Syntagmatic relationships and indexing consistency on a larger scale
| Pages | 602-615 |
| Date | 25 July 2008 |
| Published date | 25 July 2008 |
| DOI | https://doi.org/10.1108/00220410810884093 |
| Author | Hope A. Olson,Dietmar Wolfram |
| Subject Matter | Information & knowledge management,Library & information science |
Syntagmatic relationships and
indexing consistency on a larger
scale
Hope A. Olson and Dietmar Wolfram
School of Information Studies, University of Wisconsin-Milwaukee,
Milwaukee, Wisconsin, USA
Abstract
Purpose – The purpose of this article is to examine interindexer consistency on a larger scale than
other studies have done to determine if group consensus is reached by larger numbers of indexers and
what, if any, relationships emerge between assigned terms.
Design/methodology/approach – In total, 64 MLIS students were recruited to assign up to five
terms to a document. The authors applied basic data modeling and the exploratory statistical
techniques of multi-dimensional scaling (MDS) and hierarchical cluster analysis to determine whether
relationships exist in indexing consistency and the coocurrence of assigned terms.
Findings – Consistency in the assignment ofindexing terms to a document follows aninverse shape,
although it is not strictly power law-based unlike many other social phenomena. The exploratory
techniques revealed that groups of terms clustered together. The resulting term cooccurrence
relationships were largely syntagmatic.
Research limitations/implications – The results are based on the indexing of one article by
non-expert indexers and are, thus, not generalizable. Based on the study findings, along with the
growing popularity of folksonomies and the apparent authority of communally developed information
resources, communally developed indexes based on group consensus may have merit.
Originality/value – Consistency in the assignment of indexing terms has been studied primarilyon
a small scale. Few studies have examined indexing on a larger scale with more than a handful of
indexers. Recognition of the differences in indexing assignment has implications for the development
of public information systems, especially those that do not use a controlled vocabulary and those
tagged by end-users. In such cases, multiple access points that accommodate the different ways that
users interpret content are needed so that searchers may be guided to relevant content despite using
different terminology.
Keywords Indexing, Information retrieval,Classification schemes, Group behaviour
Paper type Research paper
Introduction
Consistency in indexing has long been considered essential for effective retrieval. In the
process of indexing, indexers familiarize themselves with the document, choos e what
topics to represent, and choose what to call those topics. The goal is to identify, select,
and name topics consistently so that all of the material about any given topic will be
found together. Decades of research on consistency between different indexers and by
the same indexer at different times has documented medium to high levels of
inconsistency. People simply do not identify and choose the same concepts or the same
words for those concepts and difference in any one or a combination of these will result
in inconsistency. This body of research has identified various factors that contribute to
consistency, but it has not addressed the underlying nature of consistency or of
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/0022-0418.htm
JDOC
64,4
602
Received 12 February 2007
Revised 25 September 2007
Accepted 28 September
2007
Journal of Documentation
Vol. 64 No. 4, 2008
pp. 602-615
qEmerald Group Publishing Limited
0022-0418
DOI 10.1108/00220410810884093
To continue reading
Request your trial