DEFAULT RECORD DISPLAYS IN WEB-BASED CATALOGS1
Abstract
The composition of
bibliographic record displays may have an impact on catalog use in so far as
the failure to display certain elements of description may give users
incomplete or misleading information about the item being described. In this study, thirty-eight MARC fields and
subfields are surveyed in 122 Web-based catalogs to determine whether or not
they are included in default single-record displays. Results show that some fields are displayed
in almost every catalog surveyed, including the 100 (personal author main
entry) , 245 a and b subfields (title proper and other title information), and
260 (publication, distribution) fields.
Other fields are displayed considerably less frequently, particularly
the 490 (untraced series title) and 020 (ISBN) fields. Similar types of fields, such as title
fields, are sometimes treated inconsistently, in that some are displayed very
frequently, and some are not.
Introduction
The display of bibliographic
records in library catalogs may play a crucial role in the ability of users to
find and select items that may be useful to them. The types of information provided in
descriptions are based on objectives, principles, and standards developed over
centuries of practice in the cataloging community to meet the needs of catalog
users. The International Standard Bibliographic Description (ISBD)
functions to normalize the information included in bibliographic descriptions
on an international basis [1]. It also,
to some extent, attempts to normalize the format in which this information is
presented. In card catalogs, the
provision of standard displays is supported by convention and common practice
in both the selection and presentation of types of bibliographic information on
cards. In the online catalog
environment, however, deviation from this standardization has occurred.
While flexibility has proven to be one of the great advantages
of the online environment, it may also, however inadvertently, contribute to
deviation from the standards for description supported by cataloging principles
and objectives. One reason for this is
that flexibility allows individual catalog designers to “customize”
bibliographic records displays, showing as much or as little of a description
as they deem necessary. Customization
resulting in omission of meaningful elements of description may undermine
catalog principles and objectives by displaying records that give users
incomplete or misleading information about the item being described.
This study poses critical questions
that will serve as a means of investigating the potential of customization in
an online environment to obscure – if not undermine – the function of
standardized description as it relates to the fulfillment of the intended
principles and objectives of the catalog; principles and objectives designed to
meet the needs of catalog users. First,
to what extent do current default displays of bibliographic records meet the
standards and objectives set out for them by the cataloging community? Second, and more specifically, to what extent
do current default displays of bibliographic records not include specific types
of data specified in standards regulating bibliographic descriptions? Third, in light of previous research findings
regarding user preferences for elements of display, to what extent do current
default displays of bibliographic records accommodate users’ expressed desires
for the presence of specific elements?
The research presented in this paper
addresses these questions by investigating a basic set of elements of
description in the form of MARC fields and subfields to determine how
frequently they are included in default record displays in web-based
catalogs. Specific research questions
include: 1) How frequently is each of a basic set of MARC
fields and subfields displayed in default record displays in web-based
catalogs?; 2) What is the average (mean and median) number
of a basic set of MARC fields and subfields displayed in web-based catalog
default displays?; and 3) To what extent
do elements appearing in current default displays of bibliographic records
include elements that users identify as important?
Related Research
Gregory J. Wool, Bart Austhof, Anita Breckbill, and B.
Larry Mozer performed a study of bibliographic records
in which they tabulated differences between the
Wool later
investigated displays of six Library of Congress created bibliographic records
at 36 research-level online catalogs in
Some research has
been done investigating the actual use of elements of description by catalog
users. David H. Thomas [4], provides an
extensive analytical review of this research, as well as pertinent research
from the human-computer interaction literature. He concludes that little
research has been done that actually provides empirical data on which elements
of a bibliographic description should be displayed and how they should be
displayed. In an effort to provide such
empirical data, Thomas conducted an extensive controlled experiment testing
four different record display options.
In this experiment, the four different display options, including
labeled versus unlabeled displays, as well as brief versus long displays, were
presented to four different groups of study participants who were given a
particular task to perform using the test displays. Judgments as to the use of data elements were
also collected from participants.
The results of the
research indicate that the study participants, given their topically oriented
search task, made extensive use of title and subject fields. Other results suggest that, for the type of
topical inquiry conducted in the study, brief catalog displays with subject-rich
content may have certain advantages over to other types of catalog
displays.
A group of
research papers focusing on bibliographic displays was generated at the Faculty
of Information Studies, University of Toronto for the Bibliographic Elements
and Displays Project (see: http://www.fis.utoronto.ca/research/programs/displays/index.htm). Annie T. Luk
performed focus group interviews investigating various aspects of bibliographic
record displays [5]. In one part of the
study, she identified thirty-nine elements of bibliographic records and asked
study participants to rank them as to how important they were to see in a display. She studied two groups of users, Chinese
language speakers and English language speakers. Chinese and English language speakers ranked
bibliographic elements from most important to least. Table 1 summarizes the findings of the top
twenty elements identified by both sets of users. The two user groups showed much agreement, in
that fifteen elements out of twenty were identified by both groups.
Table 1
Preferred
Bibliographic Elements*
Chinese Speaker English Speaker
Bibliographic Elements Ranking Ranking
______________________________________________________________
______________________________________________________________
Title 1 1
Author 2 2
Call Number 3 4
Name of Publisher 4 9
Summary 5 3
Subjects 6 6
Type of Material 7 7
Date of Publication 8 5
ISBN 9 NA
Reading Level 10 8
Other Authors or Editor 11 10
Edition Information 12 12
Language of Text 13 NA
Genre 14 NA
Series Title 15 NA
Subtitle 16 13
Table of Contents 17 NA
Place of Publication 18 14
Document Abstract 19 NA
Number of Pages or Parts 20 11
Nature of Contents NA 15
Book Reviews NA 16
Illustration or other
physical details NA 17
Contents Notes NA 18
Keywords from the Book Index NA 19
Author or Editor Information NA 20
______________________________________________________________
*Data summarized from Luk focus
group study on preferred bibliographic elements [5].
