College and Research Libraries


FLOYD CAMMACK and DONALD MANN 

Institutional Implications of 
an Automated Circulation Study 

In February 1965 Oakland University library implemented a circu-
lation control system utilizing IBM 357 data collection units with a 
1620 computer equipped toith disk storage. An initial test run of 
analytic by-products of the system shows implications for more effi-
cient library administration and further defines the library's function 
in the educational process . 

. THIS PAPER is a report, based on initial, 
limited data, of a few of the questions 
and answers made economically possible 
through the combination of machine-
readable circulation and student records 
at Oakland University. The specific an-
swers to the questions asked in this 
study are of secondary importance, ex-
cept as their relative status may suggest 
general trends. The questions themselves, 
together with similar, more refined types 
of inquiries, are ones which become 
readily and regularly answerable with 
any circulation system that combines call 
number information with borrower iden-
tification in an easily tabulatable form. 

In the case of the Oakland system, 
this transaction record is a punched card 
produced automatically from a machine-
readable book card, a machine-readable 
borrower's card, and a time clock. Its 
primary function is input to the main 
circulation records system. By accumu-
lating these cards after their information 
has been processed through the com-

Until July 1966 Dr. Cammack was Uni-
versity Librarian and Associate Professor of 
LinguWtics at Oakland University, Roches-
ter, Michigan. He is now on leave of ab-
sence as a Fulbright Lecturer in Japan. 
Donald Mann is Assistant Director of Oak-
land's Computing and Data Processing 
Center and the designer of the system 
described. 

puter, there develops a store of book-
and-borrower data which in turn can be 

· run against student records containing 
current information on grades, curricu-
lum, and class. 

The analyses performed thus far are 
basically counting mechanisms. The ap-
proach to each analysis was to deter-
mine, within the charge data, the 
possible span of the quality- or identifier-
field to be analyzed, such as the day of 
charge, class of book charged, etc. A 
counter for each unit of the span was 
set up in computer memory, the file of 
charge cards was passed through the 
computer, which incremented the appro-
priate counter for each charge, .and the 
contents of the counters were then 
printed. 

No forecasting or statistical analysis 
was done by computer at this time. The 
study merely lends insight into the rela-
tive volumes of the various characteris-
tics of the data. By establishing the 
counters in memory prior to giving the 
data to the computer on each run, the 
necessity of large-scale card sorting was 
all but eliminated. All programs were 
written in the FORTRAN programing 
language and were run on the university 
IBM 1620 computer which has card, 
disk, and printer capacity. 

The charge cards were first separated 
from the discharge cards. Three analyses 

I 129 



130 1 College & Research Libraries • March~ 1967 

were then made on the charges: Use by 
Day (run time: 50 min.), Use by LC 
Class ( 85 min. ) , and Use by Time of 
Day (55 min.). The charges for students 
were then split out and sorted into stu-
dent number sequence. These cards were 
then run against the student cumulative 
performance file which is maintained by 
the university, and student summary 
cards were punched. These cards contain 
student number, number of charges, 
cumulative Grade Point Average, curric-
ulum, and class. The remaining four 
analyses were then run: Use by CPA 
(20 min.), CPA by Use (20 min.)~ Use 
by Class ( 20 min. ) , and Use by Curric-
ulum ( 20 min.). Copies of detailed data 
not presented here can be made avail-
able on request. 

The initial tabulations, after four 
months of operation, provided answers 
to the following inquiries. 

Question: Could personnel scheduling 
at public service desks be more effi-
ciently or economically handled? 

Answer: Yes. A tabulation of dates of 
charges showed .a striking consistency by 
which charging activity tended to follow 
a cyclical pattern, consistently giving 
certain days of the week by far the 
heaviest workloads. Desk staffing pat-
terns and shelving labor could be 
brought into better relationship with an 
unexpectedly predictable activity cycle. 
In answer to the same question, the 
number of changes per hour showed 
four regular peaks with definite impli-
cations for desk staffing and improved 
closing procedures. Unexpectedly, and 
in contradiction to the staff's impressions, 
charging activity did not tend to cluster 
around class intervals, but spread quite 
evenly throughout each hour of the 
service day. 

Question: Which portions of the li-
brary's collections are used most heavily, 
and which are used least? 

Answer: A breakdown by LC classifi-
cation letters showed English literature, 

history, philosophy, education, and eco-
nomics topping the list in that order. 
The lowest significant use areas involved 
mathematics and the physical sciences, 
the totality of which did not equal use 
in English literature alone. 

Such an answer might well raise a 
number of subsidiary questions concern-
ing adequacy of the collection for stu-
dent use, library orientation for science 
students, comparative figures for similar 
institutions, the numerical relation of 
science majors to other majors, .and per-
haps the science faculty's awareness of 
the library's resources together with the 
whole question of departmental goals 
and book budget allocations. 

Question: Students majoring in which 
subjects tend to be the heaviest library 
use·rs, and does this correlate as expected 
with the answer to the immediately pre-
ceding question? 

