Data from: Data sharing, management, use, and reuse: practices and perceptions of scientists worldwide

Background: With data becoming a centerpiece of modern scientific
discovery, data sharing by scientists is now a crucial element of
scientific progress. This article aims to provide an in-depth examination
of the practices and perceptions of data management, including data
storage, data sharing, and data use and reuse by scientists around the
world. Methods: The Usability and Assessment Working Group of DataONE, an
NSF-funded environmental cyberinfrastructure project, distributed a survey
to a multinational and multidisciplinary sample of scientific researchers
in a two-waves approach in 2017-2018. We focused our analysis on examining
the differences across age groups, sub-disciplines of science, and sectors
of employment. Findings: Most respondents displayed what we describe as
high and moderate risk data practices by storing their data on their
personal computer, departmental servers or USB drives. Respondents
appeared to be satisfied with short-term storage solutions; however, only
half of them are satisfied with available mechanisms for storing data
beyond the life of the process. Data sharing and data reuse were viewed
positively: over 85% of respondents admitted they would be willing to
share their data with others and said they would use data collected by
others if it could be easily accessed. A vast majority of respondents felt
that the lack of access to data generated by other researchers or
institutions was a major impediment to progress in science at large, yet
only about a half thought that it restricted their own ability to answer
scientific questions. Although attitudes towards data sharing and data use
and reuse are mostly positive, practice does not always support data
storage, sharing, and future reuse. Assistance through data managers or
data librarians, readily available data repositories for both long-term
and short-term storage, and educational programs for both awareness and to
help engender good data practices are clearly needed.