Using information
gathered in the Luk study, Richard W. Kopak and Joan M. Cherry evaluated three prototype record
displays using focus groups for the Bibliographic Elements and Displays Project
[6]. Findings pertinent here included a
strong preference among study participants for the inclusion of summaries in
record displays, as well as a strong preference for a text-based display,
similar to those offered in current online catalogs, as opposed to a graphical
prototype display that included a title-page like representation.
A series of
research projects at the
A second study
generated a checklist of bibliographic elements that were used in a survey of
long record displays in library catalogs.
Lynne C. Howarth and Joseph P. Cox [8] used
this checklist of twenty-four bibliographic elements, that included author,
title proper, other title information, etc., derived from the second level
description described in the Anglo-American Cataloguing Rules, 2nd
ed. (AACR2) [9]. They
searched for ten specific records in ten public library catalogs looking for
the presence or absence of the bibliographic element types in each
catalog. Specific MARC fields were not
studied. Results showed that all library
catalogs surveyed displayed the following elements: titles (title proper and other title
information), first place of publication, first publisher name, date of
publication, extent of the item, and notes about the item. Other elements were not consistently
displayed.
The last study
discussed here was identical to the second study, except that twelve academic
libraries were surveyed [10]. Results
showed that all library catalogs surveyed displayed the following
elements: authors, title proper,
parallel title, other title information, edition statement, first place of
publication, first publisher, date of publication, extent of item, statement of
responsibility relating to series, series numbering, subseries title, subseries
numbering, and notes. Only one library
did not display statements of responsibility relating to the document as a
whole.
In 1997, Jimmie
Lundgren and Betsy Simpson reported the results of a survey of faculty on the
usefulness of seventeen elements appearing in cataloging records [11]. The ten most highly ranked elements are
listed here in order of perceived usefulness4: title, author, date published, subject terms,
other authors or editors, series, summary notes, contents notes, standard
numbers (ISBN, ISSN, etc.), publisher.
Lundgren and Simpson pursued a similar type of study two years later
looking at graduate students’ perceptions of the usefulness of elements of
information for Internet resources [12].
Results of the survey indicate that graduate students rate the
usefulness of the ten top elements in the following order, most highly rated
first: title, primary
author/editor/creator, internet address, summary note or abstract, assigned
subject terms, note on level of info (full vs. partial), titles of related
print works in library, date of creation or publication, access limits (if
subscription required), and additional authors and editors.
In a similar
study, Carolyn Larson and Linda Arret investigated
descriptive elements that reference librarians view as most important to
include in descriptions of networked resources [13]. Two hundred reference librarians from a
variety of library types completed the survey.
Elements considered “essential” or “often useful” by two-thirds or more of
the respondents for both searching and display included: title, URL, subject keywords, author/creator,
date of last update, index/keyword search of resource/link to index/keyword
search of resource, subject controlled vocabulary, time period covered by
resource, language, and table of contents/link to table of contents. When the researchers opened a discussion on
the topic of elements essential for description at an open meeting at the
American Library Association Annual Meeting (July 2000), they discovered
general agreement among those present that all of the elements listed in the
survey needed to be present in record displays.
The research
reviewed in this section begins to address issues related to the composition of
displays of bibliographic records in online library catalogs. Although research
projects have investigated MARC fields displayed in library catalogs, these
projects have been limited to a small number of catalogs from academic and
research libraries only. Furthermore,
these projects have tended to focus on long record displays, as opposed to
default record displays. The research
presented in this paper attempts to expand what we know about the composition
of catalog displays by looking at a large number of catalogs, selected
randomly, and also by looking at default record displays as opposed to long
record displays.
Default record
displays are of particular interest for two major reasons. First, it is likely that since default
displays are the first and often the only record displays seen by catalog
users, they are viewed more frequently than long displays. Second, if deviation from display standards
occurs, it is more likely that it occurs in default displays than long
displays. This likelihood can be
explained by the functions accorded to each type of display. If catalog designers include both a default
display and a long display, the default display is provided specifically to
give users a briefer record to review in the course of searching. The long display, in contrast, often exists
only to reveal most or all of the MARC fields existing in the record.
Several of the
research projects reviewed in this section also asked users how important it
was to see particular descriptive elements in a record display. Although these results are far from
conclusive, they give us a preliminary glimpse at elements that may be viewed
as important by a large number of catalog users. Because the research presented here looks at
a wide variety of randomly selected catalogs, it may be used to compare the
frequency with which elements are displayed in default displays with the
perceived importance by the users surveyed in the projects reviewed here.
Methodology
Sampling
In this study, 122 web-based
catalogs were surveyed to discover which of a basic selection of MARC fields
and subfields were displayed in default single-record displays (see Study web
site with catalog names, vendors, and additional statistics and information
at: http://purl.oclc.org/net/default.records). The sample of 122 catalogs represented ten
percent of the total number of catalogs available via the Library Web-based OPACs (WebCats)
website on
Table 2
Library
Types Surveyed
Library Types Surveyed Number Percent
___________________________________________________
___________________________________________________
Academic 72 59
Public 29 24
Special 12 10
Multiple-Library Type
Consortium 7 6
School Consortium 1 1
State 1 1
TOTAL 122 101*
___________________________________________________
Note
*101 percent due
to rounding error.
Web-based catalogs were selected for the
survey because it seemed likely that they represented the most current
judgments regarding record displays in the libraries surveyed. It should be noted that because academic
libraries, and to a lesser extent, public libraries, are more likely to have
web-based catalogs than other types of libraries, the sample of catalogs
included here is not necessarily representative of catalogs in general. All vendors possible were surveyed to
accommodate variance in default displays due to vendor recommendations
regarding default displays.
