RE: EBSCO and ProQuest database content

["LENARES,DEBORAH A." <LENAREDA@apci.com>]

Donnelyn,
Thank you for all this valuable information.  
One question:  do you know if each of the databases defines "peer reviewed"
in the same way?  The number of peer reviewed titles in EBSCO's Academic
Search Premier is suspiciously high.  Another valuable assessment tool is
performing a number of typical searches and evaluating the results both
quantitatively and qualitatively.

The members of the database selection committee of the Louisiana Academic
Library Consortium have spent much time over the past year comparing full
text databases.  Part of the results of the comparison are available in an
article by J.B. Hill of Southeastern Louisiana University in the latest
issue of Issues in Science and Technology Librarianship available at
http://www.library.ucsb.edu/istl/01-spring/article4.html The article focuses
on science titles in Bell & Howell's Research Library (Core and Sciences
modules), Ebsco's Academic Search Elite, and Gale's Expanded Academic ASAP.

Deborah Lenares



-----Original Message-----
From: Donnelyn Curtis [mailto:dcurtis@admin.unr.edu]
Sent: Thursday, May 24, 2001 10:40 AM
To: liblicense-l@lists.yale.edu
Subject: EBSCO and ProQuest database content


Because I started a discussion based on superficial observations, and in
response to onlist and offlist follow-up messages, I felt obligated to do
some more in-depth analysis of the currently posted source lists for EBSCO
and ProQuest databases. It is definitely true that we need to understand
what we are buying, and how those products are changing. Of course there
are several other full text databases on the market, and other companies
offering access to aggregated electronic journals. But I am just looking
at the three databases that some of us have already been comparing in
recent postings.

Both vendors' lists are showing the effects of journal publisher
retrenchment. Database content is shifting. I hope that we will see some
other response models emerge, such as more customization of databases,
more pay-per-viewing, pricing models where we pay for electronic content
only one time and access fees that are tailored to each situation, more
incentives from publishers to help us shift from print to electronic
format so they can move more quickly into an electronic-only publishing
environment. Publishers and libraries both have high costs in maintaining
both formats.

Anyway, below are the results of a very close look at the current source
lists for the ProQuest Research Library (core and all modules), Academic
Search Elite and Academic Search Premier. All of them have an impressive
amount of content. I agree that for what we pay for these databases, we
can't expect them to be substitutes for subscribing to journals.

My main concern with embargoes and ceased full text coverage is about
sending students to databases that will not support their research papers
with current articles. EBSCO, which has thus far shown the most dramatic
shift, is trying to address that concern by enabling linking from the
databases to a library's other full text subscriptions, through EBSCO
Online or the library's catalog. SFX and other third party products make
it possible for libraries to enable their own linking and gateway
searching across packages of electronic journals. Things are getting more
complicated, and we can't expect any one product to satisfy all our needs.
It's going to take more effort on the part of libraries to coordinate
access to our resources. That's another topic, though. Back to my study.

First, I need to say that I am taking the currently posted source lists at
face value. I am trusting that they are accurate and up-to-date. I did
some but not much spot checking to verify current full text availability
and did not find any discrepancies. I had to do quite a bit of work to
standardize the lists so I could compare certain elements, and I'll be
glad to send the resulting Excel files as attachments. Contact me at
dcurtis@unr.edu. Note: only full-text titles were considered in this
study.

Full text titles on vendor's source lists:

  ProQuest Research Library (PQRL) -- 1676
  EBSCO's Academic Search Elite (ASE) -- 1459
  EBSCO's Academic Search Premier (ASP) -- 2806

Full text ends for natural reasons (title changes, ceased publication):

  PQRL -- 160
  ASE -- 115
  ASP -- 174

Full text ends, but the journal continues, and some full text is still
included in the database:

  PQRL -- 101 (6%)
  ASE -- 7 (<.5%)
  ASP -- 4 (<.5%)

Full text titles embargoed for more than 30 days:

  PQRL -- 11 (<1%)
  ASE -- 339 (23.2%)
  ASP -- 1257 (44.8%)

Full text ceases, coverage is for less than a year:

  PQRL -- 21 (9 for one month only)
  ASE -- 10
  ASP -- 17 (2 for 1 month only)

Peer-reviewed full text titles:

  PQRL -- 664
  ASE -- 961
  ASP -- 2033

Peer-reviewed titles with current full text, no embargo:

  PQRL -- 548
  ASE -- 587
  ASP -- 731

Unique full text titles [I can also send lists of these]:
Note: All ASE titles are included in ASP

