10 commandments in rdm funder compliancy
0 likes61 views
The document outlines ten commandments for research data management (RDM) compliance, emphasizing the importance of data management plans (DMPs) in scientific research. Key recommendations include collaborating with institutional support, clearly describing data types, ensuring data is findable and shareable, and establishing ethical guidelines. It encourages transparency and practical documentation practices to enhance the quality and compliance of research proposals.
10 commandments in rdm funder compliancy
- 1. 10 commandments in RDM funder compliancy Hannelore Vanhaverbeke KU Leuven https://orcid.org/0000-0002-8748-8808 A pragmatic approach
- 2. Tell me why? • What funders say: Data management is an integral part of sound scientific research Data management plans (DMPs) are a key element of good data management • What you should tell yourself: Data management is an integral part of sound scientific research Data management plans (DMPs) are a key element of good data management
- 4. 1. Don’t wait for the manna, eh… money • Go through all sections of a DMP template when writing your proposal • use DMPonline.(kuleuven).be, its templates cover all questions and offers guidance • You will • write a better section on data management in the proposal • detect possible costs that can be eligible for funding
- 5. 2. Do not wander for 40 years (or so) • Know whom in your institution can help with data questions + mention these persons/services in your application/reporting • IT, Library, Research Office, …
- 6. 3. Don’t live on a prayer • Designate a coordinator for all RDM stuff + say so in the application • esp. in a consortium • min. senior postdoc level • Tasks: • coordinates writing the initial DMP • monitors whether updates to the DMP are called for • coordinates writing the final DMP • ensures uniformity or at least minimal harmonization in practices (documentation, back up, … (I would advice to do the same for Open Access)
- 7. 4. Split hairs (not the sea) • Know precisely what kind of data you will use/produce (mind you: ‘data’ = very broad) • The data description section is the most vital part of a DMP • Make a list of all ‘input’ and ‘output’ data per WP or objective and aggregate • per WP/objective • data type • Reviewers love tabular forms & bullet style lists • Incomplete/inaccurate overview = incomplete/inaccurate DMP
- 8. 5. Where there’s smoke… Mind burning bushes • Special data categories • Everything to do with humans (participants, interviewees, bits & pieces of humans) • Everything to do with possible IP (claimed by you or imposed by 3rd parties) • Dual use issues • For some you cannot start research before legal/ethical approval • privacy registry/ethical committees • For some of these categories you cannot just choose what you do with the data (publishing, storing, …): read agreements carefully!
- 9. 6. Avoid plagues • Storage & back up: best use managed services (managed by your institution) • “fool proof”: back up is automatic • someone can retrieve the data • many services are ‘GDPR’ proof: encryption, controlled access etc Own devices: responsibility is yours cost vs quality – when budgeted cost for storage are mostly eligible • Documentation, uniform description
- 10. 7. Be FAIR • Findable – Accessible – Interoperable - Reusable • FAIR is not Open • FAIR also pertains to data you wish to reuse in the future • FAIR is not to be ‘imposed’ at the end of the ride but from the beginning • https://www.go-fair.org/fair-principles/
- 11. 8. Do not speak in tongues • Never ever write things you do not understand • Never ever promise something you cannot guarantee • rather say “at this moment we consider the use of…”, “if in the course of research this changes we will update the DMP”, “in case of IP, TTO will be contacted”, … • Know about metadata (= “tongues”): metadata creation/application of standards is the most difficult in RDM • There are no standards • It is not clear how to implement these in research workflow
- 12. 9. Go & multiply • There are a limited nr of reasons for not sharing (privacy, contractual agreements, IP, security issues) In all other cases, consider sharing: “shared data live longer” • Clearly specify how data will be shared & under what licence (CC-0?) • Make use of good data repositories (license, PID, ...) • Work is done for you (in terms of preservation & discovery) • With good metadata your data will be found & possibly reused => data citation • General repositories: Harvard's Dataverse ; EUDAT's B2Share ; Figshare ; Zenodo ; Open Science Framework • Discipline-specific repositories as advised by Nature, PLoS, F1000
- 13. 10. Lead the way • preserve DMPs and share them; they can be useful to others • within labs or research units: draw up a ‘generic DMP’
- 14. Self-assessment • DCC H2020 compliance checklist • EPSRC self-assessment rubric