A stratified random sample of web-based
catalogs was created in the following manner:
a printout was made on
The survey was limited to
A subset of MARC fields was selected for
the survey because so many MARC fields exist that it would be impractical to
survey them all. Because it was
necessary to select a subset of fields, fields were selected that were judged
by the researchers as essential or core elements of record displays. The fields selected were, thus, based in
large part on the list of elements presented in AACR2’s second level
description, and also the fields required for the core bibliographic record
standards. An attempt was made to
include several fields commonly found in serials records, although the number
of serials fields included was limited because of the difficulty of finding an
adequate number of serials records in the catalogs surveyed. Fields used for specific formats of materials
only, such as those used for cartographic materials or microforms, were not
included in the survey.
The final list of fields surveyed was also
somewhat limited because of the difficulty of finding examples of particular
fields in the catalogs surveyed. For
example, it proved to be extremely time-consuming to find records containing
711 and 785 fields. Thus, these fields, although of interest, were dropped from
the initial list of MARC fields studied.
Informal observation during the survey indicates that, in general, if a
710 field was displayed in a catalog, then the 711 field was also displayed;
and if the 780 field was displayed, then the 785 field was also displayed.
Thirty-eight MARC fields and subfields
surveyed included (see annotated list of fields in Appendix 1):
Table 3
MARC
Fields and Subfields Surveyed
Fields and
Subfields
______________________________________________________________
______________________________________________________________
020 245 ‡c 440 610 730
100 245 ‡h 490 0 611 740
110 246 490
1 630 780
111 250 500 650 830
130 260 504 651 856
240 300 505 700 call number
245 ‡a 310 520 700 ‡t
245 ‡b 362 600 710
TOTAL Fields and
Subfields: 38
______________________________________________________________
The display of call number
was surveyed, regardless of the field it was taken from. (See Appendix 1 for this list annotated with
the type of information recorded in the field or subfield.)
Data Collection
Data were collected from
sample catalogs over a two-year period:
from 1998 to 2000. In each
catalog, specific subfields were surveyed only when noted, and not
otherwise. In other words, for the other
fields, the presence of any subfield in a field was tallied as “field
displayed”, even though on a rare occasion specific subfields were not
displayed. An example of this is the ‡d in the 100, 600, and 700
fields (dates associated with personal name); in one catalog, this subfield was
not displayed, although other subfields were.
Another example is the 440/490/830 ‡v (series volume number), which was not displayed in one
catalog, although other subfields were.
The suppression of specific subfields from the default displays was,
however, noted only rarely.
To determine whether or not a particular
field was displayed in a default record display, searches were performed to
retrieve a set of pre-selected MARC records.
While it was not necessary to use a particular set of MARC records, it
sped data collection to begin the survey of each catalog with searches for
specific records, as mentioned above.
The bibliographic records in the search set, most of which were
cataloged by the Library of Congress and obtained from the OCLC database,
represented monographs, monographs in a series, serials and nonbook
materials. A search was conducted for
search set records in every catalog surveyed.
If the record was in the catalog, the default record display for that
record was examined.
A default record display was determined to
be the first display of a single bibliographic record on a screen by
itself. If no single-record display
was available, the multiple display record was considered the default record
display. If specific fields were unable
to be surveyed because the catalog lacked a MARC record in the search set, then
searches were conducted for records within that catalog that contained those
fields. For example, if no record in the
search set was retrieved that contained a 611 field (conference as subject field),
then a subject or keyword search with the term “conference” or “workshop” was
performed. Strategies for discovering
records containing fields of interest varied depending on the field
searched. Whether or not a record
contained a specific field was confirmed when a MARC display was available for
that record in the OPAC.
Unfortunately, not all catalogs make MARC
displays available to catalog users. Of
the catalogs sampled in this research, forty-six out of 122 catalogs, or
thirty-eight percent, did not make MARC displays available to catalog
users. The researchers, using their
judgment and knowledge of cataloging and MARC records, made educated guesses as
to the fields displayed in catalogs that did not display MARC fields. If available, long displays were consulted
and compared to sample set records. For
those records examined that were not in the sample set, they were compared to
the records as they appeared in the OCLC database. The records were examined closely to
determine whether or not they had exactly the same appearance as the OCLC
records. When records in the sample
catalog were, for all fields displayed, identical to the sample records, then
an assumption was made that the MARC fields displayed were also identical to
the MARC fields in the sample record.
Because of the possibility for error, two instances of a particular
field were searched in every catalog that did not provide a MARC display so
that the reliability of determinations made regarding the presence or absence
of MARC fields displayed would be increased.
If the researchers could not find an
instance of a field, or could not determine the MARC tag of the field
displayed, which was sometimes the case for series fields displayed in catalogs
that did not offer a MARC display option, the value for that field in that
catalog was not tallied; in other words, it was treated as if it did not exist
in that catalog and was coded as “not available” (NA). Treatment of NA MARC field values is
discussed in the Results section below.
In some catalogs, non-standard records
were included for journals and periodicals.
In catalogs containing this type of serial record, records for other
types of serial publications, for example, yearbooks and other annuals, were
searched. If standard records were
included for these publications, then the fields common to serials (310, 362,
and 780) were surveyed. If no standard
records for serials were included, these fields were not included for that
catalog.
Results and Discussion of
Individual Findings
Display
of Individual Fields.
In
this section, findings for the display of individual fields are presented and
discussed. In each of the tables, one
set of data is presented for all of the catalogs surveyed, and another set is
presented for catalogs that featured MARC displays only. Thus, the MARC display catalog data is a
subset of the total catalogs data. The
MARC display catalog data is presented separately because in these catalogs the
researchers were certain of the fields displayed. In catalogs without MARC displays (forty-six
out of 122 catalogs), the researchers had to make educated judgments regarding
field displayed. Readers who do not wish
to rely on the judgments of the researchers may prefer to consider the MARC
display catalog data only. However, when
the results are discussed, the discussion will consider data for all of the
catalogs surveyed.
