From: "Alma Swan" Date: 25 December 2005 08:03:30 GMT To: "'Stevan Harnad'" Cc: "'Leslie Carr'" , "'Arthur Sale'" , "'David Goodman'" , "'Sheridan Brown'" Subject: RE: Repository usage statistics - user requirements report > It looks very good, though it should not be forgotten that it > is merely an opinion poll. For example, the respondents say, > a priori, that "country" is what they are most interested in: > But why? Is this rational? Is this what a researcher and his > institution is most likely to be interested in? It's what they say they are (not most likely to be) interested in at the moment, yes. Whether it's rational or not is something we can debate, I suppose. > Is it what > will actually end up being used and valued, once OAs and IRs > grow to critical mass? Or are they just an offhand a-priori guess? I can't answer those questions. They are for the future, though I agree that this is an opinion poll. That was, however, what was asked for. The purpose was to find out what they currently say they'd like. My brief was to 'ask them what they want to know about accesses, what trends measures they think are important, about weighting certain things, do they want graphs, and so on'. So that's what I did ask, plus a bit more, and tried to get them to be as specific as possible. I agree that we could add much more richness to this in time, but the project proposal to JISC said this: "Working with partners experienced in analysis of usage statistics for electronic documents, and an international consultative committee of key OAI archive and service managers, the principal deliverables will be: - An API for gathering download data implemented for common IR platforms - A set of agreed standards defining the basis for measuring and reporting usage of materials deposited in IRs and aggregated with data from other sources where such materials can be found." I decided to solicit user views by qualitative methods because, as I pointed out, it is no good asking them about a whole list of things in a survey if they don't know what you're talking about. In some circumstances, and this was one in my judgment, a concept (or details of it) need to be explained and discussed before people can come up with an opinion on how desirable it is, or with refinements of their own. This phase of the work was just to give Les/Arthur/David feedback on what sort of measures researchers want to have about the usage of their articles, as far as they can think about it. The software development can start, writing in all those things the researchers and repository managers say are important (plus more, if we all decided on more measures) and then I'll do another round of polling. > A lot of the responses are like that, though a few of them do > have face validity. (I hope the actual studies and tools made > available will be at least as much based on a rational > analysis of what is needed and why, rather than on prior > opinions as to what might be needed and why... I said in the report that I will go back to this community (in various ways) once some of these measures are built into the software. We can ask them to test it and say what they value and what they don't. We can also add all sorts of other things and ask them about those too (things we think of ourselves - as you have below). There is also a difference between usage measures and impact measures, and it is largely the latter that you have described below, whereas the project is specifically about usage (though we could enrich it if we have time and resources). I agree that impact measures are going to prove much more interesting and useful to researchers than knowing how many people from Taiwan read their latest article last month, but an unstated aim of this present work is that by providing an array of measures that will enable a researcher to discover and monitor how much their work is being looked at it will encourage further deposition and thereby fuller archives. Consequent upon more deposition (a critical mass being reached), the availability of a matching array of new impact measures will further excite them and make the depositing habit a permanent one, one hopes. But Les required an initial 'go out and ask the users what they'd like to know about who's using their articles' phase and that's what's been done. So far. A.