Related Agendas

Research Information Management projects round two (RIM2) initial analysis and summary

1. Introduction

Arising from the JISC call 15/10 , four projects were funded under the Research Information strand of the JISC 'infrastructure for education and research programme'. The call directed that funded projects should seek to 'expand the community of HE institutions and organisations which are using the CERIF standard'. As a result, the four funded projects were:

CERIFy, UKLON and partners

The CERIFy project sought to map CERIF (Common European Research Information Format) structures onto a selection of non-CERIF institutional CRISs (Current Research Information System). It identified four common strands of activity with its partner institutions:

Pre-award management
Measures of esteem
InCites exchange
Benchmarking

The project produced business process maps for each activity before focusing more closely on measures of esteem and Incites exchange in partnership with Thompson Reuters. Surprisingly, recording esteem factors within a CERIF environment revealed unexpected synergies with the MICE project (see below).

As a result of their process mapping, the project came to the conclusion that while there are variations in RIM (Research Information Management) processes between institutions there is also a very high degree of commonality which supports the adoption of a common data standard and, more controversially, possible common business models.

BRUCE (Brunel Research Under a CERIF Environment), Brunel University and partners

The BRUCE project sought to map data supplied to the Higher Education Statistics Agency (HESA), for statutory reporting, to CERIF and develop a prototype reporting tool to provide management information for Research Information Managers.

The project had intended that the reporting tool would be developed in the open source Blacklight discovery interface but encountered very significant difficulties in establishing a stable Blacklight platform and eventually chose an alternative approach involving a purpose built Solr development. Using Solr proved beneficial and allowed the project to develop better searching facilities than those available in Blacklight. The resulting tool is now available for interested parties to download.

IRIOS (Integrated Research Input and Output System), University of Sunderland and partners

The IRIOS project designed and implemented a demonstrator system for exchanging Research Council (RC) data in CERIF format. It is widely agreed that one of the key benefits of a common standard for RIM will be to streamline communication between RCs, other statutory bodies and HEIs. The project aimed to produce a system capable of:

Facilitating the import of RC grant data
Allowing registered users to search and browse the data
Allowing selected data to be exported in CERIF format, for subsequent use by CERIF compliant systems at the partner HEIs
Allowing HEIs to link their institutional repository output data to the RC project data

In order to achieve the project goals within the timescales the project utilised the already existing system infrastructure of the Universities for the North East Information System (UNIS).

MICE (Measuring Impact Under CERIF), King's College London and partners

By addressing the problem of recording research impact systemically the MICE project addressed one of the major RIM related issues currently facing Higher Education. The Research Excellence Framework (REF) will require institutions to report on impact and a pilot, run by HEFCE in 2010, highlighted the lack of information systems to hold impact data and processes for gathering it.

By developing a CERIF based structure for holding impact data the project started to address the systems aspect of collecting, holding and reporting on research impact. By including a major CRIS supplier (Atira) as a project partner, MICE substantially increased the possibility of the outputs of the project being incorporated into commercial products and so passing into general use.

The project also highlighted the need for a common taxonomy for impact and made a start in this debate.

2. Common Themes

A number of common themes emerged from the project reports. As all the projects interfaced with CERIF in some form, unsurprisingly some of the major common themes were around the CERIF standard and EuroCRIS.

2.1 Accessibility of CERIF

A common thread through all four of the project reports was the accessibility of CERIF. The point was repeatedly made that though the existence and importance of CERIF is widely known in the sector, the detail of the standard and its structure is not. All the projects reported that they had experienced a significant learning curve with CERIF and all had engaged with EuroCRIS at some time in the project to either discuss detailed aspects of the data model or improve their understanding of it.

The point was made that CERIF is a highly complex data model and well outside the area of expertise of most research administrators. This in turn results in organisations specifying that their CRIS has to be CERIF compliant but not being aware of the potential of the standard or how best to leverage the potential advantages that having a CERIF compliant CRIS offers.

