Making data sharing the norm
Progress: some progress (issues understood, needs operational implementation)
At the national and institutional level, individual researchers and projects need to build data sharing into their daily work. Researchers are not necessarily rewarded for providing and improving raw data, but are judged largely on their publication record. This gives them an incentive to control access to their datasets until after they have published their results. As a consequence, datasets are either built for a particular project or publication and access is restricted; or sharing of data becomes a neglected effort, operated on an inadequate budget despite its great value to the community at large. The whole of this framework relies on freely available, reusable data, yet even now some new data are being published in restricted forms, so it is important to tackle this as a matter of urgency.
In some fields, such as genomic research, publication relies on the underlying data being deposited in a common data store such as GenBank. The next step will be to broaden this to other fields, so that funding or publication is compromised if the underlying data are not made permanently or openly available.
In the short term, the priority will be to implement mechanisms for citing data, including use of DOIs and data standards, and promoting recognition of data owners and data managers, in concert with changes to policy incentives to use data sharing as a criterion for awarding funding.
In the medium term, providing data and making improvements to data quality should become valued as a service to science, giving institutions and individuals an incentive to make data available.
In the long term, data sharing through permanent archives or national repositories will become part of the language of science, just as citing publications or type specimens is now.