In the tables presented in this section
for results of specific fields, the MARC field tag is presented, followed by “n”
(the number of catalogs in which data was able to be collected about that
field); “Number (#)” (the number of catalogs displaying
that field in the default record display); and “Percent (%)” (the
percent of catalogs displaying that field in the default record display.6
Findings
for Author Fields.
All
catalogs surveyed display the 100 field. Most display other 1XX author
fields. A majority of catalogs surveyed
also display the 700 and 710 fields.
This result is not surprising, given the importance of authors and other
contributors to a work and the frequency with which they are sought in
catalogs. However, given the fact that
these fields represent elements that play identical roles in item description,
it is perplexing that none of the results for these fields is the same, and
that some libraries make different decisions regarding the display of these
fields. The rationale behind a library’s
decision to display a 700 field but not a 710 field, for example, is unclear.
Display of 7XX fields is somewhat less
frequent than display of 1XX fields.
This may have important ramifications for users searching for
information contained in these fields. If users search on data
contained in fields that are not displayed, it is possible that they will be
confused as to why those records are retrieved.
Table 4
Findings
for Author Fields
Total Catalogs MARC Display Catalogs Only
MARC n # % n # %
Tag
_____________________________________________________________________
_____________________________________________________________________
100 122 122 100 76 76 100
110 122 117 96 76 71 93
111 117
113 97
72 68 94
700 122
106 87 76 62 82
710 122 103 84 76 59 78
_____________________________________________________________________
Note: Total
catalogs surveyed (122); Total MARC display catalogs surveyed (76)
n - number of catalogs in which data was collected for the
field; # - number of catalogs displaying the field
% - percent of catalogs displaying the field
______________________________________________________________
Findings
for Title Fields.
Findings
for title fields reflect a diversity of opinion with respect to how important
it is to display various types of titles.
All catalogs surveyed displayed the a and b subfields of the 245 field,
indicating the importance accorded title proper and other title information,
which may be regarded as the attributes that most closely identify the
item. However, no such endorsement
appears for the other title fields surveyed.
The absence of these fields from default
record displays may make it difficult for users to discover the identity of the
item described. For example, a main
entry uniform title (130 field) is not displayed in eight out of 120 catalogs
surveyed. In a record for an edition of
the Arabian Nights entitled The Thousand and One Nights, the
130 field carries the information that identifies it as an edition of the Arabian
Nights. The presence of a
subfield t in a 700 field makes reference to a work contained in the item or to
a work related to the item being described.
Again, if the subfield t is not displayed, users may not be given
sufficient information about the nature of the item described. In addition, title fields are frequently
indexed in keyword searches. As
mentioned above, if users are able to search and retrieve on a field not
displayed, particularly a search such as “Arabian nights”, then confusion as to
why specific records were retrieved may result.
The 245 subfield c, statement of
responsibility, is somewhat of an anomaly. Although it appears in a title
field, it carries author and other contributor information. At times, the nature or identity of the item
is made clear only in the statement of responsibility; for example, sometimes
the indication that an item is an adaptation or a translation appears in the
statement of responsibility. In
addition, if the 7XX field carries the name or names of persons who appear only
in the statement of responsibility and the statement of responsibility itself
is not displayed, it may be unclear as to why the 7XX names appear in the
record at all. Because the statement of
responsibility has a variety of critical identification functions, it is
unclear why any library would choose not to display it. It was found that eight
percent of catalogs surveyed do not display the statement of responsibility.
Table 5
Findings
for Title Fields
Total Catalogs MARC Display Catalogs Only
MARC n # % n # %
Tag
_____________________________________________________________________
_____________________________________________________________________
130 120 112 93 75 68 91
240 122 106 87 75 64 85
245‡a
& b 122 122 100 76 76 100
245‡c 122 112
92 76 68 89
245‡h 121 111
92 75 66 88
246 120
91 76 75 51 68
700‡t 117 97 83 71
53 75
730 120 95 79 75 55 73
740 118 91 77 76 54 71
780 108 89 82 68 50 74
_____________________________________________________________________
Note: Total
catalogs surveyed (122); Total MARC display catalogs surveyed (76)
n - number of catalogs in which data was collected for the
field; # - number of catalogs displaying the field
% - percent of catalogs displaying the field
Findings
for Subject Fields.
While
the title fields results show a diversity of opinion within individual
catalogs, subject field results show the opposite; in other words, within
individual catalogs, subject fields are treated as a group and are, generally
speaking, either displayed or not displayed.
Most of the catalogs (92 percent) display subject headings. Call numbers are displayed in nearly all of
the catalogs (98 percent).
Subject fields frequently convey
information critical in identifying the subject content, form, or genre of the
item cataloged. In many catalog use
studies, subject searches have been discovered to be the most frequently
performed search (e.g., [15]). In
addition, subject fields, like title fields, are often indexed in keyword
searches. It would not be surprising if
users doing subject searches were confused to find a lack of any statement
regarding subject content in the records they retrieved. Further, while the retrieval function of
subject headings is frequently noted, subject headings have a descriptive
function as well. They make known to the user the nature of the item in terms
of its subject content, form, or genre.
When subject headings are absent in displays, the presence of works
whose titles convey little or no subject information may be, at best,
mystifying to users or, at worst, misleading.
