HCIR 2012‎ > ‎

Full papers

Full papers were peer-reviewed for novelty, rigor, and quality of presentation according to standards of first-tier conferences. Links refer to full papers published in the ACM Digital Library.

Knowledge Journey: A Web Search Interface for Young Users
Tatiana Gossen, Marcus Nitsche, Andreas Nuernberger
This paper describes a new user interface for a web search engine whose main target group are young users. We explain the main challenges for this user interface and discuss design decisions we made. Our interface is audio supported, contains possibilities for both searching through text input and navigating using menu categories, has a guidance figure for emotional support and a result storage functionality to support cognitive recall. It is also colourful which is appreciated by most children. A comparative user study with 28 young users was conducted where we compared our user interface with a classic text search user interface provided by most current web search engines. We evaluated what features of both interfaces children like or do not like to further improve the interface.
The most effective strategy for finding files is to carefully arrange them into folders. This strategy breaks down for teams, where organizational schemes often differ between team members. It also breaks down when information is copied and reused as it becomes harder to track versions. As storage continues to grow and costs decline, the incentives to carefully archive old versions of files diminish. It is therefore important to explore new and improved search tools. The most common approach is keyword search, though recalling effective keywords can be challenging, especially as repositories grow and information flows across projects. A less common alternative is to use provenance –information about the creation, use and sharing of documents and their context, including collaborators. This paper presents a limited user study showing that provenance data is useful and desirable in search, and that an interface based on a graphical sketchpad is not only feasible, but efficient.

Modeling User Variance in Time-Biased Gain
Mark Smucker, Charles Clarke
Cranfield-style information retrieval evaluation considers variance in user information needs by evaluating retrieval systems over a set of search topics. For each search topic, traditional metrics model all users searching ranked lists in exactly the same manner and thus have zero variance in their per-topic estimate of effectiveness. Metrics that fail to model user variance overestimate the effect size of differences between retrieval systems. The modeling of user variance is critical to understanding the impact of effectiveness differences on the actual user experience. If the variance of a difference is high, the effect on user experience will be low. Time-biased gain is an evaluation metric that models user interaction with ranked lists that are displayed using document surrogates. In this paper, we extend the stochastic simulation of time-biased gain to model the variation between users. We validate this new version of time-biased gain by showing that it produces distributions of gain that agree well with actual distributions produced by real users. With a per-topic variance in its effectiveness measure, time-biased gain allows for the measurement of the effect size of differences, which allows researchers to understand the extent to which predicted performance improvements matter to real users.
The goal of this paper is to provide guidance to researchers investigating exploratory search behaviors and exploratory search systems. It focuses on the design of search tasks assigned in such studies. Based on a review of past studies, a set of task characteristics associated with exploratory search tasks are identified: exploratory search tasks focus on learning and investigative search goals; they are general (rather than specific), open-ended, and often target multiple items/documents; they involve uncertainty and are motivated by ill-defined or ill-structured problems; they are dynamic and evolve over time; they are multi-faceted and may be procedurally complex; and they are often accompanied by other information or cognitive behaviors, such as sensemaking. Recommendations are provided for the design of search task descriptions that will elicit exploratory search behaviors.