Fostering open science

Training for EC project officers on open access and open data in Horizon 2020

We ran four half-day workshops at the end of June as part of the FOSTER project. FOSTER aims to facilitate open science by training researchers about open access and open data. 

The courses were for project officers at the European Commission. They’re the liaison point between researchers and the EC, so our training looked specifically at the requirements under Horizon 2020. It was a really useful learning experience for me too – I came away knowing much more about open access and the support available via OpenAIRE. Below is a summary of the key points and references I took away.

Why make publications and data openly available?

Daniel Spichtinger and Monica Tarazona Rua from the Directorate General for Research and Innovation gave an overview to the EC policy background. They characterised open access as part of broader changes in science, termed ‘science 2.0’. The EC sees a real economic benefit to OA by supporting SMEs and NGOs that can’t afford subscriptions to the latest research. A study by Houghton, Swan and Brown provides quantifiable evidence of how much a lack of OA costs SMEs, both in terms of the time lost accessing documents and the delays to produce new products. 

Daniel and Monica noted that open access is passed the tipping point in several fields (e.g. biology, biomedical research, mathematics and general science & technology) whereas the social sciences, humanities, applied sciences, engineering and technology are the least engaged. These findings are based on a Science Metrix study. Some disciplines have committed to sharing data and are reaping the benefits – the research process is the fastest in High Energy Physics now due to the community practice of sharing immediately. 

The EC ran a pilot on OA under FP7, which is now being followed up with a requirement under Horizon 2020. A pilot for open data has also been introduced with an intention to develop policy in the same way.

Open access in Horizon 2020

Iryna presented some use cases on open access to show potential options and useful services. The diagram below presents the different publishing options, highlighting the gold and green routes to open access.

Birgit spoke about the specific OA requirements under Horizon 2020. The mandate does not intend to restrict publishing in any way. Researchers can publish anywhere they choose. The only requirement is that they make sure the publication is made openly available through a repository. This can be done by:

• publishing with an OA journal (that may, or may not, charge an APC) 
• publishing  with a subscription based journal and choose to deposit a copy into a repository (with open access being usually delayed by an embargo period imposed by the publisher) 
• or, if that option is provided by the publisher, pay an APC to have an immediate open access copy. 

Under Horizon 2020, a copy of the article must always be deposited in a repository, even if the gold route or (hybrid payment) has been chosen.

When researchers first decide where to publish, it’s useful to consult a service like SHERPA RoMEO to see what open access options are available. Birgit suggested that researchers start with a list of targeted journals and prioritise or use a mix and match approach based on the results of this. Although over 60% of publishers don’t charge APCs, the fees can be quite steep. The average rate is €1,020 per article for open access publishers and €1,980 for hybrid journals. (Ref: Björk & Solomon). It could be very costly to always choose the gold route and pay lots of APCs, so a mixture of gold and green approaches is probably best.

The main points to highlight about the Horizon 2020 open access requirements are:

• The requirements apply to peer-reviewed articles rather than monographs, technical reports and conference proceedings, though these can be included as desired.
• All peer-reviewed publications should be made open access via the green or gold routes.
• It is no longer sufficient to make publications available on the project website. Deposit in repositories is required in ALL cases (even under gold OA), so the bibliographic data is open and can be harvested by services like OpenAIRE.
• Currently, the EC does not impose any price cap on fees for publication costs. Researchers should plan OA from the proposal stage and write any APCs into the proposal under the dissemination budget.
• The EC recommends how their funding should be mentioned in publications. This style should be followed to facilitate indexing.

The main document to consult is:
Guidelines on Open Access to Scientific Publications and Research Data in Horizon 2020 

Open data in Horizon 2020

I spoke on the open data pilot in Horizon 2020. The EC has selected seven areas to participate in the pilot, which correspond to about €3 billion or 20% of the overall Horizon 2020 budget in 2014 and 2015. Projects in other areas can take part on a voluntary basis.

The main points to highlight about the Horizon 2020 open access requirements are:

• The pilot applies to two types of data: 1) data needed to validate the results in scientific publications; and 2) other curated and/or raw data as specified in the DMP. It is for researchers to define which ‘other’ data should be made openly available.
• Projects take part in the pilot on an opt in / opt out basis. Opt out is available at any point – at the proposal stage or mid-project. 
• Exemptions can apply at a granular level and do not need to preclude participation. Sensitive, personal or commercial data could be excluded, while other data created by the project remains part of the pilot.

There are four main requirements on projects that participate in the pilot:

1. Develop (and update) a Data Management Plan. This is a deliverable as part of the project – not part of the grant application to be evaluated.
2. Deposit in a research data repository.
3. Make it possible for third parties to access, mine, exploit, reproduce and disseminate data – free of charge for any user.
4. Provide information on the tools and instruments needed to validate the results (or ideally provide the tools if possible).

The main document to consult is:
Guidelines on Data Management in Horizon 2020

OpenAIRE

Eloy spoke about OpenAIRE. This service aggregates data from repositories, cleans it, de-duplicates entries, mines the content and enriches it. Part of the enrichment is about providing links between everything – the publications, data, funders, people, institutions etc. OpenAIRE also provides services and APIs to make the data more useful to others.

I was really impressed with the demonstration of OpenAIRE.  It very clearly demonstrated what can be done with open data to make the lives of researchers, research managers and funders easier. There are services to help researchers locate suitable repositories, to claim publications, link research results and generate citations to standard styles. There are also a number of apps that allow researchers or projects to dynamically incorporate a list of publications on their website, or generate a publications list for EC project reporting. These tools and the statistics provided would clearly be useful to help project officers to monitor compliance too. 

Questions?

There was lots of engagement in every workshop with a pleasing range of questions and discussion. Some of the most common queries were:

• Does the OA mandate not just give our competitors an advantage?
  - What is the position in China, Japan & USA? Is OA a global or European movement only? What are the benefits?
• Which publications does the OA policy apply to?
  - Just peer-reviewed papers? What about other deliverables? And how does it apply when only a small part of the research findings related to the Horizon 2020 funding?
• Does the policy restrict the choice of publisher to open access ones?
  - What if they want to publish in a high-impact journal that doesn’t offer ‘gold’ options? Fear that researchers will publish under another source of funding instead.
• Why are researchers required to deposit in repositories?
  - What is the benefit to doing this if the publication is already open access? Are we not asking them to do the same thing twice?
• How will compliance be monitored?
  - When potential penalties or sanctions are there? Who is responsible for implementation and checking? Are there guidelines and processes for this?

Initial feedback on the training has been positive. We’ll be following up on it with future events and guidance materials. Slides will also be available shortly on the FOSTER website