In working to understand the effects of Journal Data Policies I’ve been spending time talking to some of the real experts – the support staff that advise on and deal with these policies every day. Last week – after a call for interested participants – I met with Bev Jones (Lincoln), Laurence Horton (LSE), Marta Teperek (Cambridge), Sarah Middle (formerly Cambridge, now a PhD student at the Open University), Kevin Sanders (St Mary’s) and Neil Jefferies (Oxford).
I’m aiming to have further meetings and visits to help me feed in to an RDA interest group on journal data policies. Publishers are already aware of the difficulties that such a complex system of policies causes and are looking to standardise – the valuable role that Jisc can play is to feed in the experiences of those who work with the policies every day in UK universities. If that describes you get in touch! I’d love to talk further.
We looked in detail at a variety of data policies from publishers including Springer Nature, Wiley and PLOS. But the focus of conversation concerned how the sharing of data that underpins a paper is just one point in a much longer process.
In an ideal world, a researcher will have been managing and curating their data (using the widest possible definition of data) from the start of a project (or other activity). There would be a Data Management Plan in place for the whole project, and consultation of a journal data strategy would simply be to ensure that the shared data that lives in a subject area or institutional repository is linked to the submission in the required way.
In practice, and in particular for unfunded/non-project research, the point of article submission may be the first time a research has encountered the need to manage or share data. This, clearly, is not an optimal position.
In bridging the divide between ideal and pragmatic use cases, journal data policies have tended towards the latter – setting out, in legalistic terms, what is expected and where exceptions may be permitted. Clarity in expression supports journal editorial teams, who are often required to make calls on exceptions or variances without having had the training or experience which would support them in doing so.
In some cases, publishers have begun to put in place systems which assume this pragmatic use case – auto-depositing a supplied dataset on Figshare, for example. But this could cause issues where data also existed elsewhere, and prompted awkward questions on authors’ ability to update data, or to de-duplicate where multiple papers rely on the same dataset.
As is common in academia there is a suitable metaphor in religion – the ability to make a decision based on a well understood set of core principles is far superior to the application of scriptural rules! We liked, in particular, the COPDESS model of a subject area (in this case Earth and Space Sciences) statement of principles to which journals could sign up – a useful point of negotiation with editors.
Recommendations from this group centred around the need for support and training for journal editors. Whilst in general research data support staff would expect researchers to understand and relate to their own disciplinary data norms, we need editors to support and understand these norms too.