In light of the recent posting by Mitchell Dunkley at DMU, I thought it might be useful to share some of our recent experiences about trying to track down ejournal content. We share what seems to be a similar problem : that of actually finding out what holdings we have – and particularly for ejournals, there are different issues than with ebooks.
Our main point of contact for ejournals data (as opposed to journal titles in databases) has been our subscription agent Swets, and following a recent account meeting with them we flagged several inconsistencies between content available Swetswise Online Content (SWOC) and content available via some publisher’s sites. Swets are still looking into this for us but uncovering some of these problems has raised several issues that I think are generic and the examples below apply across the board. (I have used screen shots from http://screencast-o-matic.com to amplify some of these points – in this the small set of journals happen to be from Oxford Journals.)
1) Differences in holdings between subscription agent and publisher. There seemed to be often a wide variety of conflicting data depending where we looked: for example we found 37 OUP titles on SWOC but only 29 listed on Oxford Journals site. I found downloading information from SWOC problematic and unfriendly – we had to break up a download into several spreadsheets and couldn’t download one spreadsheet for all our holdings.(see this screenshot : http://screencast-o-matic.com/watch/clf0XPCns). We have no idea of how often our subscription agent and the publisher update each other – these conflicts may be a simple mistake, or a reporting error that has lasted for years. Again OUP was only one example, we know of at least two other publishers where this is happening.
2) Publishers approach the problem of ejournal data in different ways. If we turn to the publisher, the Oxford site in this case seems to be structured around a volume issue-based system – which is great for an individual user but access entitlements are shown as being an long HTML through which we had to scroll down http://screencast-o-matic.com/watch/clf0IlCel. It was difficult to work out our holdings start and enddate from this, and as far as I can see an Excel list of holdings was only available on request from Oxford’s help desk.
Not all publisher administrative sites are the same – and in fact access to ejournal holdings may be reported differently depending on whether the publisher is showing holdings via our subscription agents entitlements or via a different account.The package under which a group of titles is accessed or set up may also efffect access – for example we also get Oxford titles through Oxford University Press Archive via JISC http://www.jisc-collections.ac.uk/Catalogue/Overview/Index/1171. Bundles of titles tend to be reported better than individual ongoing ejournal subscriptions.
3) Technical reasons :any discrepancies about content entitlement are often compunded by technical confusion – because of an IP-check the publisher’s site will often say the user is recognised as belonging to the University but then is prompted to login : see http://screencast-o-matic.com/watch/clf022Cek (incidently this is a article we can currently access via SWOC but not via OUP ).This is often compounded when the user logs in off-campus – we have licenses with other publishers where off-campus accesss has not been made available.
4) Every institution has a different subscription history : ‘retrospective’ entitlements to content may be complicated by insititutions not maintaining a print run in the past – broken runs or cancellations can lead to an interruption in electronic access. This similar to the problem that has been mapped by the KB+ project : http://knowledgebaseplus.wordpress.com/2012/03/19/historical_entitlements/ ; and the reality is (like in most things) there is no single source of truth.
5) Free access – is often used by publishers as a marketing tool, which leads to difference between what content the library says is available and what the publisher is actually offering. There is usually no clear statement on how long the offer is for. Publishers vary in how they signal it.
The national work being carried out at KB+ (a JISC project led by @liamearney) is relevant here, but the key question for us is that of scale. When these issues are scaled up per publisher, the inaccuracies can be too resource-intensive to deal with en masse, especially in the light of implementation of a resource discovery system such as Summon. This adds another layer of dependency into the the mix : for example our Elsevier Freedom collection titles also appear in SWOC and we initially found that there are around 390 titles (about 18%) in Summon’s KnowledgeWorks’ definition of the Freedom Collection that don’t appear in our Swets holdings.
It may be that in implementing a resource-discovery system we have to review where we get the data from, and who best to trust. And also be preprared to be flexible. There’s no guarantee. Put up the best that we have, when we have it then take it down later. Journal holdings, like clay pidgeons, never stay still.