  In PQRL but not in ASE or ASP -- 842
    Peer-reviewed in PQRL but not ASE or ASP -- 265

  In ASP but not PQRL -- 1974
    Peer-reviewed in ASP not in PQRL -- 1556

  In ASP but not in PQRL or ASE -- 1202
    Peer-reviewed in ASP but not in PQRL or ASE -- 1013

  In ASE but not in PQRL -- 737
    Peer-reviewed in ASE but not in PQRL -- 533

Format of full text:
Page images for articles in journals:

  PQRL -- 1587 (94.7%)*
     Current titles -- 1308 (78.0%)*
     Peer-reviewed -- 649 (97.7%)**
     Current peer-rev. -- 533 (80.0%)**

  ASE -- 1093 (75.0%)*
     Current titles -- 1093 (75.0%)*
     Peer-reviewed -- 800 (83.2%)**
     Current peer-rev. -- 800 (83.2%)**

  ASP -- 2201 (78.4%)*
     Current titles -- 2201 (78.4%)*
     Peer-reviewed -- 1952 (96.0%)**
     Current peer-reviewed -- 1952 (96.0%)**

* percentage of all FT titles in database
** percentage of all FT peer-reviewed titles in database

Depth of full text coverage:
[Note: I include more detail here because this is hard to pull from title
lists, and it can be sliced and diced various ways, so I would rather not
slice it. I hope the formatting stays intact. If not, I can send it in
Excel].

Years of coverage for all full text journals in databases:
[note: for this part of my study, embargoed titles were considered to have
an ending date, depending on the length of the embargo. Percentages
relate to all full-text titles in each database.]

               PQRL     	ASE     	ASP
14 years	 91	 5.4%	0	 0.0%	0	 0.0%
13 or more	237	14.1%	0	 0.0%	0	 0.0%
12 or more	251	15.0%	0	 0.0%	0	 0.0%
11 or more	281	16.8%	379	26.0%	399	14.2%
10 or more	324	19.3%	409	28.0%	429	15.3%
 9 or more	425	25.4%	457	31.3%	485	17.3%
 8 or more	448	26.7%	525	36.0%	560	20.0%
 7 or more	609	36.3%	662	45.4%	721	25.7%
 6 or more	738	44.0%	751	51.5%	827	29.5%
 5 or more	893	53.3%	979	67.1%	1108	39.5%
 4 or more	1152	68.7%	1071	73.4%	1270	45.3%
 3 or more	1525	91.0%	1243	85.2%	1659	59.1%
 2 or more	1615	96.4%	1394	95.5%	2372	84.5%
 1 or more	1655	98.7%	1449	99.3%	2725	97.1%
  < 1 year	  21	 1.3%	  10	 0.7%	  81	 2.9%

Years of full text coverage, peer-reviewed journals in databases:

               PQRL	       ASE	       ASP
14 years	31	4.7%	0	0.0%	0	0.0%
13 or more	77	11.6%	0	0.0%	0	0.0%
12 or more	83	12.5%	0	0.0%	0	0.0%
11 or more	94	14.2%	260	27.1%	266	13.1%
10 or more	104	15.7%	283	29.4%	290	14.3%
 9 or more	143	21.5%	305	31.7%	314	15.4%
 8 or more	148	22.3%	342	35.6%	352	17.3%
 7 or more	234	35.2%	425	44.2%	441	21.7%
 6 or more	297	44.7%	464	48.3%	487	24.0%
 5 or more	375	56.5%	620	64.5%	673	33.1%
 4 or more	492	74.1%	679	70.7%	784	38.6%
 3 or more	636	95.8%	811	84.4%	1118	55.0%
 2 or more	661	99.5%	926	96.4%	1738	85.5%
 1 or more	663	99.8%	955	99.4%	1991	97.9%
  < 1 year	  1	 0.2%	  6	 0.6%	  42	 2.1%

I will let readers draw their own conclusions from this data. I hope some
of it will be useful in comparing content of competitive databases. Of
course, content is only one of the aspects of these databases. There are
things we like and dislike about the search interfaces and the browsing
capabilities of full text databases. Both companies are working on
enhancements. But what is reflected here is the content we get if we
subscribe or renew a subscription today.

After working with these lists and those from other vendors at other
times, I would recommend the EBSCO source list format as a model for
describing database content. I appreciate the inclusion of publisher
information and peer review status, and to have one starting and ending
date for full text, regardless of format (text or PDF). I appreciate that
the ProQuest lists have separate columns for ending dates and that there
are notes describing recent changes.

Donnie Curtis
University of Nevada, Reno Libraries


+


+