Theme HCI Ways of Knowing, Part II: "How We Know What We Know"
Location Snow Mountain Ranch, Fraser, Colorado
Date June 14 - 18, 2011
Boaster Presentations
Can Crowdsourcing Improve HCI?
Serge Egelman (NIST)
+ Abstract - Abstract
Much of HCI research focuses on improving the user experience by using data from human subjects experiments. Designing a laboratory study, observing participants, and compensating them is a very expensive process, in terms of both time and money. Due to these costs, sample sizes tend to be relatively small, which in turn has an effect on the related confidence intervals. However, new crowdsourcing technologies, such as Amazon's Mechanical Turk, allow researchers to conduct human subjects experiments in much less time, on much larger sample sizes, and for less money. In this paper I describe several previous studies that I have performed using crowdsourcing'some prior to joining NIST'and explain how they would have been time and cost prohibitive without crowdsourcing technologies.
Storytelling as a Conceptual Design Method
Kim Gausepohl (Virginia Tech)
+ Abstract - Abstract
NIOSH encourages the exploration of methods to foster a dialogue between engineers and healthcare stakeholders. I previously explored the use of storytelling for medical device requirements and found that storytelling elicits a robust requirements set. I intend to further explore the impact of storytelling during the conceptual design stage.
Understanding task intent for web search
Robin Jeffries (Google)
+ Abstract - Abstract
Query-level search intent categorizations alone cannot capture the complexity that spans search tasks. We come to this conclusion after an extensive analysis of real users labeling their own web search tasks. Analyzing this data, we derive a search-intent classification that is task-based (rather than single-query-based) and far more descriptive than other classifications. To derive the taxonomy, we conducted a diary study with 36 participants for two weeks. Each participant self-segmented their web activity stream into tasks and labeled each task with their intent. result was a twelve category categorization of task intent organized along two main axes: (a) specificity: degree of specificity of the user's desired end goals, and (b) decision- related: whether the user was in the process of making a decision based on information found during the task.
Designing Effective End-User Interaction with Machine Learning
Saleema Amershi (University of Washington)
+ Abstract - Abstract
Recent work has demonstrated the potential of end-user interactive machine learning in specific applications. However, we still lack a generalized understanding of how to design effective human interaction with these systems. We show that designs balancing the needs of humans and machines significantly impact the effectiveness of interactive machine learning.
Linguistic Analysis and Ontological Representation of Twitter Communications During Crisis: Contributions to Situational Awareness
Sean Munson (University of Michigan)
+ Abstract - Abstract
The popular microblogging service Twitter hosts 160 million users who send almost one million 140-character messages (known as tweets) per day. In times of mass emergency, many use Twitter to gather and disperse information. With so many people using Twitter, and so many tweets being sent at any given time, locating and organizing timely, useful information during these safety- critical situations is a task best suited for automatic methods. However, training machines to correctly identify and subsequently extract tweets that contribute to situational awareness'an overall picture of what is going on'is a task that involves understanding what types of information people tweet during mass emergencies, how information may differ depending on the type of mass emergency a population is experiencing, how various types of information are linguistically constructed, and how we can represent this knowledge in a way that is computationally tractable. This research confronts the problem of automatically identifying and categorizing tweets that contribute to situational awareness during mass emergency. This process involves a macro-level, behavioral analysis of Twitter communications, as well as a micro-level analysis that identifies how different types of information are linguistically constructed.
Lost in Translation: Bridging the User-evaluator Gap in Evaluation Instrument Design
Diana Kusunoki (Drexel University)
+ Abstract - Abstract
Eliciting users' perceptions and understandings of digital libraries involves guiding users and evaluators lost in translation toward the same wavelength before forging ahead. This paper outlines the initial framework and methods for the participatory design of digital library evaluation instruments that bridge meaning between users and evaluators.
Leveraging Crowds and Clouds: An Intercommunity Approach to Improving Web Accessibility
Jeffery Hoehl (University of Colorado)
+ Abstract - Abstract
This paper analyses the trends in web accessibility research by discussing historical approaches including technical, information consumption, information contribution, and community based. An analysis yields the potential to explore intercommunity-based approaches and proposes methods to combine cloud-computing and crowd-sourced tools to improve web accessibility for those with cognitive disabilities.
Designing Effective Virtual Organizations with Patterns
Jing Wang (The Pennsylvania State University)
+ Abstract - Abstract
Virtual organizations are increasingly pervasive in knowledge work, despite significant challenges with respect to support for awareness, sense of community and problem solving. We are investigating the hypothesis that such organizations need to and can better understand and manage specific sociotechnical affordances of information infrastructures and tools. We analyze two effective organizations through a patterns-based approach, characterizing their practices and design schematically. Such an approach contributes practical guidance to the design of effective virtual organizations.
