College and Research Libraries


By D O R O T H Y E. C O L E 

Some Characteristics of 
Reference Work' 

THERE ARE several generalizations which emerge from an examination of the 
literature on the measurement of reference 
work. Evaluation per se does not enter into 
the earlier literature, which is concerned in 
the main with a description of the tech-
niques, tools, and intellectual rewards of 
reference librarianship. T h e earliest men-
tion of reference records was made by 
Marilla W . Freeman in 1913.2 Since that 
date a number of librarians have attempted 
to cope with the problem of reference meas-
urement and have approached the problem 
in several different ways. A number of 
plans for classifying reference questions 
have been developed both in theory and in 
actual library practice. These classifica-
tions have been based o n : ( 1 ) the subject 
matter of the questions, ( 2 ) the amount of 
skill or knowledge necessary to answer 
them, or a combination of the two. T h e 
element of time has frequently been used 
to distinguish between simple and complex 
questions; frequently the latter are classified 
as complex because the process of answering 
them consumes more time. Examples of 
these classifications may be found in the 
articles by Swift,3 Alexander,4 Hazeltine,5 

and Barlow.6 

1 Condensed from an M . A . thesis, " A n A n a l y s i s of 
A d u l t R e f e r e n c e W o r k in L i b r a r i e s , " submitted to the 
Graduate L i b r a r y School, U n i v e r s i t y of Chicago, in 
I943-

2 Freeman, M a r i l l a W a i t e . " S c i e n t i f i c Management, 
and the R e f e r e n c e Department as a B u r e a u of In-
f o r m a t i o n . " A.L.A. Bulletin 7: 331-36, July 1913. 

8 S w i f t , I v a I. " C l a s s i f y i n g Readers' Questions." 
Wilson Bulletin 8 : 274-75, January 1934. 

4 A l e x a n d e r , Carter. " T h e Technique of L i b r a r y 
S e a r c h i n g . " Special Libraries 27: 230-38, September 
1936. 

6 Hazeltine, M a r y Emogene. " F u n d a m e n t a l s of Ref-
erence S e r v i c e . " Wisconsin Library Bulletin 15: 
85-90, A p r i l 1919. 

M a n y of the analyses of actual reference 
questions have been made in the public li-
brary field. Attention should be directed to 
the studies of Kingman,7 Conner,8 Guer-
rier,9'10 and Darsie.11 In some studies the 
questions were classified by subject, while 
in others an attempt was made to classify 
them according to their content or com-
plexity. W h e n questions are classified by 
subject, the field of the social sciences is 
usually found to be the subject group in 
which most questions fall, while studies 
based on the nature of the questions show 
that the informational or fact-finding group 
is the largest. T h e time spent in answering 
reference questions seems to suggest a basis 
for estimating the cost of Reference work 
and the skill of the reference worker. 

The Nature and Scope of Investigation 

N o precise and exact formula for evalu-
ating reference work has yet been devised; 
furthermore, the nature of the reference 
process seems to preclude the possibility of 
such an accomplishment. Certainly it was 
not the intention of the writer to aim at 
such a goal. On' the other hand, previous 
studies of reference work based on certain 

8 Barlow, S. H. " S u g g e s t i o n for Estimating U s e of « 
the R e f e r e n c e L i b r a r y . " Library World 4 1 : 29-30, 
August-September 1938. 

T K i n g m a n , Marion C. " R e f e r e n c e W o r k in a.^ 
Branch L i b r a r y . " A.L.A. Bulletin 32: 834-35, O c t . 
IS, 1938. 

8 Conner, Martha. " W h a t a R e f e r e n c e Librarian 
Should K n o w . " Library Journal 52: 415-18, A p r . 15, 
1927. 

9 Guerrier, Edith. " M e a s u r e m e n t of Reference Serv-
ice in a Branch L i b r a r y . " A.L.A. Bulletin 29: 632-37, 
September 1935. 

1 0 " T h e Measurement of R e f e r e n c e S e r v i c e . " 
Library Journal 6 1 : 529-31. J u l y 1936. 