Table 6
Findings
for Subject Fields
Total Catalogs MARC Display Catalogs Only
MARC n # % n # %
Tag
_____________________________________________________________________
_____________________________________________________________________
600 122 113 93 76 69 91
610 122 112 92 76 69 91
611 118 108 92 73
66 90
630 120 110 92 76 69 91
650 122 113 93 76 70 92
651 122 113 93 76 70 92
Call no. 122 119
98 76 74 97
_____________________________________________________________________
Note: Total
catalogs surveyed (122); Total MARC display catalogs surveyed (76)
n - number of catalogs in which data was collected for the
field; # - number of catalogs displaying the field
% -
percent of catalogs displaying the field
Findings for Series Fields.
Results for series fields are similar to title fields results in that there
is a relatively wide divergence in display practice for the various
fields. Almost 90 percent of the
catalogs surveyed display the 440 field, which is a series title traced as it
appears on the item, while only 76 percent of catalogs surveyed display field
the 490 _0 field, which is a series title transcribed from the item that is not
traced.
Series tracing practice is somewhat
complex. Presumably, only two series
fields, the 440 and 830 fields, are traced or indexed so that they are
searchable in a catalog. The divergence
between the results for the 440 field and the 830 field is, therefore, somewhat
perplexing, considering that in this respect the status and purpose of the two
fields is identical. Why some catalog
designers decided to display the 440 field and not the 830 field is
unclear.
The 490 _0 is a series title that is
transcribed, but not intended to be used as an access point. If this field is not displayed, users have no
indication that the item is indeed part of a series. Because a large percent of catalogs surveyed
decided not to display the 490 _0 field, it must be presumed that catalog
designers making this decision do not consider an item’s identity as part of a
series to be significant. The 490 _1 is
a series title that is transcribed, but is traced in its authorized form in the
830 field. It is somewhat bewildering
that more catalogs display the 490 _1 field than display the 490 _0 field,
because in records with a 490 _1, the 830 field is also often displayed, so for
these records there may already be an indication that the item is part of a
series. Most bewildering of all are the
fourteen catalogs that display the untraced series field (490 _1) but do not
display the authorized form of that same series in the 830 field. It can only be surmised that the designers
making such decisions for display of series fields do not understand the
function of these fields, or are compelled for some other reason not to display
the 830 field.
Series fields are the most difficult of
all field types to survey in catalogs that do not make available a MARC
display. This absence of a MARC display accounts for the low number of catalogs
in which data was collected for the field (n), particularly for the 490
_0. One reason for surveying
difficulties is that series tracing practice is one of the least standard areas
of cataloging. As a result, if one were
looking at a Library of Congress (LC) record and comparing it to a local
catalog record, one could not be certain that local catalogers made the same
decision with respect to series tracing as LC.
In record displays where only a single series field displayed, it is
sometimes impossible to tell whether that field is a 440, a 490 _0, or an 830. Thus, during the data collection process it
was common for us to be unable to make a determination, particularly with
respect to the 490 _0.
Table 7
Findings
for Series Fields
Total Catalogs MARC Display Catalogs Only
MARC n # % n # %
Tag
_____________________________________________________________________
_____________________________________________________________________
440 121 108 89 75 64 85
490 _1 114 78 68 73 45 62
490 _0 83
54 65
69 49 71
830 116 88 76 73 55 75
_____________________________________________________________________
Note: Total
catalogs surveyed (122); Total MARC display catalogs surveyed (76)
n - number of catalogs in which data was collected for the
field; # - number of catalogs displaying the field
% - percent of catalogs displaying the field
Findings for Notes Fields.
Results for notes fields reflect varying opinions with respect to the importance of displaying notes fields. Given the varying purposes of these fields, this finding is not wholly surprising. The 500 and 504 fields were least likely to be displayed; 21 and 24 percent of libraries, respectively, chose not to display them in their default record displays. However, these two fields are quite different in nature. The 504 field is a highly circumscribed field, containing information about bibliographies and indexes only. On the other hand, the 500 field, a general note, contains a wide variety of information, from a note identifying the source of the title proper to a note indicating the nature, scope, or artistic form of the item. Given the different nature of the two fields, the similarity of the results is surprising.
Because one of the functions of the
general note is to cite previous editions or works, a decision not to display
the 500 field often obscures the nature or history of the item. For example, the 500 field may be the only
place in the record identifying the item as an adaptation or a
translation. Depending on the extent of
subject indexing present in a record, the 500 field may also be the only place
in the record indicating the form of the item, for example, identifying the
item as a comedy or a documentary.
Furthermore, not displaying the 500 field may hide the only field in a
record that differentiates this item from other items in the library’s
collection. Differentiation of items is
a critical function of a bibliographic description. The decision not to display the 500 field is
a significant one given the essential nature of the information it may contain.
The 505 (contents) and 520 (summary)
fields, both of which indicate the subject contents of an item, are displayed
more frequently than the other notes field studied. However, given the enormous pressure for
catalogers to improve subject access to items by providing 505 or 520 notes for
all items, it is surprising that even as many as 10 percent of the catalogs
surveyed chose not to display these fields.
This result is even more significant given the findings of the
Table 8
Findings
for Notes Fields
Total Catalogs MARC Display Catalogs Only
MARC n # % n # %
Tag
_____________________________________________________________________
_____________________________________________________________________
500 122 96 79 76 52 68
504 122
93 76
76 54 71
505 122
105 86 76 61 80
520 122 108 89 76 65 84
_____________________________________________________________________
Note: Total
catalogs surveyed (122); Total MARC display catalogs surveyed (76)
n - number of catalogs in which data was collected for the
field; # - number of catalogs displaying the field
% - percent of catalogs displaying the field
Findings for Other Fields
Perhaps not unexpectedly, display practice for the remaining fields varies
widely. Like notes fields, these fields
have a variety of functions, and their importance may be viewed differently by
catalog designers. The most frequently
displayed field is the publication, distribution field (260); 99 percent of
catalogs surveyed display it. Given the
significance of the date of publication in dating the content of an item, and
the fact that without the display of this information the item could be
unrecognizable, it makes sense that only a single catalog does not display
it. Another critical field, the 250
field (edition statement), is displayed in 85 percent of the catalogs
surveyed. Without the edition statement,
one item may be indistinguishable from another item having virtually the same
description. Again, one of the fundamental
functions of a bibliographic description is to distinguish one item from
another. Without the information this
field provides, catalogs are less able to fulfill this basic function.
