Ready to draft your institution’s RDM guidelines but not sure where to start?

But those efforts are already bearing fruit: “Funder mandates and government strategies are driving research institutions to engage with RDM – it has become a big topic for us. And institutions play an important role in translating those requirements for their researchers,” explains Julian. “We knew that developing the right guidelines for ETH Zurich would help us provide our researchers with the support they need to achieve RDM best practice. And importantly we’ve managed to create guidelines that aren’t primarily there for researchers to comply with – they also enable them.” 

Ensuring that the document can serve that dual function has required the team to provide clear information on how researchers can best publish, archive and preserve their research data – all critical steps in RDM. Advice includes: 

1. Researchers should include RDM in the planning of their activities.  

To help with this, the guidelines define and support RDM as a scientific activity. Julian explains: “It’s not just plugged on top of research; we really see it as an integral part of our research work.”  

2. A data management plan with clear timelines is expected for every research project.  

Plans should outline what data researchers plan to collect, and how they will be managed, shared and stored (both during the project and after). Planning RDM ahead of time is crucial to enable, for example, efficient research workflows and data publishing in line with the FAIR data principles.  

3. Research data must be published in a FAIR repository, generally at the time of publication of results. 

The FAIR data principles(opens in new tab/window) require digital assets, such as research data, to be Findable, Accessible, Interoperable and Reusable. While, in some instances, FAIR data is also ‘open data’ that isn’t always the case; for example, when there are legal or ethical restrictions. Julian adds: “The FAIR requirement isn’t only something that ETH Zurich as an institution has agreed upon; funders require it as well. Importantly, deciding which data are directly relevant for a specific publication is subject to established community standards.”  

4. All publications of results must include a Data Availability Statement. 

These statements tell the reader whether the research data associated with a paper are available and, if so, where and how they can be accessed. According to Julian, “If there’s a good reason why the researchers can’t openly share the data, they can use the statement to explain that too.”  

5. Research data must be retained for an extended period - generally a minimum of 10 years. 

This guideline applies to data whether they are stored in the university’s systems or published in an external FAIR data repository. If researchers opt for a third-party repository, they need to ensure it can comply with the required timeframe. The 10-year minimum storage period also extends to unpublished research data at ETH Zurich. 

 6. Project members should determine as early as possible how data will be shared externally and how they can be used by researchers leaving the project team. 

Julian explains: “This last point is really important to enable researchers who move from one institution to another to continue working with their precious research data if they haven’t fully exploited or published them yet.”  

> View the ETH Zurich RDM guidelines in full(opens in new tab/window)

Creating a shared understanding is an important first step

The process for the creation and discussion of the guidelines was managed by the Office of Research on behalf of the Executive Board. Several units at the university then contributed to the drafting phase. An agreement was reached on who would be responsible for RDM support at ETH Zurich – in this case, the Library and the Scientific IT Services in close collaboration with further units – and the role that researchers and their group leaders would play. And they established the vocabulary that would be used to refer to RDM and open research data.  

They also quickly determined that you don’t always need to look outside your institution for examples of best practice. Julian explains: “At ETH Zurich we see certain research groups who are already at the forefront of embracing good practice principles in RDM and we also have established RDM support services. The Office of Research drew on these experiences when drafting the guidelines.”  

He adds: “We were clear from the start that good RDM practice should build on existing community standards. A large technical university like ours has diverse disciplines. These come with the challenge that the data they generate are heterogenous and so are their community practices; for example, there are disciplines which share a simulation model, rather than the terabytes of data created in a model. With no one-size-fits-all solution available, it is crucial to find where that community knowledge is located and lift its potential.”  

But for Julian and his colleagues, formal requirements such as the RDM guidelines ETH Zurich has developed are only one side of the coin. “Incentives and continuous cultural change are also important. And you need institutional commitment and an ability to provide the right infrastructure and support if you want to get this right.”