1 1 Darsie, Helen H . " M e a s u r i n g the Results of 
R e f e r e n c e S e r v i c e . " A.L.A. Bulletin 29: 604-05, Sep-
tember 1935. 

JANUARY, 1946 45 



quantitative or measurable aspects have in-
creased our understanding of the nature of 
reference practice in libraries. T h e aim of 
the present investigation, then, is to aug-
ment this fund of knowledge. 

Previous investigations of reference work 
have given scant attention to certain seg-
ments of the process. Analyses of reference 
questions, as heretofore mentioned, have 
been made largely in the public library field, 
yet a great deal of important reference 
service is being rendered in academic and 
special libraries. W h a t are the elements of 
similiarity and difference in these three 
types of reference service? 

At the present time we have but slight 
knowledge of the relation between the 
reader's occupation and his reference re-
quests. W h i l e in the academic and special 
libraries the occupational distribution of the 
readers may be determined by the nature of 
the institution, the public library, in theory 
at least, should be meeting the needs of 
readers of all occupations. T h e present 
study includes an analysis of the occupa-
tional background of users of the reference 
department. 

It has been observed that many library 
readers have difficulty in expressing their 
reference needs. Another area of investiga-
tion, then, is the difference between what 
reference patrons want and what they ask 
for. One of the tasks required of reference 

f librarians is that of translating vague re-
quests into meaningful terms. Generaliza-
tions regarding typical behavior patterns of 
patrons should be helpful for actual refer-
ence practice. A n analysis of unanswered 
reference questions will also be undertaken. 

T h e final aspect of reference work 
covered in this paper is that of the sources 
of reference information. W h a t books and 
other materials were found useful in an-
swering reference questions ? Into what pat-
terns do they fall? T h e findings should be 
of interest to those responsible for book 

selection in libraries and to those who teach 
basic courses in reference work. 

Gathering the Data 

T h e questions on which this study is 
based were recorded in thirteen libraries 
during the month of April 1941. Not all 
of the questions received by these libraries 
during the period were reported, since li-
brarians found it impossible to record them 
because of pressure of business. Eleven of 
the libraries are in the Chicago area, one in 
central Illinois, and one in the western 
section of the United States. In addition a 
number of questions received at the St. 
Louis Public Library on one day in January 
1941 were used.12 T h e following types 
of libraries were represented: large, 
medium, and small public libraries, junior 
and liberal arts colleges, university depart-
mental libraries, and special libraries in the 
fields of social science, business and tech-
nology, and general reference. T h e form 
for recording the questions provided space 
for the following items: 

1. What patron asked for 
2. Exact. statement of the question 
3. Occupation of the patron 
4. Specific books and other tools used in 

finding the answer 
5. Other functions performed 
6. Was question answered satisfactorily? 

If not, indicate reason. 

All of the libraries were furnished copies of 
instructions for filling out the forms, in 
order to insure uniform and comparable re-
sults. 

Limitations of the Study 

First, the study does not attempt to 
establish a technique for measuring the 
reference work of an individual library, 
although a library might gain a clearer 
understanding of its reference work by 

12 Information Please: A Typical Day in the Refer-
ence Departments of the St. Louis Public Library . . . . 
S t . L o u i s , P u b l i c L i b r a r y , 1941. 

46 COLLEGE AND RESEARCH LIBRARIES 



keeping similar records. Second, the study 
does not consider the element of time in 
classifying reference questions. Finally, it 
does not include questions from large re-
search libraries which are used primarily 
by scholars. 

T h e 1026 reference questions upon which 
this study is based were classified by subject, 
since this method offered a convenient 
means of comparison with earlier studies. 
It is obvious from the subject distribution 
that no field of knowledge, at least at the 
level represented by the second summary of 
the Dewey Decimal Classification, is beyond 
the requirements of library reference serv-
ice. T h r e e areas stand out with particular 
emphasis: the social sciences, useful arts, 
and history. W i t h i n these groups were con-
centrated approximately 72 per cent of the 
questions. T h e following fields were well 
represented: political science, economics, 
law, administration, associations, medicine, 
engineering, biography, and modern history. 