The 300 field (physical description)
provides information about both the physical nature of an item, such as number
of pages or physical parts, and the intellectual or artistic content of the
item, for example, whether it is in black and white or in color. Because this field contains information about
the intellectual or artistic content of an item, it may also be the only field
in the item to distinguish one item from another. While most catalogs (89 percent) display this
field, it is surprising that any catalog would not display it, given its
functions of distinguishing one item from another and revealing essential
content information about the item.
The ISBN (020) is displayed in only 56
percent of the catalogs. While the ISBN
may be used to distinguish one item from another, usually other types of
information are used for this purpose.
It is unclear how the absence of this field would affect users. It is certainly a field used frequently by
library professionals, and when available, some users do use it for
searching. Obviously, many catalog
designers believe that it is not an essential element of a record display.
Two fields, the 310 (current publication
frequency) and 362 (dates of publication/sequential designation) fields, are
included in the study specifically because they are associated only with
records for serials. Serials are
commonly held items; and, we were anxious to learn whether or not fields
associated with serials were displayed.
In the survey, one quarter of the catalogs included did not display
these fields. Again, one could argue the
essential nature of these fields.
Serials are generally distinguished one from another by the use of
varying uniform titles. Thus, one could not argue that the 310 and 362 fields
are needed to distinguish one serial from another. However, they provide for the user
information about the nature and identity of the serial that might be more
recognizable than differences in uniform titles.
The 856 field (electronic location and
access) was surveyed with great interest on the part of the researchers because
it is a relatively new field, and its presence indicates an attention to
providing access to Web resources. The
856 field is displayed in three quarters of the catalogs surveyed. Given both
the newness of the field and its function, this is, perhaps, an unexpectedly
high number. It would be a mistake to assume that this finding entails that 75
percent of the libraries surveyed are actually selecting and cataloging Web
resources. In some catalogs surveyed, it
was extremely difficult to find an 856 field; only one or two records were
found containing it. This indicates that
these libraries are not actively collecting and cataloging Web resources.
Rather, they are using records for items that have a reference to a Web
resource only. However, in many catalogs
surveyed, it was obvious that Web resources were being collected and cataloged
actively.
It is curious that libraries would choose
to include 856 fields in MARC records and then not choose to display them. Because for many items, the 856 field is the
only source of information about the location of the item, it would seem that
not displaying it would prohibit users from actually finding and using the
item. This practice defeats the purpose of having included the record in the
catalog in the first place. However, 12
percent of the catalogs surveyed included the 856 field in their records, but
chose not to display it.
Table 9
Findings
for Other Fields
Total Catalogs MARC Display Catalogs Only
MARC n # % n # %
Tag
_____________________________________________________________________
_____________________________________________________________________
020 122 68 56 76 40 53
250 122 104 85 76 59 78
260 122 121 99 76
75 99
300 122 109 89 76 66 87
310 112 82 73 72 49 68
362 109 73 67 71 47 66
856 93
82 88 56 50 89
_____________________________________________________________________
Note: Total
catalogs surveyed (122); Total MARC display catalogs surveyed (76)
n - number of catalogs in which data was collected for the
field; # - number of catalogs displaying the field
% -
percent of catalogs displaying the field
Summary
Field Findings.
A summary of the results for
individual fields presented throughout this paper appears in Table 10. Eighteen catalogs are not included in this
table because of the large number of fields in these catalogs for which data
was not able to be collected. In Table
10, fields are listed in descending order of frequency displayed; in other
words, the fields displayed most frequently in catalogs surveyed come
first. Overall, 53 percent of fields
surveyed are displayed in 90-100 percent of the catalogs surveyed. Not surprisingly, these fields are among
those that might be regarded as essential for retrieval and identification of
the item. The 1XX and 245 fields are
frequently used for retrieval, and would be crucial for identification of the
item. The 260 fields are also crucial
for identification of the item. The 6XX
fields are essential both for understanding the subject content of the item and
for retrieval based on subject content.
Call number, essential for location of the item, is also displayed in 90
percent or more of catalogs surveyed.
Series, notes, and author and title added entry fields are among the
least frequently displayed fields.
Table 10
Fields
Displayed Across Catalogs
Fields Displayed % of Catalogs Number of Percent of
___________________Displaying These Fields______Fields_________Fields___
___________________________________________________________________
100, 245a, 245b, 260, 90-100% 20 53
call no., 110, 111, 130,
600, 650, 651, 610, 611,
611, 630, 300, 440, 245c,
240, 520, 245h
700, 856, 505, 710, 250, 80-89% 11 28
700t, 730, 246, 500, 740
830, 504, 310,
490 _0 70-79% 4 11
362, 490 _1 60-69% 2 5
020 50-59% 1 3
___________________________________________________________________
Total catalogs tallied:
104. Eighteen catalogs were not
included in the overall statistics because of missing field information.
Summary Catalog Findings
A
summary of results for the total number of fields displayed per catalog appears
in Figure 1. Again, eighteen catalogs
were not included in this tally because of missing field information; results
for 104 catalogs are reflected in the figure.
In addition, eleven catalogs are not included in the Figure because the
distribution of fields displayed is highly skewed. Approximately 90 percent of the 104 catalogs
displayed twenty-seven or more fields.
The remaining eleven catalogs displayed between twenty-six and nine
fields. Grouping these eleven catalogs
in a single column would disproportionately reflect their significance in the
Figure; as a consequence, they were left out.
