Using metadata record graphs to understand controlled vocabulary and keyword usage for subject representation in the UNT theses and dissertations collection

An important function of metadata for electronic theses and dissertations (ETDs) is supporting the discovery of related documents through linking of data values in the fields of metadata records. While benefits of the ETD format allow for full-text searching, metadata is still an important and necessary component of the global ETD infrastructure because it is often not possible to share the full documents in aggregations such as the Global ETD Search for the Networked Digital Library of Theses and Dissertations. The metadata field that has the most potential to assist users in discovery is the subject field used to represent what a resource is about. Over the years there has been much discussion of the value of author-generated keywords versus adding subject terms from controlled vocabularies by information professionals as documents are submitted to the University repository. This research seeks to explore this problem with the help of network analysis method not used for such analyzes before by building and analyzing metadata record graphs for the University of North Texas theses and dissertations. This paper reports on the characteristics of keyword-based and controlled-vocabulary-based metadata record networks and discussions insights that can be gained from this approach to metadata quality analysis. This research seeks to explore this problem with the help of network analysis method not used for such analyzes before by building and analyzing metadata record graphs for the University of North Texas theses and dissertations. This paper reports on the characteristics of keyword-based and controlled-vocabulary-based metadata record networks and discussions insights that can be gained from this approach to metadata quality analysis. This research seeks to explore this problem with the help of network analysis method not used for such analyzes before by building and analyzing metadata record graphs for the University of North Texas theses and dissertations. This paper reports on the characteristics of keyword-based and controlled-vocabulary-based metadata record networks and discussions insights that can be gained from this approach to metadata quality analysis.