Ron Sandland, ANDS’ Steering Committee Chair
Evaluating the impact of research has been a thorny problem for some time. There have been a number of surrogate measures suggested.
One is the revenue generated from collaborative research with industry. In 1987, the CSIRO was required to adopt an arbitrary target of generating 30 percent of revenues from external sources. This led to some significant unintended consequences, as those of us who were in CSIRO at the time can attest.
The most striking of these was that significantly more than 30 percent of the organisation’s effort went into meeting these targets, which were difficult to achieve in some areas with underdeveloped industry bases. The result was a reduction in focus on the longer term problems of national significance that had characterised CSIRO’s mission-driven research from its earliest days. This was addressed by the national flagships initiative commenced in 2003.
Another approach was counting the number of citations of research papers and the number of papers in journals with the highest impact factors. The problems with metrics of this kind are manifold; in particular they favour certain areas of research over others (for example, medicine versus the mathematical sciences), and they channel efforts into areas of research that are more likely to lead to publications rather than longer-term research that may lead to important societal outcomes.
Research data, a critically important output of research, provides a natural avenue for multiplying the impact of research per se.
The ability to effectively store, manage, transform and share data provides a significant avenue for research collaboration both nationally and internationally. It also provides a major route for research-industry collaboration (which needs to be a two-way street), and the substrate for collaboration to address urgent national and international problems.
The impact of research data is multiplied when it is interoperable, shareable and properly prioritised and managed. It is gratifying to see that this has been explicitly recognised in the National Research Infrastructure Roadmap released late last year. It is difficult to imagine a more constructive step that could be taken to enhance the impact of research data and indeed of the whole research enterprise.