Four other studies, all concentrated in 
the public library field, show a very close 
similarity, in the distribution of questions, 
to the findings of the present investiga-
tion.13"16 

At this point it is interesting to note that 
the specific character of the various libraries 
receiving the questions was apparently not 
a factor in the distribution. W h e n the 
classification . is restricted to libraries 
roughly homogeneous in function the same 
emphases appear. Thus, of the questions 
recorded by the public libraries, 69 per cent 
fell in the social sciences, useful arts, and 
history; the same proportion of questions in 
the college and university libraries fell in 

13 Charters. W . W . " C o l l e g e Preparation f o r R e f e r -
ence W o r k . " School and Society 27:150-52', 1928. 

14 Rozendal, Hendrine. " S t u d y of the R e f e r e n c e 
Service of the Davenport, Iowa, Public L i b r a r y f o r 
the P e r i o d f r o m 1925-1935." Unpublished master's 
thesis. L i b r a r y School, U n i v e r s i t y of Illinois, 1936. 

15 Conner, op. cit. 
16 Chait, William. " B o o k s and Other Sources U s e d 

in A d u l t Reference S e r v i c e in Branches of the Brook-
lyn Public L i b r a r y . " _ Unpublished master's thesis, 
School of L i b r a r y Service, Columbia U n i v e r s i t y , 1938. 

these fields; and of the questions asked 
in special libraries, 80 per cent were so 
classified. 

T h e occupational classification of refer-
ence patrons used in this study was one 
employed by Haygood in his analysis of the 
use of the N e w York Public Library.17 

Since some questions came from representa-
tives of organizations, a category was added 
to provide for them. O u t of 1026 ques-
tions asked, it was found that the concen-
tration appeared among students (356) and 
professionals ( 2 1 0 ) . T h e student group 
asked the largest numbers of questions in 
political science, economics, law, medicine, 
agriculture, literature, biography, and 
modern history. T h e fields in which pro-
fessional workers asked the most questions 
are political science, administration, lan-
guage, and engineering. Although the 
small number of questions asked by other 
occupational groups is scattered widely over 
the subject fields, we find that the questions 
of skilled laborers fell largely within the 
useful arts group, that most of the ques-
tions from shopkeepers and salesmen were 
about topics classified as social sciences and 
the useful arts, and that questions from 
clerical workers fell chiefly in the social 
science group. 

In a further attempt to define the char-
acteristics of reference work, a new classi-
fication was devised, one based on the 
historical period in which the questions fell. 
For this purpose the following divisions 
were set u p : 

I Questions pertaining to events, tech-
niques, aesthetic or technical creations, or 
persons, in the current year 

II Questions pertaining to the twentieth 
century 

I I I Questions pertaining to modern times 
(i.e., from the Reformation to the end of the 

1 1 Haygood, William Converse, ed. Who Uses the 
Public Library: A Survey of the Patrons of the Circu-
lation and Reference Departments of the New York 
Public Library. (Studies in L i b r a r y Science.) Chi-
cago, U n i v e r s i t y of Chicago P r e s s , 1938, p. 72. 

JANUARY, 1946 47 



O C C U R R E N C E O F R E F E R E N C E Q U E S T I O N S , AS R E V E A L E D BY S E V E R A L E N Q U I R I E S 

D.C. Classes Present Study Charters Rozendal Conner Chait 

ooo 
ioo 
200 
300 
400 
500-600.. 
700 
800 
900 

5 - 5 % 

1 - 5 

1.8 
3 7 - 3 

3 - I 
22.9 
6.7 
4 - 7 

16.5 

2 - 5 % 
1 - 5 
2.2 

22.7 
2.4 

18.0 
11.0 
12.5 
24.2 

4 - 2 % 
0.9 
1.8 

21.7 
4.9 

12.3 
8.1 

11.4 
27.2 

7.8* 

2.0% 
1.0 
2.8 

21.2 
2.0 

3 3 . 0 
6.8 

11.0 
17.0 

3 - O % 

0.5 
I - 7 

2 9 - 5 
1.2 

22.0 
6.2 
4 - 3 

31.6 

* Local and library questions 

nineteenth century) 
IV Questions pertaining to the middle ages 
V Questions pertaining to ancient times 
VI Questions pertaining to no one time 

period; those which cover several periods; 
those which are not readily classifiable by a 
time scheme 

V I I Questions pertaining to future events. 

