---------- Forwarded message ----------
Date: Fri, 13 Apr 2001 16:23:04 -0400
From: David Goodman <dgoodman@PHOENIX.PRINCETON.EDU>
Subject: Re: E-J statistics (Minh-Thu Nguyen)
I think your method, besides not getting journals that your users have
bookmarked, will also not get journals that your users access directly
from the links in other journals or from indexes. It's my impression that
these may represent from 50 to 90% of the use, at least in the sciences.
A number of other people have also suggested answers based on this method,
all of which have the same objection.
In any case, we are specifically interested in every individual journal.
A number of other people have also suggested ways of counting those
journals whose counts are easy to get to.
So far no one has told me that they are trying to measure on a regular
basis, every access from their users to each of the titles.
That does not mean that partial measurments are without value, but we are
trying to get a complete detailed measurment.
David Goodman, Princeton University Biology Library
dgoodman@princeton.edu 609-258-3235
On Fri, 13 Apr 2001, Marcia Tuttle wrote:
> ---------- Forwarded message ----------
> Date: Fri, 13 Apr 2001 14:52:35 -0400
> From: Minh-Thu NGUYEN <mtnguyen@COURRIEL.POLYMTL.CA>
> Subject: Re: E-J statistics
>
> At Ecole Polytechnique, Montreal, we have been taking "usage" statistics of
> electronic resources for about a year.
>
> In order to do this, our URLs have 2 parts:
> - part 1 is the address of a local perl script which increments the count,
> then redirects to the address in the second part.
> - part 2 is the address of the resource.
>
> Example:
> http://www.biblio.polymtl.ca/cgi-bin/go/ojps.aip.org:80/ASMEJournals/Dynamic
> Sys/
> <-------------- part 1 --------><--------------- part 2 --------------->
>
> Every resource's address has this structure, however, we keep statistics by
> group. This means that if someone clicks anyone of the thirty or so journals
> of the ASME, the group ASME will be incremented, but we won't know which of
> the journals has been accessed. Which is kind of OK since most electronic
> resources are licensed by group, and since we cannot afford to consider
> doing this on an individual basis anyway .
>
> Advantages :
> - have SOME statistics, PLUS a uniform way of counting and presenting
> statistics for all resources.
> - have complete stats of computers in the home domain
>
> Disadvantages :
> - need to insert the script's address in each resources to count.
> - link checking programs do not check redirected addresses (only report the
> redirection), so we have to produce a list of such addresses, take out the
> srcipt's address then run the checker program.
> - users probably bookmark their favorite journals which may or may not
> contain our script.
> - every click is counted, not the sessions (the sessions are on the
> publishers' servers)
> - need to manipulate the stats to address the following problems:
> . robots and people from outside (they might or might not be able to
> access after all but the clicks are counted)
> . people using Internet services and our proxy server.
>
> I would have to say that all this is quite time consuming ! We try to limit
> this control to paid resources, and as mentioned, as groups.
>
> Hope this helps.
> Minh-Thu Nguyen
> Ecole Polytechnique, Montreal
> mtnguyen@mail.polymtl.ca
>
> ---------- Forwarded message ----------
> Date: Wed, 11 Apr 2001 13:28:13 -0400
> From: David Goodman <dgoodman@Princeton.EDU>
> Subject: E-J statistics
>
> Are any of the readers of this list maintaining a month-by-month count
> of the total use by their library of _every_ ejournal they receive? (Not
> just the journals from a particular publisher or distributor.) I have
> been asked to provide an estimate of the work involved.
> Please contact me off list.
>
> --
> David Goodman
> Biology Librarian
> and Co-chair, Electronic Journals Task force
> Princeton University Library
> Princeton, NJ 08544-0001
> phone: 609-258-3235
> fax: 609-258-2627
> e-mail: dgoodman@princeton.edu
>
