed more than 30,000 research articles, researchers found rates of sharing standardised data are as low as one in 20 in some disciplines.
The study comes after IoPP, the publishing arm of the?UK’s Institute of Physics, began asking authors in 2022 to state whether they had shared their data in line with the?, which aim to make data sets findable, accessible, interoperable, and reusable.
In 2023, it introduced a policy requiring researchers who are unable to, or choose not to, share their data publicly to disclose the reason why.
According to the paper,??64 per cent of articles published in 2023 indicated that data was freely available to access via some direct route, which is up from 50 per cent in the prior year.
However, only 11 per cent of articles in IoP journals in the past 12 months declared that their data was shared in a Fair way, while the figure was 8 per cent for 2022.
The highest level of sharing took place in environmental science, where 81 per cent of papers shared data in some way, with 59 per cent meeting Fair criteria. For physics, the figures were 72 per cent and 18 per cent.
For engineering and materials sciences, however, just 55 per cent shared data and only 8 per cent did so in a way that met Fair criteria.
For papers published in 2024, the most commonly cited reason for not sharing data was the lack of a suitable repository, followed by concerns over confidential or sensitive data.
Other reasons for not sharing data includes not holding data in a format that is suitable for sharing and the time and costs of sharing data.
However, the reasons for not sharing data differed between continents and countries, the study says. In China and India, researchers were most likely to state that they were unfamiliar with repositories that they could use to share their data, while in the UK, the primary concerns revolved around data confidentiality and the accessibility of the data format to others.
For researchers in the US, the most common obstacle was the concern that the data format would not be accessible even if shared, highlighting a need for better standards and tools, and the need for a culture of data management?that seeks to enable wider data access and interoperability.??
Daniel Keirs, head of journal strategy and performance at IOPP and lead analyst of the study,?said that it would “take a concerted effort across the scientific ecosystem” to ensure the Fair principles, which were introduced in 2016, were more widely used.
“What we’re seeing is that the barriers to open data are numerous and diversified, even within the physical sciences. We need to rethink several key areas, such as the standardisation of data formats, proprietary data issues, data repositories and how we incentivise data sharing,” he said.
