This paper provides an overview of the purpose and uses of an abstract;
and the guidelines and processes which are helpful in creating abstracts.
An abstract is an abbreviated representation or surrogate of a data
or information resource. It contains a concise and significant summary
of the resource, and is generally intended to serve as a stand-alone
description. Coupled with pertinent bibliographic information such as
title, it provides users with supportive information for evaluation
of a data or information resource. An abstract with its accompanying
citation is a metadata record--a work of scholarship in its own right
which can be indexed, organized, accessed and used as an information
resource for research purposes.
Purposes and Uses of an Abstract
The 'traditional' purposes and uses of abstracts, either alone or
aggregated within an abstract journal or online directory, have changed
very little over the centuries that abstracts have been used. The purposes
focus on the ability to assemble, access, and evaluate relevant materials
more easily and efficiently, using summaries of source items instead
of the original items themselves.
Some specific uses include:
Thus, abstracts have evolved to serve two additional purposes:
- Serving as a current awareness source, providing rapid and accurate
communication of professional knowledge. In some cases abstracts are
available prior to release or publication of the source data and information.
In other cases, abstracts may be the only information which is accessible.
- Presenting a condensed and collective view of selected material
for retrospective inquiry and access. Abstracts can enhance access
to rare or historical data and information.
- Providing native language summaries of text-based foreign language
materials, which helps users to overcome language barriers and to
select appropriate materials which can then be fully translated.
- Providing textual descriptions of oftentimes non-textual information,
such as numerical data.
- Allowing users to access and evaluate a larger volume of source
data and information by providing summarized surrogates to search.
- Allowing users to evaluate individual source materials more easily
by providing summaries of key data and information. Evolving Uses
While abstracts have not changed much from these traditional purposes,
they have evolved into increasingly important tools for retrieving
information in an electronic, rather than a print, environment. Over
the past several years, abstracting and indexing publications and
services have proliferated, both in print and online, which provide
abstracts and bibliographic information (a metadata record) about
data and information resources. CIESIN's Gateway is one example of
an online information service. At the same time, full text indexing
has developed to allow rapid searching of these abstracts, using either
controlled or natural language terminology, or both.
- serving as a source of terminology which can be used to index and
search data and information resources. This is particularly valuable
in new or emerging fields of study; and
- providing the basis of free-text search systems, in lieu of, or
as a preliminary step prior to, searching a complete data or information
However, associated with these newer uses of abstracts, difficulties have
arisen. As more systems support free-text and other searching techniques,
and as distributed systems become interoperable, it becomes increasingly
difficult for metadata creators to anticipate what search mechanisms will
be used. For example, SEDAC's Gateway provides both controlled vocabulary
searching and free-text search capabilities, and provides additional access
to directory entries through a World Wide Web server, which supports full
text searching and browsing capabilities. This example highlights the
challenges in creating metadata which will maximize retrieval using diverse
search systems. Since all or part of directory entry text may be used
in the search process, metadata writers have an increased responsibility
to create quality content, and directory developers must work to insure
that the content and the access mechanisms work together to enhance retrieval.
To help create this metadata text, it seems appropriate to review the
'traditional' guidelines and processes for writing quality abstracts.
The basic principle to keep in mind while writing an abstract, is that
the abstract may be the only text that users search and consult, if they
choose not to retrieve the original data or information. Remembering this
may help the abstractor focus on the key elements and select terminology
to be included. Types of abstracts There are two primary types of abstracts
and selection of the appropriate type is dependent upon the nature of
the material being described, and the intended use of the abstract. The
indicative abstract contains generalized statements to convey (indicate)
to the user what the described item is about. It is very brief, and does
not contain specific findings. Its purpose is to acquaint users with the
subject content of the resource and to help them decide whether or not
to consult the original source. The informative abstract is generally
longer and contains more complete and precise information about the resource,
such as scope, objectives, and purpose of the work, methodologies used,
key quantitative data, and any conclusions and recommendations. Its purpose
is to serve as a true substitute for or surrogate of the resource.
Regardless of the type, an abstract should be:
- complete in and of itself; that is, it should be able to serve as
a stand-alone description which provides a complete picture of the
resource at the selected level of specificity;
- comprehensive in its representation of the key concepts or significant
content that are present in the resource;
- concise, precise, and accurate in its use of terminology;
- written in a clear, terse, non-critical style; and
- logically structured in its presentation of the selected data and
Data owners, principal investigators, and original authors are often
excellent sources of quality abstracts. These individuals have subject
expertise and knowledge of the data and information to be described.
However, they may not have the objectivity of someone who is less familiar
with the data, and may introduce biases or emphases which can hinder
retrieval by users in other subject areas.
Professional abstractors, who are trained and experienced in writing
However, these individuals may not have the required subject expertise
or a working knowledge of the data and information. While a single individual
may not necessarily possess all of the requisite skills and knowledge,
quality abstracts can be and are being generated through cooperative
interaction among subject specialists, abstractors, principal investigators,
data owners, and others. The key is to develop and maintain direct contact
or consultation between these individuals.
Other individuals can be excellent information sources while writing
abstracts. For example, users of data may publish articles about the
data, the methodologies used to collect the data, evaluations of the
data quality, etc.
Users of directory access systems can be excellent information sources
for creating and refining abstracts. Examining system use statistics,
such as search query successes and failures, and tracking user feedback
comments will help to create abstracts which are more effective in terms
The Abstracting Process
There are several steps involved in the process of writing an abstract
of a data or information resource:
- Record citation information as required by local rules.
- Conduct content or subject analysis to determine the key concepts,
characteristics, or significant content within the resource which
need to be represented. This is the most difficult aspect of the process;
the abstractor must represent the multiple viewpoints of diverse users
in order to capture all information that may be pertinent and which
will serve anticipated user needs. Experience with the user audience
and access to feedback studies can help at this stage.
