College and Research Libraries


ELI M. OBOLER 

Academic Library Statistics Revisited 

Claims are made that the latest (Library Statistics of Colleges and 
Universities, 1965-66: Institutional Data) report on national academic 
library statistics contains "great" and "frequent" errors. Numerous ex-
amples are given. Excessive use of individual reports which are 
rendered almost useless by qualifying footnotes and disregard of what 
was asked for is noted. Since these national figures are used for many 
significant purposes (such as replies to Congressional inquiries), it is 
urged that ACRL once more take up the task of compiling them. 

IN NovEMBER 1964 CRL published an 
article1 concerning academic library sta-
tistics which was intended as a one-time 
study of "certain imperfections and mis-
leading inclusions and omissions that 
deserve some attention and analysis." 
The recent publication of Library Sta-
tistics of Colleges and Universities, 1965-
66: Institutional Data compels a return 
to what were, in 1964, believed to be 
strictures which would not need repeat-
ing as soon as three years later-par-
ticularly in light of the aegis under which 
this newest study was issued, as "Com-
piled by the Library Administration Di-
v_isio;; of the American Library Associa-
tion. 

It appears, however, that whether the 
federal government or the American Li-
brary Association prepares college and 
university statistics, the pitfalls of these 
publications are so great and likely 
errors so frequent that the profession 
might be better off without any so-called 
"national" cumulation of figures at all. 
No lengthy research is necessary to sup-
port this point. 

The earlier article particularly em-

1 Eli M. Oboler, "The Accuracy of Federal Academic 
Library Statistics," College and Research Libraries 
XXV (November 1964), 494-97. 

Mr. Oboler is University Librarian of 
Idaho State University, Pocatello. 

phasized the "rather odd figures" given 
for "volumes" and for "volumes added," 
especially in relation to expenditures in-
dicated for "books and other library ma-
terials." If 1962-63 was considered 
"rather odd" in this report, then 1965-66 
was, to say the least, weird. 

Item: a library with 59,000 volumes at 
the end of 1963-64 reports holding well 
over 100,000 volumes as of July 1, 1966, 
despite adding only a few more than 
2,500 volumes during 1965-66. This 
would, of course, mean that during 1964-
65 this same library had to have added 
about 40,000 volumes! 

Item: another academic library, with 
somewhat more than 90,000 volumes as 
of 1963-64, records nearly 184,000 as of 
1965-66. 

Item: a third, with slightly more than 
350,000 volumes in 1963-64 and "volumes 
added" that year of a little more than 
19,000 suddenly becomes "big-time," 
with nearly 550,000 volumes in 1965-66, 
despite the fact that they indicate fewer 
than 35,000 volumes added in the same 
year. One must assume that this insti-
tution had around 515,000 in 1964-65 
having added the prodigious figure of 
160,000 (almost 50 per cent!) during 
the year. · 

Or do these, and many similar compar-
ative figures, reflect not errors in re-
porting, but rather (what could be inter-

/ 407 



408 I College & Research Libraries • November 1967 

esting) the fact that since 1963-64 there 
have been substantial changes in the 
definition of a volume. An examination 
of definitions, however, indicate that 
there has been no such change. For 
1963-64 a volume is defined2 as "A 
physical unit of any printed, typewrit-
ten, handwritten, mimeographed or proc-
essed work contained in one binding or 
portfolio, hardbound or paperbound, 
which has been classified, cataloged, or 
otherwise prepared for use. The term 
includes bound periodical volumes and 
all non-periodical government docu-
ments. All forms of microtext are ex-
cluded." 

In the 1967 report is found3 the fol-
lowing definition: "Volume. A physical 
unit of any printed, typewritten, hand-
written, mimeographed or processed 
work contained in one binding or port-
folio, or otherwise prepared for use. The 
term includes · bound periodicals [sic] 
volumes and all non-periodical govern-
ment documents. All forms of microtext 
are excluded." This does not look like a 
very substantial change. Actually, the 
only difference between the two defi-
nitions is the omission of the words 
". . . hardbound or paperbound, which 
has been classified, cataloged .... "Thus, 
there seems to be no evidence of justi-
fication for «errors" by change in defi-
nition of the disputed term "volume." 

Can such variations in reporting be 
explained in another way? In a preface 
to the 1967 volume, Theodore Samore 
says: 4 "The substantial number of in-
completed forms necessitated the appli-
cation of valid editing procedures to re-
trieve data which would have otherwise 
been unavailable. Hence, the consider-
able increase in the number of foot-

2 U.S. Office of Education, Library Statistics of 
Colleges and Universities, 1963-64 Institutional Data. 
OE-15023-64. Circular No. 769. (Wash., D.C.: U.S. 
G.P.O., 1965), p. 4. 

8 ALA Library Administration Division, comp. Li-
brary Statistics of Colleges and Universities, 1965-66 
Institutional Data (Chicago: ALA, 1967), p. 3. 

' Ibid., p. 2. 

notes, especially those marked <esti-
mate.' " Since the matter of «estimated" 
figures is supposedly taken care of by 
Mr. Samore's explanation, let us look 
over the actual use of «estimated" figures. 

As was pointed out in the 1964 article, 
the use of rounded-off figures is a fairly 
clear indication of the use of estimation, 
rather than actual physical or biblio-
graphical count to determine results. 
The 1965-66 records show use of «esti-
mated" as a footnote (on the matter of 
volume count) twenty-three times; yet 
thirty-five other academic institutions 
show «000" at the end of their volume 
figures, without stating that their figures 
are estimated. 

