Inter‐indexer consistency in graphic materials indexing at the National Library of Wales

Abstract

Purpose -This paper seeks to report a project to investigate the degree of inter-indexer consistency in the assignment of controlled vocabulary topical subject index terms to identical graphical images by different indexers at the National Library of Wales (NLW). Design/methodology/approach -An experimental quantitative methodology was devised to investigate inter-indexer consistency. Additionally, the project investigated the relationship, if any, between indexing exhaustivity and consistency, and the relationship, if any, between indexing consistency/exhaustivity and broad category of graphic format. Findings -Inter-indexer consistency in the assignment of topical subject index terms to graphic materials at the NLW was found to be generally low and highly variable. Inter-indexer consistency fell within the range 10.8 per cent to 48.0 per cent. Indexing exhaustivity varied substantially from indexer to indexer, with a mean assignment of 3.8 terms by each indexer to each image, falling within the range 2.5 to 4.7 terms. The broad category of graphic format, whether photographic or non-photographic, was found to have little influence on either inter-indexer consistency or indexing exhaustivity. Indexing exhaustivity and inter-indexer consistency exhibited a tendency toward a direct, positive relationship. The findings are necessarily limited as this is a small-scale study within a single institution. Originality/value -Previous consistency studies have almost exclusively investigated the indexing of print materials, with very little research published for non-print media. With the literature also rich in discussion of the added complexities of subjectively representing the intellectual content of visual media, this study attempts to enrich existing knowledge on indexing consistency for graphic materials and to address a noticeable gap in information theory. IntroductionInter-indexer consistency is a quantitative measure of the degree of agreement between different indexers indexing the same document within an information retrieval system. Although the literature shows only a tentative correlation between indexing consistency and retrieval effectiveness, many see consistency as a benchmark of quality both in service provision to information seekers and potent resource discovery. Previous consistency studies are abundant yet have almost exclusively investigated the indexing of print materials, with very little research published for non-print media. With the literature also rich in discussion of the added complexities of representing the intellectual content of visual media, this study focuses on indexing consistency for graphic materials in an attempt to address a noticeable gap in information theory.

Reengineering of MeSH thesauri for term selection to optimize literature retrieval and knowledge reconstruction in support of stem cell research

“…MeSH is an authoritative controlled vocabulary that is highly effective for bibliographic retrieval for systems that use it. However, studies and literature reviews on inter-indexer consistency often find variability and inconsistency, with consistency rates in the 20–40% range [ 25 – 27 ]. Our method takes this into consideration and, in part, addresses this problem to enhance term specificity by looking at how these terms posit in the subcategories within the thesaurus.…”

Social tagging and commenting: theoretical perspectives

“…20 The problem with the proposition is that the evidence from formalised subject description and inter-indexer studies, of course only partly relevant for tagging and commenting, suggest the opposite. 21 The paradox of both tags and comments are that their rhizomatic nature means that at an aggregate level, when tags and comments are not used by a single individual or a single static close-knit community, they risk being simultaneously too generic and too specific over time.…”