Undergraduate Research Highlight: Helping Computer Science Research by Improving Online Surveys

By Alejandro Velasco Dimate (CRA-E Fellow, College of William & Mary) and Emma McDonald (CRA-E Fellow, University of Alberta)

This Q&A highlight features Ye Shu, an Honorable Mention for the 2024 CRA Outstanding Undergraduate Researchers Award. Ye finished his undergrad from Williams College with a double major in Computer Science and Philosophy and a concentration in Cognitive Science. They are now pursuing a PhD in Computer Science at UC San Diego. 

How did you get involved in computing research?

During the summer before my sophomore year, I heard that Prof. Daniel Barowy had developed an autograder system, which reminded me of a project I was working on to automate testing and deployment processes when someone pushed code changes to a git repo. I approached Dan and asked what he thought about these automations. It turned out that we saw eye to eye and he took me on as a research assistant to implement a git action-triggered autograder. By next summer, I was convinced that I liked doing research and wanted to try doing it full-time, so I applied to Williams College’s summer research program and was unsurprisingly paired with Dan.

Can you tell us about your research?

One of the projects I worked on addresses the difficulties that can arise when implementing an online survey as part of a research project. We designed a programming language that allows scientists to describe their surveys using a template with potential synonymous word choices, similar to the game Mad-Libs. Then, during runtime, the system distributes the survey, each time selecting a different way of wording. The idea is similar to a randomized controlled trial, where each correspondent sees the same survey but with slightly altered wording, thus mitigating the potential bias introduced by word choices in the aggregated result. 

I also discovered that authentic online survey results will stand out from the random background noise of inattentive respondents who use tactics like selecting random answers. I built our algorithm to filter out these inattentive responses. This work will help scientists with little CS or statistical background to design and automatically conduct robust online surveys. I presented the preliminary results at PLATEAU 2024, a workshop on the intersection of HCI and PL. We are preparing the full paper for submission to a conference.

 What challenges did you encounter throughout the research process?

One unique thing about attending a small liberal arts college is that there are no graduate students, so undergraduate students like me get to work with faculty members directly and take charge of the research projects. While this gives us more space to explore and learn as independent researchers, it also presents challenges early on as there are no senior graduate students to ask for help. Fortunately, my faculty advisor, Dan, helped me close my knowledge gaps with lots of hands-on mentoring. As I slowly gained experience from trials and errors, Dan also gave me more independence to explore and steer the project into directions that we both were interested in.

How has your interest in philosophy influenced your research path?

During my philosophy training, we were rigorously trained to read broadly, understand other people’s ideas, and write extensively about our own ideas. These skills have proven to be extremely useful for my research, as these are also the three things I do: read relevant literature to know what others are doing, identify gaps in current works, write about our ideas, and communicate them with peers. In this regard, my philosophical training prepared me well for a career in research. It also allows me to look at many computing problems from an entirely different perspective. When I was applying for IRB approval for the survey project, my formal training in applied ethics allowed me to better reason about potential consequences of my research and design informed disclosure and anonymized data processing procedures. Ethical considerations become ever more relevant to me, because my PhD research is in the fields of security and measurement, where we have to interact with real-world systems and impact real people behind them.

What do you think are the most important qualities for a successful researcher?

Resiliency and ability to handle setbacks. Research is not straightforward and often lacks clear indication of progress (although the philosopher inside me is questioning what the word “progress” really means). Instead, the research experience is full of visible failures. Maybe an idea does not work out or your paper is rejected. As a researcher, it is important to learn to handle these setbacks and learn from them. The path of research is going to be nonlinear.

How do you stay motivated and inspired in your research?

The interesting questions themselves keep me motivated, but when I run out of steam I go hiking with friends. The beautiful natural views of Western Massachusetts work magic to generate new thoughts and ideas.

Do you have any advice for other students looking to get into research?

If you are unsure whether you want to pursue a career in research, start by getting your feet wet! Start with a small project, explore different areas, and work with different people. Most importantly, try to have fun!