It should be noted that the time periods are 
by no means equal in length. Whereas 
types I and I I combined cover less than half 
a century, types I I I , IV, and V each cover 
several hundred years. W h e n the questions 
are classified by this method, it becomes 
clear that most patrons ask about events 
or creations relevant to the present century. 
Seventy-two per cent of the questions were 
classified as types I and I I . I t is evident, 
also, that in the subject fields where the 
largest numbers of questions are asked, the 
"time" emphasis is found in the most recent 
categories. T h u s , 343 of the 383 questions 
in the social sciences were classified as types 
I and I I , and 145 of the 179 useful arts 
questions were classified likewise. In the 
third field, history, there is somewhat less 
emphasis on events, creations, or personali-
ties of the present century, although 93 of 
the 172 questions fall within the two most 
recent categories. 

Certain areas of inquiry do not fit into 
a time classification at all. For such ques-

tions, type V I was established. Thus, 20 
of the 31 questions on language did not fit 
into any one period. Nearly half of the 
questions on pure science were not classifia-
ble by time. 

Among the existing classifications of 
reference questions the one devised by 
Carter Alexander18 permits the most ac-
curate description of the nature of the 
question, apart from the subject matter of 
the question itself. H e lists the following 
types of reference questions: 

1. Fact type 
(a) Meaning type 
(b) Numerical or statistical type 
(c) Historical type 
(d) Exact wording type 
(e) Proper name type 

2. How to do type 
3. Trends type 
4. Supporting evidence type 
5. "All about" type 
6. Evaluation of reference type 
7. Duplication of previous work type. 

W h e n the attempt was made to classify the 
questions according to this scheme it was 
found that many questions did not fit any 
of the categories suggested by Alexander. 
In fact, only little more than half could be 
classified by this scheme. I t was necessary, 
therefore, to extend the classification to 

18 A l e x a n d e r , op. cit. 

48 COLLEGE AND RESEARCH LIBRARIES 



accommodate the remaining questions. T h e 
following additional categories were 
adopted: 

i. 
( f ) Addresses of individuals and so-

cieties 
(g) Facts about books and publishing 
(h) Reviews of books, plays, and mo-

tion pictures 
( j ) Facts about specific individuals (as 

opposed to type I (e) which was 
used for such questions as, "What 
is the name of the ruler of Iraq?" 
This new type covers biographical 
facts about persons whose names 
are known to the reader.) 

(k) Requests for illustrations 
(1) Geographical facts 

8. "General information on subject" type. 
This category was adopted in place 
of Alexander's "all about" type, 
which was designed to cover gen-
eral information questions which 
are phrased in a particular way 

9. "Review" type (for example, material 
for a review of high school algebra) 

10. Historical discussions (opposed to I (c) 
historical facts) 

11. Requests for bibliographies 
12. Requests for the translation of pas-

sages 
13. Requests for the preparation of reports 

and questionnaires. 

T h e four types in which the overwhelming 
proportion of questions fall are: 1. Fact 
type, 55 per cent; 2. H o w to do type, 10 
per cent; 4. Supporting evidence type, 8 
per cent; and 8. General information of 
subject type, 20 per cent. Together they 
account for 93 per cent of the questions. 
Illustrations of the major types follow: 

1. Fact type 
What cities have daylight saving time? 
Who invented the quadratic equation? 

2. How to do type 
How to convert a furnace into an oil 

burner 
How to remove stains from clothing 

4. Supporting evidence type 
Do colonies pay? (i.e., are they profit-

able to the mother country?) 