Answer: Secondary education majors 
in modern foreign languages and English 
head the list for average number of 
charges per person, followed closely by 
liberal arts majors in the same fields. 
Similarly high averages support history 
and philosophy use patterns mentioned 
above. Majors in physical and biological 
sciences average below the mean, but 
again with their secondary education 
counterparts somewhat more active than 
students in the College of Arts and 
Sciences. 

The suggested implications of these 
tabulations would appear to be that stu-
dents tend to re.ad largely within their 
own fields of study (although a further 
breakdown by individual students' 
charge records would be required to 
confirm the conclusion) and that, at 
Oakland at any rate, students planning 
teaching careers use the library more 
than arts and sciences students with the 
s.ame subject specialization. 

Question: Do upperclassmen tend to 
use the library more than lowerclass-
men? 



Answer: Decidedly yes. Of those who 
used the library during the list period, 
juniors and seniors averaged almost 
twice the number of charges for fresh-
men and sophomores. Somewhat more 
enlightening are the implications of the 
class breakdown shown in Figure 1 
below. The freshman year would appear 
to be .a «non-library" year, in at least a 
portion of which only 10 per cent of the 
class found it necessary to check out a 
library book. By the sophomore year, 
the number has increased sharply and 
by the junior year, apparently almost all 
students are library borrowers. With the 
implementation in 1966 of a new lower 
division curriculum at Oakland, it would 
seem worthwhile to establish a detailed 
analysis of freshman library activity (or 
inactivity) in cooperation with the ad-
ministrative officers responsible for the 
curriculum. The inclusion of librarians 

F ig. 1 

1000 t------r---.--- --------

200 

F res hm a n Soph omores J uni ors Sen io rs 

D L ibrary Users D Non-Use rs 

Automated Circulation Study I 131 

within the freshman advising program, 
and the presentation of library orienta-
tion programs, apparently require exami-
nation and increased emphasis. 

With an over-all library non-use figure 
of 42 per cent during the test period, it 
appears likely that freshmen are the odd-
men-out. Unless their «in-house" use of 
library facilities is considerably higher 
than other indications would imply, 
major attention should be focused on the 
library's relations with lower classmen. 

Question: What is the relationship be-
tween library use and academic achieve-
ment? 

Answer: E xtremely close. As illustrat-
ed in Figure 2, even a limited body of 

10.0 

9.0 

8.0 

7.0 

6.0 

5. 0 

0 

0 

2.0 

0 

Fo g. 2 

/ r--.... 

L ~ v 
,.~ 

..,...... 

1.5-1.7 18-2.0 2 1-2 .3 2.4 -2.6 2.7-2 .9 3.0 3.2 3.3 3, 3.6 3.8 3.9 4.1 
Cumulative G r ade Poult Average G roups 

data shows a direct, positive correlation 
between the borrowing of library rna te-
rials and cumulative grade point average. 
Using the test period as a sample, the 
cumulative GP A of students who bor-
rowed at least one book during the time 
was 2.73, while that for students who 
borrowed no books during the period 
was 2.54; implying again that regular 



132 1 College & Research Libraries • March~ 1967 

library borrowers are likely to rriake bet-
ter grades than do nonborrowers. It may 
also be that entrance examination scores 
are higher for the former group. A se-
mester-by-semester tabulation of these 
relationships could begin to establish 
profiles of student achievement by class-
year and subject. Further sorting of data 
could provide individual «reading pro-
files" for individual students and groups 
of students, answering in part the old 
question «What does a 'good' student 
do that a 'bad' student doesn't?" -and 
vice versa. 

Sidestepping for the moment any dis-
cussion of causal relationships between 
academic achievement and library use, 
the library's position (if not function) 
within the undergraduate educational 
process would appear, simply on the 
basis of circulation figures, to be demon-
strably significant. When this correlation 
is viewed in connection with freshman 
and sophomore use patterns, perhaps 
academic librarians (at least at Oak-
land) should set out to «recruit" users 
in their first two years of college work. 
It might well be that college librarians 
should take another look at the public 
library's operation and staffing of young 
adult collections. 

As the store of transaction data con-
tinues to increase as a by-product of the 
circulation system, it will soon reach the 
point at which reliable answers can be 

found to a number of additional ques-
tions. The ability to identify nonbor-
rowers quickly and accurately should 
provide useful information for academic 
advisers. The borrowing records for com-
muters, dormitory students, students in 
specific courses, part-time students, and 
faculty members are within easy access, 
allowing runs of the total accumulation 
of data or of any desired sample. While 
most manual circulation systems gener-
ate the same type of data, machine-
readable data allow easy and rapid 
tabulation at a fraction (estimated at 
1/5) of the cost of the same information 
developed manually. Initial faculty and 
administrative reaction to the availability 
of this type of information has very 
quickly answered the question, «Who 
cares?" Requests for special runs have 
ranged from the professor who asked for 
periodic «traffic reports" on his assigned 
reading lists to .a department chairman 
who expressed a near-unethical degree 
of interest in the reading habits of his 
faculty members in connection with pro-
motion decisions. 

With the system designed to go online 
as soon as practicable, the provision of 
remote terminals could put the answers 
to these and many other questions 
literally at the fingertips of librarians, 
administrators, and faculty members, 
with data always up-to-date and in-
stantly . available. • •