Out of a total possible thirty-eight
fields/subfields studied, the mean number of fields/subfields displayed across
catalogs is thirty-two. The median is thirty-five, and the mode
thirty-six. Obviously, most of the
library catalogs surveyed display most of the fields studied. Forty-four, or 42 percent of catalogs
surveyed, displayed nearly all (thirty-six to thirty-eight) of the fields
studied, and twenty-six, or 25 percent of catalogs surveyed, displayed
thirty-three to thirty-five fields. Of
interest are the catalogs displaying twenty-six or fewer fields. The number of fields displayed for these
eleven catalogs ranges from as few as nine fields to as many as
twenty-six. Nine out of the eleven
catalogs in this category displayed sixteen or fewer fields.
Fig.
1.—Total MARC Fields Displayed in Catalogs
General Discussion
and Directions for Future Research
In the discussion of specific
results in the previous section, unusual or seemingly inconsistent choices in
the display of fields in individual catalogs were noted. Many of these choices have to do with
selecting different display treatments for fields that have either very similar
or identical purposes. For example, in
some catalogs, 130 fields (uniform title as main entry) are displayed, but not
240 fields (uniform title); 700 fields (personal author added entry) are
displayed, but not 710 fields (corporate body added entry); 740 fields (title
added entry) are displayed, but not 246 fields (title added entry). These choices are puzzling given the similar
functions of the fields involved.
Another issue involves the display of only
part of a field. For example, catalogs
displaying 245 subfields a and b (title proper and other title information),
but not subfield c (statement of responsibility); 700 subfield a (personal name added entry),
but not subfield d (dates associated with a name; although the 700 d subfield
was not studied specifically, we noted that one catalog did not display the
subfield d with the name); 700 name subfields, but not the t (title)
subfield. In this latter case, the
meaning of the added entry is changed from being an analytical or related work
added entry to being simply a name added entry.
For some records, dropping the subfield t in a 700 field means that
several fields are displayed that look identical but, in reality, are not. One wonders what criteria, if any, catalog
designers are using when they chose to display one type of field, but not
display another field that fulfills the same function, or to display parts of
fields only, particularly when they change the meaning of the information
provided by those fields.
The third research question posed in this
study investigates the relationship between the elements displayed in catalogs
and those identified by catalog users as important or useful. While it is extremely difficult to determine
those elements that are in fact used and needed by catalog users, studies
reported in the Related Research section of this paper reveal elements rated by
users as important to include in displays of cataloging records. Eight elements may be identified as the most
frequently included in the top ten list of elements reported in the studies
discussed above. These eight elements
are: title, author, subjects, date of
publication or creation , call number/URL/Internet address (all associated with
location of item), summary, publisher, and other authors or editors.
In all of the studies, participants
identified title, author, subjects, and date of publication or creation among
the most important elements to include in a cataloging record. Location (call number and URL or Internet
address) and summary were identified in the top ten list of elements in four of
the five studies. Publisher and other
authors or editors were identified in the top ten list of elements in three of
the five studies.
Examining these elements with respect to
the frequency with which they appear in default displays, one finds that user
preferences for title (if one includes only the 245 field (subfields a and b
only), author, date, publisher, and call number are met by almost all of the
catalogs surveyed. Call number is not
displayed in two percent of catalogs surveyed; corporate and conference primary
“authors” are not displayed in four and three percent of the catalogs surveyed,
respectively. Publication information,
including date of publication, is not displayed in only a single catalog
surveyed.
Subjects, rated among the top ten most
important elements in all of the user preference surveys, are not displayed in
seven to eight percent of catalogs surveyed.
Summaries, rated among the top ten in four out of five user preference
surveys, are displayed in even fewer catalogs; eleven percent of catalogs
surveyed do not display them. Other
authors or editors, rated among the top ten elements in three out of five
studies, are not displayed in thirteen to sixteen percent of catalogs surveyed.
Given the importance users accorded these
last three types of elements, catalog designers may wish to consider changing
display options for them if they do not appear currently in default
displays. This is particularly
recommended for subjects, which appear among the top ten preferred elements in
all five user preference studies reported here.
It is also worth noting that ISBN, the element displayed least
frequently of all elements surveyed (displayed in only 56 percent of catalogs
surveyed), appears among the top ten preferred elements lists in two of the
five studies.
A variety of further areas of
investigation arise from questions surrounding the variability of information
included in default displays. Future
research could, for example, investigate reasons why different decisions are
made regarding fields that have similar functions; what criteria are being used
to judge one field as being more informative or more important than another,
such that one is deemed necessary to display and another one not; whether or
not non-displayed fields are searchable and if so, whether or not users are
confused when a search on a particular author name or other element retrieves a
record in which that element is not displayed.
Conclusion
The results of the research presented
above demonstrate vividly the impact that the flexibility inherent in
computer-based catalogs may have on standardized description. The frequency with which the fields studied
here are displayed varies from 100 percent (100, 245 $a & $b) to 56 percent
(ISBN). Large variability in displays
sometimes exists even among similar types of fields. For example, the least frequently displayed
field (246, varying form of title) appears in 76 percent of the catalogs, while
the most frequently displayed title field (245 $a & b, title proper and
other title information) appears in 100 percent of the catalogs.
Although little is known of the actual
impact of the presence or absence of elements of display on individual catalog
users, the results of this study indicate that elements users identify as
important are missing in a significant number of catalogs’ default record
displays. In addition, these displays
sometimes reflect a lack of knowledge or consistency on the part of catalog
designers.
