Grant: Discovery Grants Program - Individual

The success of commercial web search engines has led to a common expectation that searching is a solved problem. Such search engines work extremely well for look-up tasks where the searcher can clearly explain what they are seeking in just a few words, and the search engine can provide a ranked list of high-quality documents that contain the required information. A simple search interface following the "query box and 10 blue links" design pattern is all that is required to support such search activities. However, when the search task is complex, a simple search interface that is tuned for look-up tasks is no longer sufficient to support the searcher. One specific setting in which complex search tasks are commonplace is searching within academic digital libraries. Academic searchers (students, researchers, faculty) are often faced with search tasks that are ambiguous, multi-faceted, and open-ended, and which require the synthesis of information from multiple sources. Rather than trying to look-up specific pieces of information, a more appropriate approach is to conduct an exploratory search. This consists of a two-stage process, in which the searcher first seeks to discover, learn, and investigate aspects of their search task. Then, as they become more certain about what it is they are seeking, they focus their search efforts on formulating queries, examining the results, and extracting information. Such exploratory search activities are time consuming and often span multiple search sessions. The problem academic searchers face is that the search interfaces used by academic digital libraries only provide limited support for cross-session exploratory search processes, as they are strongly influenced by the simple search interfaces used in web search settings. The focus of this research program is to conduct a systematic study of academic digital library search interface features that support cross-session exploratory search activities. Novel search interfaces will be devised and implemented that enable specific aspects of cross-session exploratory search, leveraging available search metadata and using visualization methods to convey patterns and relationships. New approaches for presenting search results to a searcher that encourage discovery, learning, and investigation will be developed. New ways of assisting a searcher to reacquaint themselves with their previous search activities will be devised. Promising approaches will be studied in both controlled and natural search settings to both determine their utility in supporting the search activities and to assess how the behaviour of the searcher changes when provided with appropriate support for their cross-session exploratory search activities. Outcomes from this research will influence the design of next-generation academic digital library search interfaces, enabling future academic searchers to more readily and confidently find academic literature to support their research endeavours.