ACRL News Issue (B) of College & Research Libraries


User statistics for CD-ROM 
selection and management
By Karyle S. Butcher and Shirley R. Scott

Good data aids collection 
decisions

A n academic library entering the CD-ROM (CD) arena may anticipate the popular­
ity of CD access for its users and the incr

visibility it may bring to the library on cam­
pus.1 Such library services as online searching, 
bibliographic instruction, and interlibrary loan 
will be affected by the addition of this technol­
ogy.2 Moreover, several important management 
issues must also be considered. A statistical pro­
gram can be useful in resolving these issues; 
this is especially true for those decisions which 
revolve around staffing and collection manage­
ment. This paper addresses the usefulness of 
CD-ROM user statistics in making selection/ 
deselection decisions and in managing a CD- 
ROM installation. Included is reference to the 
authors’ experience.

Statistics m anagem ent
In 1987 the O regon State University (OSU) Li­
braries established a 12-station CD reference 
center. O ur belief w hen w e established the 
center w as that it w ould be used widely by 
undergraduates and graduates and that as its 
use increased, other library services and col­
lections w ould be affected. To see if this was 
true, w e developed a simple system w hich al­
lowed us to track usage. The system required 
users to sign in and indicate their academic 
status and which databases they intended to 
search. Statistics w ere totaled monthly or as 
time allowed. As usage increased, w e moved 
to a more sophisticated statistics-keeping pro­
gram. In 1990 the library purchased menuing 
software from Saber Software Corporation to

e

manage record-keeping and to customize the 
menu format. Data is now  recorded automati­
cally as the user logs onto the CD workstation 
or network. The program allows for the creation 
of a dBase data file which can then be analyzed 
by a program written by our CD librarian.

The software captures all menu choices into 
astehde  dBase file as well as time logged in and 

out, and includes an automatic logoff function. 
Although w e do not presently restrict off-cam- 
pus users w ho access our CD network, the Sa­
ber software provides options for passwords 
and limitations on the num ber of users of a 
particular application. It also has the potential for 
providing help screens for users at remote sites.

The setup is configured so that the initial 
m enu screen is the login screen. Here, the user 
chooses either the local drives or netw orked 
databases. Because the system requires a dif­
ferent login prom pt for each location, the soft­
ware allows us to know  w hether a user is in 
the library or outside the library. Local and 
netw orked databases are listed on separate 
menus, thus informing offsite users that the li­
brary has databases w hich are not available on 
the network. After selecting a database, users 
are required to indicate their group status (fresh­
man, sophom ore, junior, senior, graduate, fac­
ulty, staff, other) before going to the next 
screen, w hich requests departm ent or other 
affiliation. Next is a seven-subject menu. The 
final menu is a database list appropriate to the 
subject area chosen.

The information provided by this program 
is useful for many management and research 
purposes. We have analyzed use statistics to 
determine w hen the center is most heavily used 
and have reduced staff hours during slow peri­
ods and increased staffing at busy times. The 
program tells us w ho our users are and which

Karyle S. B utcher is assistant university librarian f o r  research a n d  reference services a n d  Shirley R. Scott is 
h ea d  o f  Research Services a t Oregon State University

September 1 9 9 2 /5 0 5



databases they are using. Analysis of this infor­
m ation helps us to determine w hether users 
are selecting the appropriate databases for their 
indicated subject interests. If, for example, w e 
determine that graduate students in agriculture 
are using the General Science In d ex instead of 
Agricola, w e will emphasize in our library in­
struction courses the difference betw een the 
two. In addition, it provides information on the 
average time a user searches a database, al­
lowing us to know  how  often and how  long 
each database is used p e r hour, per day, per 
w eek, and per month. We are able to deter­
mine the cost per search of each database, and 
have used this information in budget presenta­
tions to the university administration.

Collection m anagem ent
User statistics are also helpful in making col­
lection management decisions. The increase of 
information on CDs and declining budgets are 
forcing libraries to look at a variety of methods 
for making collection m anagem ent decisions. 
Libraries must balance the decision to purchase 
CDs with that of purchasing materials in tradi­
tional formats. A statistical package can be use­
ful in making these decisions.

