July is proving to be a fast and furious month in terms of presenting on SALT, crunching the remainder of the gargantuan amount of data we’re receiving from our partners at JRUL, finessing the API, implementing the developments in the Copac prototype (and showing it off to people at the CILIP Umbrella conference), and running various workshops and user testing sessions to test our hypothesis (as much as we can in this time frame) and see whether the ‘shared service’ side of all this might actually scale.
We’re working intensely, but reflecting too. The workshop hosted by the RISE project earlier this month was an excellent opportunity to step back and see the bigger picture, and reflect on the benefits we’re aiming to realise for libraries and their users. A reminder of what we’re attempting to find out:
Library circulation activity data can be used to support humanities research by surfacing underused ‘long tail’ library materials through search
Also… how sustainable would an API-based national shared service be?
And can such a service support users and also library workflows such as collections management?
What we already know
We know that arts and humanities students and academics borrow books.
Research conducted in-house by Mimas, and also by others (for example Carole Palmer ) also highlights the differences in search methodologies between this demographic and their STEM counterparts. In short, humanities researchers tend to search centrifugally, ‘berry picking’ from various trails. Mimas’ recent research with Mindset and Curtis and Cartwright indicates that newer postgraduates tend to work in quite an isolated way – asking few if any for advice on where to search (supervisors feature heavily in this regard, whereas subject librarians do not at all) and sticking with a few ‘known’ resources). While these users are typically suspicious of the idea that allowing other users to annotate, tag, or rate items would be of benefit to them, there is generally a positive response when asked about the usefulness of a recommender function; in fact Amazon is used significantly in this regard to help users find related materials that are not surfacing through a ‘traditional’ library search.
Library recommendation systems are already achieving benefits for undergraduates – University of Huddersfield being the obvious example. Indeed, Huddersfield’s system helps students move beyond the ‘nose’ to the long tail of library collections – and there is some evidence that new borrowing patterns are emerging, with students taking out books from outside what is assigned. In humanities research especially, the long tail is obviously much more relevant.
So what additional benefits might we realise through this work, especially if we move on t
o aggregating data from additional libraries?
- Key to our hypothesis is the belief that such systems can help surface and hopefully increase the usage of hidden collections. Obviously circulation data is only going to offer a partial solution to this problem of discoverability (i.e. many ‘hidden gems’ are of course non-circulating) but nonetheless, we believethat the long tail argument borne out by Chris Anderson can also hold true for libraries – that the collective ‘share’ or recommendation of items can turn the Pareto Principle on its head. For libraries this means being able to demonstrate the value of the collections by pointing to increased usage. It might also give libraries a better sense of what is of value to users, and what perhaps is not.
- For users, particularly those in the humanities, a recommender function can help providing new routes to discovery based on use and disciplinary contexts (not traditional classification). In other words, what you areviewing through ‘recommenders’ are patterns of real usage, how other users with similar academic interests are aggregating texts. This is particularly useful for finding conceptually related groupings of texts that cut across differentdisciplines, and which will not ordinarily sit together in a standard results set.
- It also means we can support humanities users in their preferred mode of discovery, powering ‘centrifugal searching’ and discovery through serendipity. The downstream benefits of this concern the emergence of new, original research, new knowledge and ideas.
Last week we sat down with a group of collections managers from JRUL as well as Leeds University and talked with them about other possible benefits related to library workflows which we weren’t yet seeing. Here are the potential benefits we came up with:
- Aggregated activity data could support activities such as stock weeding by revealing collection strengths and allowing librarians to cross check against other collections.
- By combining aggregated collection data and aggregated activity data, librarians will see a fuller picture. This means they can identify collection strengths and recognise items that should be retained because of association with valued collections. We thought about this as a form of “stock management by association.” Librarians might treat some long-tail items (e.g. items with limited borrowing) with caution if they were aware of links/associations to other collections (although there is also the caveat that this wouldn’t be possible with local activity data reports in isolation)
- Aggregated activity data could have benefits for collection development. Seeing the national picture might allow librarians to identify related items – “if your collection has this, it should also have…”
- This could also inform the number of copies a library should buy, and which books from reading lists are required in multiple copies.
- Thinking more outside the box, we thought it might also inform digitization decision-making – i.e. if you digitized this, you might also consider digitizing…
- Aggregated activity data could inform stock purchase – allow librarians to see what’s new, what’s being used elsewhere and therefore what’s worth buying.
- This could also have benefits when discussing reading lists and stock purchases with academic staff, and thus enhance academic engagement
Over the next couple of weeks we’ll have quite a bit more to report as we analyse sustainability issues with other libraries, and perhaps most importantly, put the recommender itself in front of postgraduate humanities researchers to see if our hypothesis is likely to be proven true — at least based on what they tell us (the true test will happen over the next year as we monitor impact via JRUL and Copac).