Tagging behaviour with support from controlled vocabulary

The paper investigates how knowledge structures from a controlled vocabulary affect tagging. The study is a comparative analysis of tags assigned in two tagging systems a simple tagging system (control system) that provides suggestions from two tag clouds (all users tags and my tags), and an enhanced tagging system (experimental system) that additionally offers suggestions from the Dewey Decimal Classification system (DDC). In the experimental study, 28 political students completed four tagging tasks, each comprising 15 documents. The focus was to examine how suggestions from the enhanced tagging system affect tags as regard tag specificity, exhaustivity, and novelty. Generally, there were no big differences between assigned tags from the two systems. The largest difference was a higher degree of tag specificity in the enhanced system indicating that suggestions from a controlled vocabulary might help taggers in being more specific in their tagging, allowing more precise information searching based on user tags. In addition, the results indicate that structured controlled suggestions might encourage taggers to add synonym variations enhancing the variety and number of access points. Furthermore, controlled vocabularies might be useful for automatic spell checking. Future study should explore in what direction the different kinds of suggestions lead the tagger and whether it is possible to identify scope or patterns between related tags from the two systems. Introduction Social tagging is a way for users to provide user-oriented access to information resources on the Web. In information retrieval, tagging may be complementary to traditional indexing methods. However, many of the existing social tagging applications have not been designed with information retrieval in mind. Users use tags to organize their own private documents, and tags are rather personal than social (Guy & Tonkins, 2006). Folksonomies often lack basic control of spelling variants, synonyms, homonyms, and polysemes (Guy & Tonkins, 2006; Spiteri, 2007). Studies of the