---------- Forwarded message ----------
Date: Tue, 26 Nov 1996 08:41:09 -0500
From: Steve Black <blacks@ROSNET.STROSE.EDU>
Subject: Re: ProQuest (Rosalee McReynolds)

> From: Rosalee McReynolds <reynolds@BETA.LOYNO.EDU>
> Several months ago, it was brought to my attention that UMI includes a
> statistics module in its ProQuest software that counts the number of
> times users have printed out citations from the ProQuest indexes. . .
> .  Has anyone else used these statistics in this way, or in any other
>way?

Rosalee,

We get similar statistics from our Academic Abstracts CD-ROM (an EBSCO
product).  We are also in the midst of a journal use study.  A few months
ago I correlated journal use for one semester with journal "hits" on
Academic Abstracts.  My statistical analysis of the results indicated that
43% of the use of journals indexed in AA was due to their being cited in
AA.

  I never did anything with these figures, because the whole point of
multiple regression analysis is to determine the impact of multiple
factors on a phenomenon.  Since I only had one factor (hits in Academic
Abstracts), I couldn't say anything worthwhile about why journals get
used.  But the 43% figure "feels" right to me, and reinforces my gut
feeling that CD-ROM indexing strongly influences journal use.  We all know
there's a strong influence, but I've never seen any published proof of it.

  Anyway, one possible use of the stats is to compare them with journal
usage.  Let me know if you want details of my analysis methods, I'll be
glad to share.
************************************************************************
Steve Black
Reference, Serials and Instruction Librarian
Neil Hellman Library
Albany, NY  12203                                  "Cogito eggo sum"
blacks@rosnet.strose.edu                     (I think, therefore I waffle)