Arguments for and against the con-
solidation of railroads 

8. General information on subject type 
Information on cosmetics 
Material on structural engineering 

T h e sources used in answering these types 
of questions are discussed below. 

All reference librarians are aware of the 
great difficulty patrons have in making their 
exact and specific wants known. T h i s 
difficulty is not limited to the uneducated 
but extends to the more sophisticated mem-
bers of the population. An attempt was 
made in this section to determine the extent 
of the tendency and to point out specific 
types of confusion between what was re-
quested and what was wanted. 

T h e number of public library readers 
who failed to ask for exactly what they 
wanted is exceedingly small. Only 5 read-
ers out of 467 were guilty of this. In the 
academic libraries 45 readers, or nearly 21 
per cent of the total, did not ask in specific 
terms for the material which they needed. 
Several persons asked for specific titles 
or kinds of publications which they mis-
takenly thought would contain the answer 
to the question. For example, a student 
asked for "reports of the Bureau of Stand-
ards." H e wanted material on the quanti-
tative analysis of part-wool samples; this 
material was supplied by means of the 
Chemical Abstracts. T h e error made most 
frequently by students, however, was to 
request material on a large subject when 
some particular aspect was desired. Nearly 
11 per cent of readers in special libraries 
had similar difficulty in expressing their 
exact needs. 

Only a small proportion of the questions 
asked in the three types of libraries were 
unanswered. About 4 per cent of the pub-
lic library questions, 9 per cent of the 
college and university questions, and 12 per 
cent of the special library questions were 
unanswered or only partially answered. 

JANUARY, 1946 49 



T h e reasons given most frequently by refer-
ence librarians are: 

The material was not in the library 
The material desired was too recent to be 

in print 
The question was outside the scope of the 

library. 

Sources of Reference Information 

Since the answers to many questions can 
be found in several places, from a functional 
point of view there is no source which can 
be considered better than any other except 
in terms of the needs of specific individuals. 
Furthermore, when we examine the books 
used by reference librarians in answering a 
group of questions, we must realize that the 
factors of accessibility and personal prefer-
ence play a part in their selection. T h e 
only assumption upon which a description 
of books used in reference work can be 
based is that the books represent types of 
material generally found to be useful. 

W h e n the sources used to answer the 
present sample of questions were classified 
by type, it became evident that reference 
books represented about half of the sources 
used in answering factual questions. Five 
hundred and thirty-nine sources were used 
to answer 408 such questions; 298, or 55 
per cent, were classified as reference ma-
terials. T h e dependence upon reference 
books varied in the groups of factual ques-
tions. Reference books represented one-
third of the sources used for statistical 
questions; in the groups of historical ques-
tions, exact wording questions, and proper 
name questions, reference books represented 
half, or slightly more than half, of the 
sources used. Reference books were used 
most heavily to answer questions relating to 
books and publishing and to furnish bio-
graphical facts. In the latter group, refer-
ence books constituted three-quarters of the 
source material. Non-reference, or "circu-
lating," books accounted for 18 per cent of 

the source material. Periodicals and docu-
ments each accounted for one-tenth of the 
material, while newspapers represented 3 
per cent and pamphlets 4 per cent of the 
total sources used. 

Half of the group of "how to do" ques-
tions were answered by one source of in-
formation. Fifty-three per cent of the 
sources used to answer these questions were 
non-reference materials. However, Post's 
Etiquette a n d t h e World Almanac w e r e 
frequently used, and the Readers' Guide 
to Periodical Literature was the title most 
often referred to. 

Only 21 of the 74 "supporting evidence" 
questions were answered through the use 
of one source of information. T w o hun-
dred and five sources were used in answer-
ing these questions. Of these, 31 were 
classified as reference books, 83 as non-
reference books, 64 as periodicals, 14 as 
documents, 3 as newspapers, and 10 as 
pamphlets. Non-reference books and 
periodicals combined accounted for nearly 
three-fourths of the source materials. T h e 
average question required the use of three 
sources of information. 

