Understanding help seeking within the context of searching digital libraries

Abstract

To date, there has been little empirical research investigating the specific types of help-seeking situations that arise when people interact with information in new searching environments such as digital libraries. This article reports the results of a project focusing on the identification of different types of help-seeking situations, along with types of factors that precipitate them among searchers of two different digital libraries. Participants (N = 120) representing the general public in Milwaukee and New York City were selected for this study. Based on the analysis of multiple sources of data, the authors identify 15 types of help-seeking situations among this sample of novice digital library users. These situations are related to the searching activities involved in getting started, identifying relevant digital collections, browsing for information, constructing search statements, refining searches, monitoring searches, and evaluating results. Multiple factors that determine the occurrences of each type of help-seeking situation also are identified. The article concludes with a model that represents user, system, task, and interaction outcome as codeterminates in the formation of help-seeking situations, and presents the theoretical and practical implications of the study results.

Predicting searcher frustration

“…Xie and Cool [17] explored helpseeking or problematic situations that arise in searching digital libraries. They identified fifteen types of help-seeking situations that their 120 novice participants encountered.…”

“…However, multi-class prediction may also be useful, using either regression or a multi-class machine learning method. We also focus on general frustration, but predicting types of frustration may also be useful, e.g., predicting the fifteen types of frustration outlined by Xie and Cool [17].…”

“…Depending on the level of frustration, we may wish to change the underlying retrieval algorithm or the user interface. For example, one source of difficulty in retrieval is a user's inability to sift through the results presented for a query [13,17]. One way that a system could adapt to address this kind of frustration is to show the user a conceptual breakdown of the results: rather than listing all results, group them based on the key concepts that best represent them [13].…”

Using Transaction Logs to Better Understand User Search Session Patterns in an Image-based Digital Library

“…To date, user interaction studies involving DLs have been carried out using direct search observation (Xie and Cool 2009;Joo and Xie 2012) and resource usage logs (Wolfram and Xie 2002). click-through data for results assessment (Smith, Brien, and Ashman 2012).…”

To hint or not

“…Bilal and Kirby [3] reported that about half of the participants of their user study felt frustration when searching. Xie and Cool [10] demonstrated that most of the time users have problems with formulating and refining search queries. Besides good retrieval performance, a successful search requires users to possess certain skills.…”