I am deeply interested in the dynamic between people and technologies: how our behaviors inform the design of technologies and how technologies change our everyday behaviors. The Human Computer Interaction (HCI) class that I took during my undergraduate study initiated this passion. I learned more about the field when conducting the research required by the BSc(Hons) degree, where I picked an HCI topic. However, the real breakthrough came when I volunteered with the local Special Olympic community. I witnessed the challenges that people with special needs and their caretakers face and the joy that they express, often upon some minor things. I started to think more profoundly about the relationship between technologies and their users. This growing curiosity led me to pursue a Ph.D. through the Department of Informatics, University of California Irvine in 2011.
Since then, I have contributed to more than ten HCI projects, including five projects that I led from start to finish. My reading and hands-on experiences allowed me to gain an in-depth understanding of various research methods including their strengths and shortcomings. Along the way, I have also developed my own identity as a researcher. Information technologies have social impact, especially for people who are disadvantaged by their abilities or social context. I wish to use my skills to develop information technologies that can improve the quality of life for these individuals, and maybe the seemingly small improvements can ripple and unlock the full potential of their lives.
In the past a few years, my main focus was on my dissertation research, which comprised three projects/phases. In this research, I looked at an often neglected problem that blind users encounter: the communication, awkward at times, with sighted users in regard to interacting web pages. I used both qualitative methods and quantitative methods to investigate the root of the issues and assess a potential design solution. Through this research, I have gained interesting insights and produced various design suggestions at each phase. I successfully defended my dissertation on May 16th, 2018.
Currently, I am in the job market. So please get in touch if you have good suggestions. I do not know what life will take me next. But I do know that I want to dedicate my career to research and work on good topics that can impact people's lives. I can’t imagine anything else more satisfying than that!
My experience provided me an in-depth understanding of a wide range of qualitative and quantitative research methods, including the strengths and shortcomings of each method.
I am also familiar with various methods used for usability testing and UI/UX designs. I have applied them in my own research and guided senior students in project or design classes to use them. An incomplete list includes the following:
One important aspect of doing research is applying rigorous data analysis methods to distinguish signals from noises. I have experiences processing both qualitative data and quantitative data. Below are a few methods that I have used:
I have a background as a software engineer before becoming a researcher. In graduate school, I have developed web-based and Android-based prototypes for my own research and projects that I contributed to. Some highlights of my technical skills include:
Performed generative and evaluative user studies to inform project teams with insights about user behaviors.
Presented workshops and seminars on various user research methods.
Conducted independent research in the area of Explainable AI, Human-Agent Interaction.
Mentored student and intern projects on topics such as Voice User Interface, iOS application, and visualization.
Led 5 projects from start to finish; responsible for all activities in a research project cycle.
Collaborated with teams of 3-5 researchers from across multidisciplinary departments in 4 projects: planned research agendas, developed research prototypes, such as:
Served as Teaching Assistant for 18 classes
No man is an island. Communication with others is essential when performing many tasks. However, prior research has reported that blind web users experienced difficulties when discussing web pages with sighted web users. One plausible explanation is that blind users use screen readers to access web pages, which convey content via synthesized speech. This is dramatically different from how sighted users perceive web pages. This leads to communication barriers when describing web pages. In this research, my goal was to clarify what the specific issues were. Such an understanding is necessary as it informs what a design solution should address.
I conducted a content analysis study. The idea was to collect written instructions for simple web-based tasks and analyze what wording and patterns were problematic for screen reader users. I distributed online surveys on Amazon Mechanical Turk and collected 50+ unique responses.
I found two consistent, problematic patterns among other minor patterns from the data. First, spatial terms, such as “top left,” “center,” were prevalent when sighted users described web elements. Since screen readers do not convey layout, blind users would not be able to make sense of these terms. Second, sighted users tend to use inconsistent terms to describe web elements. For example, they might describe the same top menu using “links,” “tabs,” or “text.” Since screen readers always use the technical web element names, blind users could get confused when receiving instructions featuring such terms.
Web Audio API is a W3C work-in-progress specification that is supported by all major browsers. Using Web Audio API, a web developer can easily edit one or more audio sources during runtime and create sophisticated audio effect. I was particularly interested in its spatialization features as spatial audio has the potential to convey layout to web users with visual impairments. However, there is no known evaluation of the effectiveness of Web Audio API’s spatialization features. In addition, moving audio can provide more design options and a more animated experience. But compared to other design spaces, we know little about how to design good moving audio, e.g., do different movements lead to different recognition rates? Do spatial properties, such as distance, speed, play a role in recognition accuracy?
To answer both questions, I designed and implemented two lab experiments: one focused on stationary spatial audio and the other focused on moving audio. I recruited 18 participants. Each session took about two hours and the participant completed recognition tasks of audio stimulus with various spatial properties. At the end, I collected more than 4000 data points. Using ANOVA and non-parametric methods, I compared and identified regions that contribute to better or worse recognition rates. I also built multi-level general models to evaluate the role that different spatial properties play.
Based on the analysis, I produced a few main spatial audio design guidelines, as well as a summary of observed patterns. The main findings include the recognition rates of stationary audio spatialized in the horizontal space for different resolutions, the recognition rates of various movement types and the impact of a moving audio’s length or direction, the ranges of less sensitive area in the audio space, and a recognition bias that users exhibited.
You can find a poster summarizing the findings here.
My early research has found that sighted people tend to use spatial terms when describing web pages, but spatial terms are problematic for blind screen reader users to comprehend. One potential solution is to convey layout information via spatial audio feedback. Currently, screen readers have not utilized spatial audio much and user attitudes about such designs are unknown. Therefore, it is necessary to conduct formative research before any major development.
I developed a proof-of-concept screen reader prototype that supports basic screen reader features and additional spatial audio features. Then I conducted user studies with 20 blind or legally blind participants. Each one-on-one session took about two hours. The user study includes usability testing, surveys, and interviews. The focus was on whether or not users could interpret relative spatial terms or perceive the overall web page layout using the prototype, and their direct feedback on the design concept.
Quantitative data collected from the usability testing and surveys shows that spatial audio features enabled participants to accomplish tasks that are hard or impossible to complete with regular screen readers. However, I also found that a user’s familiarity with screen readers had a strong impact to their experience with spatial audio features. For users who are new to screen readers, they had a hard time processing the additional spatial audio feedback. For experienced screen reader users, they responded positively to the spatial audio features in general, but a few were reluctant to incorporate new techniques into their existing interaction routines. From the interview, I also learned usability issues with my current design and various scenarios where spatial audio feedback could make positive impacts. These findings allowed me to compile a list of recommendations on how future developments should be focused.
You can find a case study of this project here.