Various standards, including the ISBD,
as noted earlier in this paper, and codes of cataloging rules such as AACR2, have been promulgated to ensure that a
minimum amount of information is included in record displays. While the particulars of these standards are
also largely unsubstantiated by user research, they do reflect the knowledge
gained by catalogers accumulated across international boundaries and centuries
of practice. Deviation from these
standards, causing the creation of inconsistent, incomplete, or
incomprehensible displays, has the potential to impede the success of a user in
successfully identifying items of interest in the catalog. Catalog designers are, thus, urged to
consider carefully decisions resulting in record displays that do not meet the
requirements set out for them by the cataloging community.
APPENDIX
1: List of MARC Fields Surveyed
020 ISBN
100 Main entry – Personal name
110 Main entry – Corporate name
111 Main entry – Meeting name
130 Main entry – Uniform title
240 Uniform title
245 Title statement, including $a, $b, $c, and $h
246 Varying form of title
250 Edition statement
260 Publication, distribution, etc.
300 Physical description
310 Current publication frequency
362 Dates of publication and/or Volume
designation
440 Series statement/added entry – Title
490 Series statement, including first indicator 0
and 1
500 General note
504 Bibliography note
505 Formatted contents note
520 Summary, etc. note
600 Subject added entry – Personal name
610 Subject added entry – Corporate name
611 Subject added entry – Meeting name
630 Subject added entry – Uniform title
650 Subject added entry – Topical term
651 Subject added entry – Geographic name
700 Added entry – Personal name
710 Added entry – Corporate name
730 Added entry – Uniform title
740 Added entry – Title
780 Preceding entry
830 Series added entry – Uniform title
Call number (no specific
field examined)
REFERENCES
1. International Federation of Library Associations, ISBD Review
Committee Working Group. ISBD
(G): General International Standard
Bibliographic Description: An Annotated
Text.
2. Wool, Gregory J.; Austhof, Bart; Breckbill, Anita; and Mozer, B.
Larry. “Cataloging Standards and Machine
Translation: A Study of Reformatted ISBD
Records in an Online Catalog.” Information Technology and Libraries
12 (December 1993): 383-403.
3. Wool, Gregory. “The Many
Faces of a Catalog Record: A Snapshot of
Bibliographic Display Practices for Monographs in Online Catalogs.” Information Technology and Libraries
15 (September 1996): 173-86.
4. Thomas, David H. “The
Effect of Interface Design on Item Selection in an Online Catalog.” Library Resources & Technical Services
45 (January 2001): 20-46.
5. Luk, Annie T. “Evaluating Bibliographic Displays from the
Users’ Point of View: A Focus Group Study.” Masters thesis. Faculty of Information Studies,
6. Kopak, Richard W. and Cherry, Joan
M. “Bibliographic Displays and Web
Catalogues: User Evaluations of Three
Prototype Displays.” The Electronic
Library 16 (October 1998): 309-23.
7. Cherry, Joan M., Williamson, Nancy J., Jones-Simmons, Carol R.,
and Gu, Xin. “OPACs in Twelve
Canadian Academic Libraries: An
Evaluation of Functional Capabilities and Interface Features.” Information Technology and Libraries
13 (September 1994): 174-91.
8. Howarth, Lynne C. and Cox, Joseph
P. Facilitating Access to Electronic
Resources: Matching Bibliographic Record
Content with Client Preferences.
1996. (http://www.fis.utoronto.ca/research/programs/displays/elements.htm#desig)
9. Joint Steering Committee for Revision of AACR. Anglo-American Cataloguing Rules. 2d
ed., 1998 revision.
10. Howarth, Lynne C.; Cox, Joseph P.;
Richard, Trina; and McConnell, Mary. Bibliographic
Elements in OPAC Displays: Monographs in
Canadian Academic Libraries.
Bibliographic Elements and Displays Project Technical Report Series, No.
1.
11. Lundgren, Jimmie and Simpson, Betsy.
“Cataloging Needs Survey for Faculty at the
12. Lundgren, Jimmie and Simpson, Betsy.
“Looking Through Users’ Eyes:
What Do Graduate Students Need to Know About Internet Resources via the
Library Catalog?” Journal of Internet
Cataloging 1, No. 4 (1999): 31-44.
13. Larson, Carolyn and Arret, Linda. “Descriptive Resource Needs from the
Reference Perspective: Report on a
Survey of US Reference Librarians for the Bicentennial Conference on
Bibliographic Control for the New Millenium.” In Bicentennial
Conference on Bibliographic Control for the New Millenium,
Confronting Challenges of Networked Resources and the Web,
14. LibDex, The Library Index
(was Library Web-Based OPACS or WebCats). (http://www.libdex.com/)
15. Matthews, Joseph R.,
Footnotes
1The
authors gratefully acknowledge the assistance of the following persons: Elizabeth S. Knight for her work on data
collection; Matthew L. Saxton for his invaluable advice regarding the
statistical analyses; Lisa M. Fusco and Harry Bruce for excellent editorial
suggestions; the anonymous referees; and the many librarians, library staff,
and vendor representatives who patiently answered questions.
2Assistant
Professor,
E-mail
infogirl@home.com.
4Results were not reported in
this manner in the article. The survey
asked faculty members to rate their level of agreement with the statement
that: “The following parts of the
catalog record for library materials are useful and should be displayed to the
user”. Levels of agreement included
“strongly agree”, “agree”, “no opinion”, “disagree”, and “strongly
disagree”. In order to create a ranked
list of elements, percentages in “agree” and “strongly agree” columns were
added together.
5A stratified random sample
was conducted because, although libraries are, in general, free to select
fields that they prefer to be displayed in their default record displays,
vendor influence may be present in that vendors frequently recommend a list of
fields to be displayed in the default display.
It was assumed that these lists of vendor recommended fields, which vary
from vendor to vendor, could have some influence on a library’s selection, and
we wanted to account for it.
6Confidence levels for all
results are extremely high given the large number of catalogs included in the
sample, the binomial nature of variables studied, and the generally limited
variability in the results. As a result,
they are not reported here.
© 2002 by The