Scientists and Archivists Call for Robust Networks to Protect Research Data from Loss and Political Interference

Concerns over the disappearance of research data due to political and technical issues are growing among scientists, librarians, and archivists. These fears are not unfounded; a 2023 study by Dorothea Strecker and colleagues at Humboldt University of Berlin revealed that out of more than 3,000 research-data repositories, 191 had shut down. Alarmingly, 90 of these closures resulted in the loss of data or a failure to name a successor repository, highlighting the significant risks to data permanence. Data repositories like PubMed, the Web of Science, and Scopus are crucial for enhancing data accessibility, reliability, and reusability. However, a March 2023 outage of PubMed demonstrated that even these robust resources are not immune to failures. Earlier studies, including a 2015 analysis of 326 databases, found that over 60% had become non-functional or had limited functionality within 18 years, with only 14% being archived. Jonas Recker, an archivist at the Leibniz Institute for the Social Sciences in Mannheim, Germany, emphasizes that repository closure risks are often overlooked. Common reasons for shutdowns include lack of funding, technical difficulties, and organizational shifts. Recker points out that the recent removal of data sets related to gender and diversity from U.S. government websites underscores the importance of building resilient networks with diverse geolocations, technologies, and funding sources to safeguard scientific knowledge. To mitigate these risks, Strecker advocates for a more distributed approach to data preservation, involving the formation of networks that can take over data curation if a repository fails. This approach would reduce the pressure on individual repositories. For instance, a group of repositories from the German Network of Educational Research Data is developing a "cessation checklist" to guide repositories on preparing for potential data transfers. Key steps include creating formal agreements and clarifying transfer conditions, ensuring a smooth transition in emergency situations. Preservation networks for text publications already exist, offering models for data preservation. One such network is LOCKSS (Lots Of Copies Keep Stuff Safe), an open-source program developed by Stanford University. LOCKSS creates multiple copies of content and stores them on library servers across the globe, funded by annual fees from participating institutions. Another option is the CLOCKSS (Controlled Lots of Copies Keeps Stuff Safe) dark archive, which only grants access during specific crises, such as publisher shutdowns or prolonged content unavailability. Martin Eve, a researcher in literature, technology, and publishing at Birkbeck, University of London, calls on publishers to engage with these services. His recent study of nearly 7.5 million research papers found that more than one-quarter lacked proper digital preservation. Without such measures, Eve warns, future researchers will face broken links and a fragmented knowledge transmission system. However, economic disparities pose a significant challenge. Wealthier publishers are more likely to implement robust preservation measures, often leaving smaller, less-funded journals vulnerable. Juan Pablo Alperin, a scholarly communications researcher at Simon Fraser University in Vancouver, Canada, notes that the costs of services like CLOCKSS can be prohibitive for journals in under-resourced regions, particularly in the global south. Thus, it is crucial to develop accessible and affordable alternatives to ensure comprehensive data preservation for all. In response, various initiatives are emerging to address these issues. Strecker and Recker recommend forming networks and preparing cessation plans. Eve encourages publishers to prioritize digital preservation, while Alperin stresses the need for inclusive solutions that consider the economic constraints of diverse publishers and repositories. The collaborative efforts of these professionals aim to fortify the scholarly record against potential threats, ensuring that valuable research data remains accessible and usable for future generations. Industry insiders highlight the urgency of this issue, emphasizing that a distributed and networked approach to data preservation is essential. LOCKSS and CLOCKSS are seen as leading examples, but they must be complemented by more affordable and accessible options to support a wide range of research communities. Both Strecker and Recker are actively involved in creating practical tools and guidelines, while Eve and Alperin advocate for broader policy changes and financial support to ensure the sustainability of data repositories and the integrity of the scientific record. The Humboldt University of Berlin, where Strecker conducts her research, is a leading institution in open science and data management. Similarly, the Leibniz Institute for the Social Sciences, home to Recker, is committed to advancing social and behavioral sciences through robust data preservation practices. These institutions and their researchers play a crucial role in shaping the future of data preservation in the scientific community.