Scientific Data and Nature host an event that explores how different stake holders can collaborate with researchers to publish better science through better data management.
Data, without a doubt, are the foundation of science. If you’re a researcher, your life is data: you spend your days generating it, analysing it, and writing papers about it. You share it with colleagues and collaborate on projects that will build on it and find new and exciting things. But policy makers, funders and universities are also involved in the conversation – each trying to solve the problem of managing the increasing influx of data whilst keeping the integrity of science high.
Last Friday, PhD students and postdoctoral researchers came to the Nature offices to learn about how research data affects a scientist’s ability to publish and get research funding. The event, Publishing Better Science through Better Data, consisted of a series of talks from editors, data curators, software developers and funding body representatives, all giving their perspective on how data affects scientific research and publishing.
The editor’s perspective
Philip Campbell, Editor-in-Chief of Nature and its sister journals provided an editor’s perspective and shared how Nature journal was handling the reproducibility problem: “It mostly consists of things that are bad or sloppy science, not fraud.” To minimize the amount of “sloppy science” being published in Nature, editors have put a check-list in place for scientists that they submit along with their papers, making the research process more transparent. “It’s improving the reliability of the design of experiments, which is what we want to see happening.”
Dr Monica Contestabile, Senior Editor at Nature Climate Change gave an interdisciplinary editor’s perspective on the importance of good data collection using real examples from the journal’s peer-review process, highlighting just how important that check0list can be. “Often we see papers that are editorially very interesting… and we send them to referees. They pass them back to us because of unsuitable data,” she said. Her advice: full transparency on data sources, collection, sampling and manipulation methods upon submission goes a long way.
In Scientific Data, scientists can publish detailed descriptions of their data and get credit for their work, said Andrew Hufton, the Managing Editor. He made the point that before data can be shared, reused and built upon, it needs to be well described. So he shared some practical advice on when, where and how to submit your research data for publication.
The funder’s perspective
David Carr, Policy Advisor at the Wellcome Trust, highlighted how funders are doing their part. This includes asking scientists to create data management plans as part of their grant applications, which clearly indicate how data will be collected, managed, curated and shared. But they also do a lot of research into the impact of open-access data, and the cost-benefits that it brings in the hope of creating more beneficial policies in the future. What Carr highlighted was that all funders are struggling to meet the challenge of building resources and culture to support data sharing.
The data curator’s perspective
Sally Rumsey from The Bodleian Libraries at the University of Oxford, showed that much support is available for scientists within libraries, especially when it comes to answering questions about data storage and using institutional repositories.
Susanna-Assunta Sansone gave a passionate explanation of how to submit to Scientific Data, what a Data Descriptor (Scientific Data’s primary article type) really is and what the background editorial processes are. “A Data Descriptor is the narrative component that complements the data. It’s not a hypothesis or a conclusion,” she said. “It’s meta: it describes data.” She emphasised the importance of good metadata, as this means your data can be reused and easily discovered.
Veerle Van den Eynden, a Research Data Manager at the UK Data Service, shared her experience and guidance on dealing with sensitive data. This is particularly important for those working in the social sciences and in health care, where mountains of personal data are collected. The UK Data Service, referred to by several of the speakers, provides access to data resources to facilitate social and economic research and education.
The software and tools perspective
But there are others, as Figshare’s (an online data repository for private and open sharing) founder, Mark Hahnel told the story of Figshare’s creation out of his own frustration at not being able to share the majority of his research: videos. He wants researchers to get the most out of their data by sharing it and getting ahead of the trend. He said that the reason people share their data is because can lead to improve their research integrity, raise their profiles and to increase their citation rate.
Overall, the resounding conclusion of the event was that good data management and sharing is important, but there aren’t enough resources – or awareness of available resources – for this to happen efficiently. Several bodies (libraries, universities, funding bodies and publishers) can all chip in and work together to make this easier for the researchers to create solid data, and hence, solid research.
Due to high levels of demand for tickets, the event was filmed and each speaking session will shortly be available to view online. More to come soon!