- Using the abstracting guidelines noted earlier, construct a narrative
incorporating all of the elements identified during content analysis.
- Edit the narrative to remove extraneous language which may hinder
effective retrieval. For example, the phrase "This study was conducted
in order to measure the vegetative cover in Northern Michigan....."
may be shortened to "Vegetative cover in Northern Michigan was measured...."
- It is likely that while creating the abstract, the abstractor will
select or suggest appropriate keywords which will be used for indexing
and retrieval. Depending upon the search system to be used, these
keywords may need to be used either in a separate keyword field, in
the abstract, or in both locations.
- Complete the metadata record, (fill in the remainder of the directory
entry fields) according to local rules. This should include documenting
authorship of the abstract, as well as referencing the sources of
information used in creating the abstract and other metadata. It should
be clear to users where obtain additional detail regarding the metadata.
- Examine the abstract from a system-wide viewpoint (that is, from
the directory or the metadata collection level), for consistent use
of terminology within abstracts residing in the same directory or
system. Effective retrieval is enhanced by use of similar terms representing
similar concepts throughout a directory. Guidance for this may be
provided by internal procedure manuals, authority lists, or external
standards that have been adopted by the organization. In addition,
presentation of information in a logically structured and consistent
order within a metadata collection (directory) will allow users to
anticipate where information will occur, and to scan directory records
more quickly for pertinent information.
When translating concepts and characteristics present in the data into
an appropriate textual description, one must make decisions about including
and excluding specific information. While not intended as an exhaustive
list, these elements should be considered for inclusion in an abstract:
- Objectives and scope (purpose) - why the data were collected, why
the study was conducted, or why the author wrote the paper.
- Subject area(s) or parameters- what information or data are present
in the resource; general description of the data.
- Methodology - how the data were collected or the work carried out;
what the data sources were; and how data were handled or analyzed.
- Results -- what relationships or correlations were reported; whether
the data reported were raw, collected data only; or, if experimental
data were included, whether there were single measurements or replications
of data; and factors which may have affected the validity, reliability
or accuracy of the data.
- Conclusions- what evaluations, applications, or implications related
to the objectives were presented.
- Other significant supportive information; which is any other information
deemed important, including tangential information.
Knowing what to exclude from an abstract is equally important as knowing
what to include. Some examples of information to exclude are:
- General introductory matter that occurs in the resource.
- Historical summaries and background information related to the data
which are widely known to the expected user audience.
- Supplementary information which is easily obtained elsewhere or
which is more appropriately described in a separate field.
- Uses of the data. An abstract should discuss what the data are about,
not what the data are used for, unless specific uses are key to understanding
- Extraneous information and terminology which are not supportive
of the significant content of the resource. Such extraneous information
inhibits effective retrieval in a free-text search, and makes it more
difficult for users to quickly scan the abstract for key information.
Abstracts are not dumping grounds for miscellaneous information. There
are various sources which may provide additional guidelines for creating
abstracts, such as the American National Standard for Writing Abstracts,
ANSI Z39. 14-1979.
This is a useful summary of basic abstracting processes and principles
which are applicable regardless of the source or format of the data
Abstracts are stand-alone surrogates or representations of data or information
resources. Users may rely on abstracts as primary sources of information,
or they may use them as a step in the evaluation and access of source
data and information. Applying traditional guidelines and processes for
abstract creation within a directory entry, will improve its retrieval
performance, particularly in free-text search systems. These guidelines
discourage use of an abstract as a 'dumping ground' for information which
does not fit elsewhere in a directory record. Additional enhancements
include use of consistent terminology, and organization and presentation
of information into logical consistent categories. Applied throughout
a directory or across cooperating directories, these enhancements will
help users retrieve metadata.
American National Standards Institute (ANSI). 1979. American National
Standard for Writing Abstracts (Tech. Rep. No. Z39.14-1979). New
York, NY: ANSI.
Bhatia, Martha T. and Cheryl J. Burley. 1992 (unpublished). "Expanding
The Directory Interchange Format To Accommodate Human Science Metadata."
University Center, MI: Consortium for International Earth Science Information
Burley, Cheryl J. 1993. "Creating Metadata Abstracts: Purpose, Guidelines
and Processes." Paper presented at the Catalog Interoperability/NASA
Science Internet Workshop, April 27-29, 1993, San Diego, CA. University
Center, MI: Consortium for International Earth Science Information
Cleveland, Donald B. and Ana D. Cleveland. 1983. Introduction to
Indexing and Abstracting. Littleton, CO: Libraries Unlimited, Inc.
Collison, Robert L. 1971. Abstracts and Abstracting Services. Santa
Barbara, CA: American Bibliographic Center--Clio Press.
National Aeronautics and Space Administration (NASA).1991. Directory Interchange Format Manual, Version 4.0. Greenbelt, MD: NASA/GSFC National Space Science Data Center (NSSDC).
The Joint Steering Committee for Revision of AACR. 1988. Anglo-American
Cataloging Rules- 2nd edition, 1988 revision . Gorman and Winkler,
eds. Ottawa: CLA, London: LA, Chicago, IL: ALA.
Pao, Miranda Lee. 1989. Concepts of Information Retrieval.
Englewood, CO: Libraries Unlimited, Inc.
Thomas, Sarah E. 1990. Bibliographic Control and Agriculture.
Library Trends 38(3)542-561.
Tibbo, Helen R. 1992. Abstracting Across the Disciplines: A Content
Analysis of Abstracts from the Natural Sciences, and the Humanities
with Implications for Abstracting Standards and Online Information Retrieval
. Library and Information Science Research 14(1):31-56.