All the projects made the point that there is a need for readily accessible educational material on CERIF. It is possible that this could be in the form of online material as well as more traditional delivery and that the content should be developed by JISC and EuroCRIS. A point made implicitly by the projects is that there are different audiences for CERIF education; for example the needs of a research administrator are significantly different to a systems developer but both need to understand the potential offered by the data model.

2.2 Development of the CERIF Model

All four of the projects analysed the CERIF model to assess whether it was capable of holding the information they required. One (IRIOS) concluded that the model 'as is' was capable of holding their data set.

The other three projects all proposed extensions to the existing CERIF 2008 model. By good use of the CERIF semantic layer UKLON were able to propose that the extensions that they were requiring for 'esteem' measures could be superimposed on the extensions that King's College were proposing for 'impact' indicators and a more generic model developed. The BRUCE project proposed some structural changes to CERIF in order to map the data model onto their Solr, Apache based indexing tool.

2.3 Exchanging CERIF Data

Three of the projects actively exchanged data and although experiencing some issues all succeeded in effecting the exchanges required. Interestingly the transport mechanisms were all different:

" The BRUCE project was feeding the information into their Solr tool for indexing and searching via the Solr import handler. The project had to make some minor compromises with the CERIF structure to achieve this but nevertheless achieved good results
For the purposes of the IRIOS demonstrator the project elected to load information from Microsoft Excel
CERIFy generated CERIF-XML for its interactions with the InCites service and successfully created records for up to fifty researchers and several hundred research outputs

2.4 Optimising existing technology

All four projects placed an emphasis on utilising existing information infrastructure and each project made successful and cost effective use of existing systems and data sources to produce and exchange CERIF compatible information with other HEIs, Research Councils and commercial organisations. The CRISs of the originating and receiving systems were a mixture of CERIF and non-CERIF systems demonstrating that CERIF can be successfully utilised without the need to discard or redevelop any existing infrastructure.

3. Conclusion

All four projects successfully demonstrated that CERIF is a viable and stable mechanism for RIM data exchange and that the exchange can be with any type of organisation within the RIM domain. This reinforces the conclusions of earlier projects such as CRISpool and significantly advances the sector's understanding of the strengths of CERIF and its issues. The range of activities covered by the projects demonstrated both the strength and versatility of the CERIF model and confirms that it is fit for purpose in UK Higher Education research information management. However all projects experienced some difficulty in understanding the CERIF model and the current lack of in-depth understanding of the model is a significant barrier to further progress in embedding the standard. For the advantages of CERIF to be fully leveraged, understanding of the model must be significantly improved. JISC, Eurocris and other interested bodies need to develop training courses and workshops. Suppliers of CRISs should consider whether they can support a CERIF 'app store' where institutions can deposit and readily access solutions to common needs.

The projects have also demonstrated that the similarities of processes and systems within HEIs substantially outweighs the differences, and this should lead to a greater willingness by institutions to invest in joint development ventures.

4. Recommendations

For JISC

JISC should consult with HEIs, Research Councils and statutory customers to develop a clear understanding of the CERIF training needs of the sector
Having established the CERIF training needs JISC in conjunction with EuroCRIS should facilitate the development of appropriate learning materials
JISC should consider how best to share the experience already gained by the RIM programme with the wider sector (possibly as part of the upcoming study of the impact of the RIM programme)
JISC (via the Research Information Management Group and other avenues) should seek to increase engagement with CRIS suppliers, bibliographic service suppliers etc, with a view to increasing commercial compliance with the CERIF standard
JISC should consult with King's College London to agree how best a definitive taxonomy of impact indicators can be developed

For institutions

Once a structure is in place to develop a definitive taxonomy for REF impact indicators, institutions should actively engage with the process in order to develop the best possible framework
Irrespective of their internal RIM systems, institutions should adopt CERIF as their default standard for RIM data interchange
Institutions should seek to expand their understanding of the CERIF model with a view to realising greater benefits from their investment in the CERIF standard
Given the high degree of common business processes institutions should actively seek to develop multilateral partnerships for the development of RIM facilities

5. Further information

For further advice and guidance on research information management based on previous JISC-funded RIM work and related resources of value to this agenda please go to www.jiscinfonet.ac.uk/research.