All entries for Tuesday 16 September 2008
September 16, 2008
I went to an interesting event at Cranfield last week. It's a pre-cursor to one that Miggie Pickton of NECTAR and I are planning, to be held in December at Northampton, both on themes of advocacy. The star presenter last Tuesday, in my opinion, was Dr Bruce Jefferson, an academic at Cranfield who has done a lot of investigation into what it takes to become highly cited. In his opinion, repositories will make little difference to citation levels, and repository managers should find evidence to prove any claims that are made about raising profiles and citations. I've also been frustrated by the lack of evidence of the difference repositories make to citations, in spite of the claims: the idea is so widely used that I thought it must be easy to find evidence, when I began on this project. But the only research done has been to prove what a difference open access makes, and only the other week we heard news of a BioMed science project which proved that open access made no difference whatsoever to citation rates.
Dr Jefferson will be conducting his own experiment, putting half of his work into the repository and half not, being careful to make sure that each half has a similar weighting in terms of current popularity and weighting. Even the results of his experiment will not be empirical, of course. But they might help to prove whether repositories make any difference to citations.
Dr Jefferson also said that there are other values to a repository, than potentially raising citations, and whilst we lack evidence about citations, perhaps we ought to focus more on those.
And the reason I titled this posting "Muddying the waters" was that he had definite views on the inclusion of anything other than published, peer-reviewed journal articles. His view was "don't". Academics will only deposit their best work if they believe it will be shown amongst a collection of the best work. Academics do not want theses they have supervised to be associated with their name in institutional repositories. Nor do they want unpublished reports or research data to be housed alongside their polished articles. This assumes that people will discover repository content from within the repository itself, rather than through a web search engine. We ought to be very clear in our message to our authors, that the repository is not a destination in itself, but more of a warehouse from which people can order content to appear in their browser.
If academics are driven by citations, and if we aim to improve citation ratings, surely the repository ought to only focus on content from publications that are indexed by Web of Science, since these are the journals in which citations are counted.
So, how pure or how muddy should an institutional repository be? I personally like the idea of having separate repositories for different types of content. They can all be cross searched at an institutional level or by search engines anyway, and in separate repository instances you can use different metadata schema to suit the needs of your content and searchers. I used Google Scholar's registration form and they wanted to know that the content of the site being registered was published journal articles. So if our repository also had data sets and learning objects, presumably I would not be able to register with Google Scholar. Although I'm unsure yet whether my filling out that form has done us any good...
It seems to me that repositories that are being established to achieve all sorts of purposes are in danger of achieving none of them. We're primarily focussed on journal articles and PhD theses, because the project aims said we would be and we don't have the resource to diversify from this if we are to achieve our targets, but even we have some reports and working papers as well, at the request of our academics. The reports in particular do not have a natural home online anywhere else. But I'm not sure yet what to recommend as a repository content policy at the end of this project.