Have been working on research data management in context of the whole research data lifecycle. Started asking question: once research data management is under control, what will be the next focus? Their answer was research tools. Produced two journal articles:
- Wolski, M., Howard, L., & Richardson, J. (2017). The importance of tools in the data lifecycle. Digital Library Perspectives, 33(3), in press
- Wolski, M., Howard, L., & Richardson, J. (2017). A trust framework for online research data services. Publications, 5(2), article 14 https://doi.org/10.3390/publications5020014
Research life cycle: Data creation and deposit (plan and design, collect and capture) -> Managing active data (Collect and capture, collaborate and analyse) -> Data repositories and archives (manage, store, preserve; share and publish) -> Data catalogues and registries
Research data repositories vary a lot. Collection or ecosystem? Open or closed? End point or part of workflow? Why is it hard to build them? Push-and-pull between re-usability and preservation:
- technical aspects
- interoperability
- lega/regulatory/ethical constraints
- one-off activity or continuous
- diversity of accessibility issues
- diversity of re-usability issues
The average number of research tools per person was 22 per person (includes Word, ResearchGate, email through to SurveyMonkey, Dropbox, Figshare, through to R and really specialised ones). Kramer and Bosman (2016) divided tools into assessment, outreach, publication, writing, analysis, discovery, preparation phases. Tools exploding as research activity scales up, collaboration increases. Large-capacity projects being funded. Data science courses upskilling researchers.
Researchers use lots of tools as part of the data workflow. The institution may manage data, but have no ownership of workflow. Since data has to move seamlessly between tools, interoperability is key – but how do we built these interoperable workflows and infrastructures?
Need to remember repository is only part of the research ecosystem. Need to take an institutional approach – or approaches rather than a single design solution. Look at main workflows and tools used – check out research communities who may already have the solutions – focus must be meeting the researchers’ needs.
Q: Will we see researchers use fewer tools as disciplinary workflows develop?A: Probably not but will see more integration between them eg Qualtrics adding an R connector.
Have a full-text repository on DSpace, building data repository on CKAN. What if we build something (at great expense) and they don’t come? We need cultural change. Eg UK seems far ahead but only 16% of respondents are accessing university RDM support services in 2016.
They have data repository, and provide support service by research office, library and IT services.
Research offices provides support in grant writing; advocates on policies; helps with internal research funding; report to senior leadership. Their toolkit:
- need to win research managers over – explain how important it is
- embedded an RDM-expert
- upskilled research office staff about data management planning and how to make a case for data management.
Look out for game changers:
- eg large collaborative research projects – produce lots of data and need to share it to be successful so more likely to listen
- DMP preview as standard procedure for proposal review and training on proposal writing. (Want data management planning to be like brushing your teeth: you do it every day and if you forget you can’t sleep.)
- adapt incentives – eg internal funding for early career researchers requires data management plans
- use existing networks – researchers go to lots of boards and meetings already so feed this as a topic like any other topic
- engage with members of DFG[German science foundation] review board – to get them to draw up criteria to reward researchers doing it
Cultural change towards open science can be supported by your research office. Let’s team up more!
RIMS – include ResearchGate, Academia, Google Scholar; ORCID, ImpactStory; PURE, Elements
ResearchGate sends out a flood of emails – good for some, a put-off for others. How can we improve our RIMS to improve researcher engagement?
Interviewed 15 researchers then expanded to survey 412 participants; also analysed metadata on 126 ResearchGate profiles of participants. Preliminary findings:
- Variety of different researcher activities in RIMS eg write manuscripts, interact with peers, curate, evaluate, look for jobs, monitor literature, identify collaborators, disseminate research, find relevant literature.
- Different levels of participation: readers may have a profile but don’t maintain it or interact with people; record managers maintain their profile, but don’t interact with others; community members maintain profiles but also interact with others etc.
- Different motivations to maintain profile: to share scholarship (most popular); improve status, enjoyment, support evaluation, quality of recommendations, external pressure (least popular)
- Different use of metadata categories: people tend to use the person, publication, and research subject catories. Maybe research experience, but rarely education, award, teaching experience, other other.
- In Person most people put in first, last name, affiliation, dept;
- Publication: Most use most of these except only 30% of readers share the file – about 80% of record managers and community member
Want to develop design recommendations to enable RIMS to increase participation.