Usage reporting (was: Seven ARL Libraries)

[Phil Davis <pmd8@cornell.edu>]

Chris Beckett wrote: "It would be interesting to understand how 
these two contrasting positions [of reporting journal usage 
statistics] can be reconciled."

Both the Research Information Network and the Rockefeller 
University reports seem correct.  The RIN report is based upon 
measuring *rare and infrequent* downloading events, whereas the 
Rockefeller report is based on the *concentration* of such 
events.

The correct question would ask which type of reporting is a more 
accurate representation of library (or institutional) value. 
And this is where things get a little more complicated, because 
each type of reporting reflects a different emphasis of value.

The concentration model implies that overall value for an 
institution can be maximized at a given expense.  Plot the 
cumulative downloads against cumulative expense and you have a 
graph that can tell you how much value (i.e. downloads or 
journals) can be purchased for a given price.  The graph 
illustrates the law of diminishing returns -- more value comes 
with acceleratingly expenses.

The rare-and-infrequent events model supports the same thinking 
that goes behind comprehensive collection development, that 
libraries should be spending money on just-in-case models because 
information requests cannot be fully predicted and libraries 
should support esoteric research.  There are many resources at 
libraries which are used even less frequently than one-in-four 
months (although I do question the arbitrary selection of 
4-months).

So both models are truthful, although they only tell a small part 
of the overall story of usage and value.  The concentration view 
looks only at the head of the distribution; the rare and 
infrequent view only looks at the long tail.  A cynic would say 
that people make up their minds first and then select the right 
statistics to report.  A pragmatist would say that both views 
should be reported.

--Phil Davis

Chris' full post:

> Its is interesting that the RLIN study cited in the RUP 
> editorial showed that:
>
> "Far more importantly, these big deals give university 
> researchers access to unprecedented numbers of titles. And the 
> evidence shows that they are making good use of this: studies 
> for JISC and others have shown heavy use of journals to which 
> libraries did not formerly subscribe. A recent study for the 
> Research Information Network found that articles from 99% of 
> the titles available in a range of university libraries were 
> downloaded over a four-month period."
>
> While at the same time the evidence cited in the editorial from 
> the Rockefeller University Library showed that:
>
> "For one of the bundles, the top 10% of journals garner over 
> 85% of the hits to the bundle from users at the University. 
> Over 40% of the journals in the bundle had no hits at all from 
> the University in 2008!"
>
> It would be interesting to understand how these two contrasting 
> positions can be reconciled.
>
> Also I would be interested to know [from the librarians on this 
> list by direct email to me please] which publishers refuse to 
> sell individual subscriptions to libraries.
>
> Many Thanks
>
> Chris Beckett
> chris@cbathome.demon.co.uk