What a survey of humanities researchers told us about data sharing

As part of the STM Association’s Research Data Program, a group of publishers (including colleagues from Taylor & Francis, Routledge, F1000, Cambridge University Press, Brill, Oxford University Press, Wiley and SAGE, chaired by Matt Cannon (Taylor & Francis),  Rebecca Grant (F1000) and Kate McKellar (Wiley) have been working together since 2020 to explore the challenges and opportunities presented by humanities data sharing policies. This led to the development of an author survey which opened for responses in March of this year. This is the first large-scale, targeted survey on humanities data sharing practices that we are aware of.

We were delighted to be invited to write up the results of our survey in the 2022 State of Open Data, which was published in October. The State of Open survey, and associated white paper, represent the longest running longitudinal survey and analysis of open data practices in the world. It’s a hugely impactful publication, and this year’s white paper included invited authors representing the National Institutes of Health (NIH), the White House Office of Science and Technology Policy (OSTP), the Computer Network Information Center, Chinese Academy of Sciences. The report was published a month ago and has already been viewed nearly 15,000 times. Anyone interested in Open Research and research data sharing should check out the full report, here: the State of Open Data 2022.

But what did we learn about humanities data sharing?

Firstly, we found that just over half of respondents (n=345) believed that “research data” is a term which is applicable to them and their research practice; although while 52% find the term appropriate, nearly half believe that it does not apply to their work either some or all of the time. Other preferred terms suggested by respondents include “research materials,” “information,” “evidence,” and “sources” (figure 1).

Figure 1

 A very high proportion (88%) believed that humanities research data should be shared with others, and a majority had shared their data, though most relied on peer-to-peer sharing methods (76% shared data by email) with only 36% sharing via a data repository, which would provide long term preservation and persistent identifiers for citation, representing best practice for data sharing (figure 2).

Figure 2

More training might be necessary to encourage best practices like repository usage:  80% of respondents stated that they had never received training on data sharing. When asked what additional support publishers could provide, guidance on selecting a suitable data repository was the most popular response (43%); in addition, 41% of respondents would like publishers to collaborate more closely with other stakeholders (institutions and libraries) to provide support for data management and sharing.

There’s a lot more to read in the full white paper, but the headline message is that our humanities authors may be more willing than we imagined to identify their sources and outputs as “data” and to share these with their peers.

The group will be building on this work with a series of interviews with humanities editors (currently underway), and we are aiming to publish a longer report on our findings. Please contact the group chairs if you would like any additional information.