In fact, to our surprise, the majority of group members were still at an early stage: some considering a CRIS/repository link, some planning to expand an existing repository to offer CRIS type functions and others just interested in the topic but with no current link plans. Although the CRIS/Repository model is becoming more common, institutions that have followed the process through and achieved full integration between the two systems are still few and far between.
Quickly shifting gears from the initial breakout group plan, we discussed some of the potential benefits and challenges of the new model.
Findings are below (transcribed from flip charts and added commentary):
- Potentially greater deposit: whether this is true or not depends on where you are starting from with your existing repository. Some are well embedded, but others have struggled to become part of everyday researcher workflows.
- One stop shop: single place of deposit but also a way to draw together many strands of research information. A CRIS can be enhanced by an OA platform and the high standards of data curation which come with it; the repository can be complemented by the administrative data in the CRIS.
- CRIS+repository may be a good model to support researcher compliance with funder OA and reporting requirements.
- CRIS benefits from repository visibility – research becomes more discoverable.
- Web page feeds may include publication lists with links to repository content - but also grants, expertise, activities, impact etc.
- Repository usage stats could be fed back to the CRIS. As well as usage, stats could show non-OA-depositors what traffic they’re missing.
- OA takes a back seat.
- Academics don’t care about the depositing system – it’s just another admin system to them. Maybe this doesn’t matter. And it’s not an issue that’s limited to the CRIS+repository model. But perhaps academics are less likely to engage with OA aspects of a CRIS if they don’t see the relevance to their own subject discipline and research.
- REF – a useful driver - but too much REF focus could lead to fewer OA deposits and more limited engagement with repository systems.
- Why have two sets of metadata? Is the repository just a file store? Does it matter?
- Data quality – building the publication database within a CRIS tends to involve importing data from a number of different sources. E.g. Thomson Web of Science, departmental databases, individual publication lists in EndNote, BiBTeX etc. Inevitably there is duplication and a range of data quality issues. Is it worth tidying the records up? Who does this? Is there any resource to do this? Is surfacing publication data on researcher web pages sufficient incentive for them to rectify any issues with their own data?
- The model helps with research publication and research data curation – funder data is tied in with compliance requirements, depositors are advised on these & there is automatic deposit or notification to required external subject / data repositories.
- Effective data exchange between systems & common data standards – probably CERIF.
- Crosswalks between systems are easy to set up and readily tailorable.
- Relevant support departments work together to create an effective system (Research Office, Library, IT, Staff Training). Effective governance mechanisms are put in place.
- Uptake by some researchers exerts peer pressure on others, raising overall take up.
So long as you have an effective system to deposit, describe, disseminate and preserve you research information, it may not matter too much what the underlying architecture looks like. However, there are many practical issues to be tackled – particularly if you have pre-existing systems which must be linked or phased out – when introducing a CRIS+repository architecture.
Many thanks to the attendees as the comments provided the project with insights that will be written up in the final project report.
submitted by: Rachel Proudfoot and Jodie Double