Towards Usability Maintenance
Sean Munson (University of Michigan)
+ Abstract - Abstract
Current usability practices focus on improving the design of software upfront through prototyping, user testing, and other activities. However, a number of problems emerge after software is deployed in the user's environment. In fact, software maintenance and software support activities incur high costs in time and resources in the post-deployment phase. At the same time, it appears that the role of usability diminishes after deployment. We propose that within the field of usability, there needs to be an orientation towards usability maintenance. Unlike the concept of software maintenance that largely focuses on the correctness and performance of the software artifact, the goal of usability maintenance is to support and improve user experience after deployment.
Practical Statistics for Human-Computer Interaction: An Independent Study Combining Statistics Theory and Tool Know-How
Jacob Wobbrock (University of Washington)
+ Abstract - Abstract
As both a student and professor, I have observed that new doctoral students in Human-Computer Interaction (HCI) often lack the statistical fundamentals necessary for conducting experimental research in HCI. Courses offered in statistics departments can be deeply theoretical but difficult for students to apply, while help manuals for statistics software packages describe the application of analyses without providing the rationale behind them. Between these extremes lies a 'practical middle,' wherein students gain enough theory to understand statistical analyses, but also learn how to handle data using current software tools. This 'practical middle' for statistics is not often taught in HCI curricula. To address this issue, I have developed a set of independent study modules designed for new doctoral students in HCI to rapidly gain the statistics know-how necessary for their research. The modules relate specifically to the kinds of data common in HCI experiments, often involving myriad trials, many crossed factors, repeated measures, and the need for nonparametric analyses. To date, 11 students have taken this independent study, and report feeling proficient in understanding and producing statistical results after completing the study.
Uncovering HIV Prevention Opportunities: Towards an Exploration of the use of CMC Applications by Young Men that have Sex with Men (YMSM) in Seeking Sexual Partners
Woodrow Winchester (Virginia Tech)
+ Abstract - Abstract
As the public health community strives to impact HIV infection rates through helping YMSM make better decisions around their sexual health, the need to explore how CMC applications, meteoric in use, are shaping safer sex negotiations is paramount. Activity theory is proposed as a framing to support such an exploration.
A Work-Centered Visual Analytics Model to Support Engineering Design with Interactive Visualization and Data-mining
Xin Yan (The Pennsylvania State University)
+ Abstract - Abstract
Traditional visual analytic tools have some limitations to support the analysis of large-scale, multi-dimensional, continuous data sets, mainly because they lack the capability to identify hidden patterns in data that are critical for in-depth analysis. Specifically, engineering designers for complex systems need powerful tools to handle overwhelming information such as numerous design alternatives generated from automatic simulating software. During the exploration within a 'trade space' consisting of possible designs and potential solutions, designers want to analyze the data, discover hidden patterns, and identify preferable solutions. In this paper, we present a work-centered approach to support visual analytics of multi-dimensional data by combining user-centered interactive visualization and data-oriented computational algorithms. We describe a system, Learning-based Interactive Visualization for Engineering design (LIVE), which allows engineering designers to interactively examine large design data sets through visualization and automatic data analysis. We expect that our approach can help designers make sense of complex design data more efficiently and effectively. We report our preliminary evaluation on our system by analyzing a real design problem related to aircraft wing sizing.
Designing Social Influence in Persuasive Technology to Motivate Behavior Change
Sean Munson (University of Michigan)
+ Abstract - Abstract
A brief overview of my research on designing social software to support behavior change, for health and wellness domain and for encouraging people to increase the diversity of political news they read.
Spatial Interaction: Exploring Interaction within the Spatial Metaphor
Alex Endert (Virginia Tech)
+ Abstract - Abstract
Visual analytics emphasizes sensemaking through interactively exploring visualizations. In this paper, we discuss how instead of adjusting model parameters to change the view, spatial interaction allows users to explore information in the visual metaphor, while the system adjusts the parameters based on the user's analytical reasoning.
Studying Long-Term Effects with Longitudinal Field Studies in HCI
Matthew L Lee (Carnegie Mellon)
+ Abstract - Abstract
Computing has the potential to change people?s lives in a very real and substantive ways such as improving their health, bringing people closer together, and supporting their ability to learn. However, these changes often occur over a long period of time, usually after individuals are able to integrate a new technology into their lives. HCI studies often study individuals using novel technologies for relatively short time periods such as a few weeks or a month. Longitudinal field studies lasting several months or years, while difficult and potentially expensive, provide the natural context for studying the longer-term effects of technology use. In this paper, I briefly discuss 1) the value of longitudinal field studies, 2) the challenges associated with using this method, and 3) some lessons learned from two long-term field studies in which I investigated the effects of lifelogging technology on human memory and self-awareness of changes in abilities as people age. Drawing on these lessons can help future researchers get the most out their field studies while helping to minimize the risks.