Thirty-eight per cent of the "general 
information on subject" questions were 
answered in one source of information. 
Classification of source material by form 
showed 26 per cent to be reference books, 
43 per cent non-reference books, 22 per 
cent periodicals, 7 per cent documents, and 
2 per cent pamphlets. T h u s , reference 
books account for about one-fourth of the 
sources used and non-reference books and 
periodicals for about two-thirds. 

Conclusion 

Classification of questions by subject is 
the method employed in five studies, in-
cluding the present one. In all of these 
studies, the Dewey Decimal Classification 
has been used for a purpose for which it 
was not originally intended. In classifying 

50 COLLEGE AND RESEARCH LIBRARIES 



the questions upon which this study is based, 
it was apparent that certain decisions or 
compromises, not entirely objective in na-
ture, had to be m a d e ; for example, does a 
question about the political ideas of J o h n 
Locke classify as political theory or modern 
philosophy? I n a college library, does a 
question about u r b a n recreational facilities 
asked by a student of sociology classify w i t h 
recreation, a division of the fine arts, or 
w i t h some phase of sociology? Such de-
cisions may alter the results of subject 
classification to a marked degree. If the 
decimal system is used for this purpose, ex-
treme care and objectivity is necessary. Of 
course, the patron's point of view will in 
some cases determine the best classification 
f o r a question. N o one is in a better posi-
tion to classify questions objectively t h a n 
the reference librarian w h o answers them. 

A problem for investigation in the field 
of reference w o r k is implied in the fore-
going. discussion; that is, the development 
of a n e w subject classification scheme for 
reference questions, or, possibly, an adapta-
tion of one of the systems of book classifica-
tion n o w in use. A comparison of, for 
example, the decimal, L i b r a r y of Congress, 
and Bliss classifications in relation to refer-
ence questions might provide guidance. 

W e have seen t h a t five classifications of 
reference questions by subject reveal that 
the fields of social science, useful arts, and 
history, contain the largest numbers of 
questions asked. O n e cannot assume that 
these areas will always be quantitatively the 
most i m p o r t a n t . Assuming that changes in 
subject emphasis will occur in time, the 
method of subject classification will con-
tinue to be useful as a means of showing 
these changes. D a t a showing the trends in 
reference w o r k over a period of years will 
be useful in planning the development of 

the book stock and in specifying the qualifi-
cations needed by the reference personnel. 
Studies of the trends in reference w o r k will 
aid library schools in keeping reference in-
struction realistic and up to date. 

I n this study an attempt was made to 
learn w h a t occupational groups made use 
of the library reference service. I t was 
pointed out that the character of the occu-
pational groups in college and university 
libraries and special libraries may be de-
termined by the n a t u r e of the institution. 
I t was, therefore, not surprising to find that 
students ask most of the questions in the 
academic libraries and that professional per-
sons are the most f r e q u e n t users of the 
special library reference service. I n the 
public library the reference service is estab-
lished to meet the needs of persons f r o m all 
occupational groups, but in the present 
study it was found that the persons who 
asked questions most frequently w e r e those 
most accustomed to the use of books, the 
professionals and students. 

T h e classification of reference questions 
by " t i m e " periods showed so clearly the 
emphasis on recent events that a refinement 
of the technique might well be applied to 
other samples of questions. I t seems par- • 
ticularly desirable to know the relationship 
of events and reference questions because of 
the implications for staff and for necessary 
resources. W h a t is the length of time 
which elapses between a " c u r r e n t event" 
and a reference question about the event? 

W e have conventionally tended to identi-
f y reference w o r k w i t h reference books. 
As has been shown, a good p a r t of refer-
ence w o r k is definitely related to reference 
books, but other materials are used as well. 
T h e r e f o r e , we might well devote increasing 
attention to the specific n a t u r e of the ma-
terials which are useful in reference work. 

JANUARY, 1946 51