At OSU, the initial selection of databases 
w as accom plished by librarians w ho recom ­
m ended purchases in their subject areas. Since 
that time, the CD librarian has established a 
dem onstration workstation in the reference of­
fice o n  w hich databases are preview ed and 
evaluated by librarians and other campus fac­
ulty. The librarians use a revised version of the 
selection criteria outlined by Nancy Herther and 
Irene Hoffman3 and make recommendations for 
purchase. Because a stagnant materials budget 
has determ ined that w e can add no new  serials 
subscriptions w ithout canceling old ones, pur­
chase decisions cause heated debate among the 
librarians responsible for collection develop­
ment.

We developed a series of steps to follow to 
aid in making selection and deselection deci­
sions for paper and CD formats. Included in 
this procedure is a comparison o f the cost and 
usage statistics for those databases on CD with 
the paper copies. Such analysis covers the cost 
of the paper index versus the CD cost, the scope 
of coverage, and information on w ho uses the 
databases, i.e, undergraduates, graduates, or 
faculty.

To determine how  w e could improve our 
current CD collection, the CD librarian analyzed

the overlap in titles indexed for a num ber of 
our subject-related databases and com pared 
costs of retaining indexes in both print and CD 
formats. The cost of the print version for some 
indexes, especially the Wilson indexes, is of­
ten comparable to or more than the CD ver­
sion. Using this information as well as usage 
statistics, the subject librarians make recommen­
dations for canceling paper or CD copies of 
the indexes. With the savings from cancellations 
of one copy of indexes in duplicate formats, we 
have purchased additional CD databases.

Conclusion
Purchasing information in electronic formats 
such as CD-ROM can be difficult due to de­
creasing budgets and the need to maintain tra­
ditional formats. Yet, increasingly sophisticated 
library users request the ease of using elec­
tronic formats. How to evaluate the use and 
cost of databases on CD w ith those in print 
format? If paper sources are d ropped are some 
users being disenfranchised? How are these 
decisions to be made?

Using available software programs, collec­
tion and analysis of user statistics are helpful in 
making these decisions. Such programs allow 
for the collection of data on use, level of user, 
choice of database, and am ount of time spent 
searching. This information can be applied to 
determine use and staffing patterns for CD in­
stallations and allow the library to make ad­
justments accordingly. Selection and cancella­
tion decisions are also m ade easier using data 
gathered on use, cost, and scope of coverage. 
Knowing comparative costs, w ho the users are, 
and w hat they are using allows the opportu­
nity to make enlightened decisions.

Notes
1Carol Tenopir, “The Impact of CD-ROM on 

Online,” Library Journal 116 (Feb. 1991): 61-62.
2Karyle Butcher and Shirley Scott, “Effects 

of CD-ROM in a University Library,” Journal o f 
E d u ca tio n a l M edia & Library Sciences 27:3 
(Spring 1990): 257-69.

3N. K. Herther, “How to Evaluate Reference 
Materials on CD-ROM,” Online 12 (March 1988): 
106-08; Irene M. Hoffman, “CD-ROM Planning 
Checklist; Collection Development Resource/ 
O rganization/M anagement,” address given at 
an OCLC Preconference, “CD-ROMBO: Com­
bat and Compromise.” Annual Conference of 
the O regon Library Association, April 5, 1989, 
Sunriver, Oregon. ■

5 0 6  /  C&RL N e w s



The small customer. The small customer 
at Ballen.

With some booksellers, a  small answ er all your questions. And an ongoing evaluation of library 
customer doesn’t rate quite the your very own shelf reserved needs so rigorous, it’s given us 
consideration th at a  large customer in our w arehouse to guard the industry’s lowest overall 
might. But a t Ballen, the single against shipment and inventory return rate -  less than 1%.
book customer receives the same mix-ups. All of which we provide with 
service and services th at a It also means access to our the understanding th at w hat are 
thousand book customer does. online, interactive BallenNet small fish today, might one day 

Which m eans no m atter w hat system for order entry and be big ones.
size your library, you can expect current status. The ability to For information or o u r new 
to have your very own customer reach a  Ballen company officer brochure, call (800)645-5237.
service representative to help ju st by picking up the phone. And