For any assistive technology to truly be helpful and beneficial, the technology itself must be adopted into regular use. However, a third to one half of all assistive technologies are abandoned after purchase. This results not only in a waste of time, money, and resources, but can also lead to disillusionment about the potential helpfulness of assistive technologies.
My dissertation is about understanding and supporting the adoption of assistive technologies for adults with reading disabilities. Using Value Sensitive Design, I am investigating the complex interactions between users, technologies, and sociocultural contexts. Through this approach, I am developing insights into the important values that assistive technologies need to respect in order to both meet the needs of the users and ensure that the technologies are adoptable.
Working from my interests involving disability and education, I conducted a series of interviews with students with disabilities about their experiences in computer science courses. I had two goals with these studies. First, I wanted to identify challenges in doing studies of university students with disabilities and then determine appropriate research methodologies. The second goal was to get an understanding of how the students' disabilities impacted their course experiences and how the courses could be adjusted (if necessary) to be more inclusive.
I have also begun to use this study and its findings as a means for teaching faculty and staff about inclusive education practices.
In collaboration with fellow UW graduate students, Tim Wright and Sarah Read from Practical Pedagogy Research, we have been exploring how to teach students to be better consumers of information from the World Wide Web. Through a multidisciplinary effort, we have developed the Q6C process to support both students and instructors in learning how to effectively evaluate the reliability and usefulness of web-based sources.
This was a continuation of work with Richard Anderson and Ruth Anderson on the Bootstrapping Project. As part of our analysis of the card sorting data from this project, we developed a method for determining the relative distance between card s orts using a defined edit metric. Using this new analysis technique, we identified clusters of closely-related sorts and examined the descriptions the sorters gave to these sorts.
From personal experience as both a student and an instructor (teaching assistant), the benefit of in-class group work for students is mixed. As part of my TA work, I became interested in forming student groups to promote participation from all students. I developed two team formation techniques: one using the Felder-Silverman styles and a variation on the classic jigsaw method—the latent jigsaw.
Example process timeline
At the Center for Engineering Learning and Teaching, I work on projects related to understanding the design processes of engineering students and professionals. In particular, I develop tools to support the analysis of verbal protocol data with a specific interest in visualization of that information. One such example are timeline representations of coded, segmented transcripts (example to the right).
The Assisted Cognition project led by Henry Kautz is an effort to develop computing technologies to increase independence for adults with mild to moderate Alzheimer's disease and dementia by providing support for activites for daily living. For my Master's work / qualifying exam, I worked on a system for recognizing human activities within a sensored environment.
In order to recognize human activities from a large plan library, a probabilistic inference engine requires not only efficiency and scalability, but also flexibility in terms of configurability and end-user programming by non-experts. My solution to these challenges was contrail filtering. Contrail filtering extends the notion of particle filtering by incorporating memory management into the structure of the inference engine and thereby allows the underlying dynamic Bayesian network structure to change as necessary.