The figures for «number of volumes 
added during year" present a similar pic-
ture. Of the 1,891 institutions which gave 
«volumes added" figures, only six foot-
noted their figures as «estimated," yet 
fifty-eight other college and university 
libraries used rounded-off figures. 

Despite the explicit instructions in the 
questionnaire which was the basis of this 
report, no fewer than nine did include 
some type of microtext as part of the 
figures reported under «number of vol-
umes at end of year," of whom three in-
cluded microtexts added under "number 
of volumes added during year." Although 
of all government document holdings 
only non-periodical government docu-
ments were to be counted as volumes in 
this survey, one university library "ex-
cludes unbound government documents," 
one college library «excludes US and 
UN documents," and one university li-
brary "excludes non-periodical govern-
ment documents." Surely these figures 
should have been omitted, rather than 
being counted with an explanatory foot-
note added, if the totals based on them 
are supposed to be reliable guides to 
current trends or situations in academic 
library resources. 

The data concerning microtext hold-
ings in this report are also interesting. 



Academic Library Statistics Revisted I 409 

The questionnaire gave this definition of 
microform: 5 "Microform. This includes 
the form of any library material which 
has been photographically reduced in 
size for storage and reproduction pur-
poses, and which must be read with the 
help of enlarging instruments. The term 
is synonymous with microtext and in-
cludes microfilm, microcard, and micro-
fiche." Column 10 of Table 1 of the 
1965-66 report calls for "number of 
physical units of microform at end of 
year." Of the respondents, forty-one 
gave "estimated" figures; three gave 
"bibliographical count only" ( although 
physical units were specifically asked 
for); one gave "volume count only"; one 
gave "microfilm only" (which probably 
means that their microform holdings of 
documents and manuscripts were not 
reported, but one cannot be sure). And, 
finally, on this matter, one indicated that 
it had so many physical units of micro-
form, and so many volumes, but foot-
noted "data for microform included ... 
volumes." It might be asked why this 
library did not simply subtract one fig-
ure from the other and report that total 
as was asked for, or why the editor did 
not do it for them? 

One must admire the candor of one 
small church-supported college which re-
ported, "All figures are estimates." In-
cluded were such data as the total num-
ber of students, the size of the library, 
the total budget, the number of staff, 
and even the beginning salary of a library 
school graduate (fifth year degree with-
out experience). Somehow, however, the 
mind boggles at any academic librarian 
"estimating" how many students are 
registered in his institution or how many 
individuals are on the library staff. 

The earlier stady of statistics referred 
to the "wonderland of academic statis-
tics," and said6 that "the further one 
goes ... , the 'curioser and curiouser' 

5 ALA, op. cit., p. 3. 
e Oboler, op. cit., p. 494. 

they get." What has been printed in the 
1967 report leads one to repeat the com-
ment. When one correlates replies printed 
under "number of volumes added during 
year" with the amounts reported as ex-
penditures for ''books and other library 
materials," one runs upon an additional 
number of unusual figures. 

Taking up Table 2, "Operating Ex-
penditures, Personnel, and Beginning Sal-
ary . . ." one finds some intriguing foot-
notes. Under "Wages," for example, 
twenty-eight give "estimated" figures, 
four state that their reports exclude 
"funds from Federal Work-Study Pro-
gram," six say "student assistants only," 
and one indicates his figures "include all 
employees paid on an hourly basis." How 
this latter differs from other reports on 
''Wages" (since the questionnaire said, 
"Amounts (including monetary estimates 
for contributed services) paid to stu-
dents and to others paid on an hourly 
basis are listed under 'wages,' " is diffi-
cult to understand. 

One of the more unusual presentations 
in Table 2 comes up where total operat-
ing expenditures in dollars are asked for. 
One institution lists a particular specific 
amount under "total" and then footnotes 
it as "estimated." Yet none of the other 
figures for this institution-salaries, cost 
of library material, binding, or "other" 
-are marked "estimated." The total of 
the four figures given is about 3 per cent 
above the total given as "estimated!" 

Surely the above examples of irrespon-
sible and confused reporting bear out 
the charges at the beginning of this 
article. If further evidence is required, 
one need only look at Table B7 in the 
1965-66 report. The figures for the total 
number of volumes at the end of the 
year for the academic library reporting 
say, most ingenuously " (includes micro-
text)." Yet the questionnaires on which 
this table is based exclude all forms of 

7 "Summary of College and University Library 
Statistics for Academic Years 1959-65," op. cit., p. 6-9. 



410 I College & Research Libraries • November 1967 

microtext in asking for "number of vol-
umes at end of year," "number of vol-
umes added during year," and "number 
of volumes withdrawn during year." 

The same Table B is footnoted to indi-
cate that all figures for 1966 (that is 
to say, 1965-66) are "estimated." The 
previous comments in this article on the 
values of estimation should be noted 
again, especially when one realizes that 
official statements on the state of Ameri-
can academic libraries now are usually 
based on this olla podrida of invalid and 
unreliable statistics. The figures given 
for "number of periodicals received" are 
footnoted. "For 1965-66, the figures are 
for Serials which includes periodicals, 
annuals, proceedings, transactions, etc." 

Then why even give 1965-66 figures, 
since they are based on such hollow 
shells of fact as are indicated throughout 
this article, obviously incapable of any 
meaningful comparison with past data? 

In sum, the statistics picture for aca-
demic libraries in the United States is at 
least cloudy, if not psychedelic. Perhaps 
the task should be returned to the As-
sociation of College and Research Li-
braries, which seemed to do a pretty 
fair job for a great many years, before 
the computers, the federal government, 
and the Library Administration Division 
took over. 

In at least one man· s judgment, bad 
-even misleading -statistics are worse 
than no statistics at all. • •