The proposed research examines the impact that adding context – via the use of subject indexes – to a query has on search results. This design science research is motivated by the need for a solution to the well-documented failure of enterprise search. Preliminary experimental data is presented that indicates that the use of subject indexes to augment full-text search may indeed be a valid solution and thereby encourages continued investigation. Continuing the research, we propose an experiment where we simulate the search for randomly selected (single) documents in a collection. We will use the comparison of search results between full-text only and full-text plus subject metadata searches to evaluate search performance. The primary dependent variable used will be the rank of the searched-for document in the result set returned by the search engine for each search.
Schymik, Gregory, "Representational Indeterminancy and Enterprise Search: The Importance of Subject Indexes" (2009). AMCIS 2009 Doctoral Consortium. 25.