Button to scroll to the top of the page.

Spotlights

 

Student Spotlight: Sukyung Park

sukyung16 December 2014—Meet Sukyung Park, a first-year PhD in Statistics student in the Department of Statistics & Data Sciences. 

Tell us a little bit about yourself—educational background, previous work experience, etc.

I’m from South Korea. I received an MS in statistics from UT Austin and a BS in Industrial Engineering from KAIST, Korea Advanced Institute of Science and Technology. Before starting my master’s program in statistics here, I studied behavioral and quantitative marketing methods in the KAIST Graduate School of Management. I left the KAIST Graduate School of Management in 2011 because I found my vocation in statistics, especially for experimental design and causal inference.

What attracted you to the PhD in Statistics program at UT?

I am interested in research on experimental design and causal inference using Bayesian approaches. At UT, there are many prominent professors who have expertise in Bayesian analyses and there are many classes that focus on various Bayesian methods. The curriculum of the doctoral program in statistics is flexible so that a student can build knowledge on statistical theories as well as substantive areas of application. Since I had already taken several core courses required for the doctoral program when I studied for my master’s degree, the fact that I could start research soon was another appealing factor of this university.

Tell us about a project or piece of research you have worked on while attending UT. 

For my master’s report, I compared various adaptive designs for clinical trials in terms of statistical power, assuming the situation where multiple experimental treatments are tested in multiple stages. Adaptive designs allow for flexible design adaptation in the middle of a trial by conducting multiple statistical tests on accumulated data. However, multiple testing inflates overall type I error rate and maintaining this error rate at a certain level often results in a reduction in statistical power. Including both frequentist and Bayesian designs, I conducted a simulation study that revealed that Bayesian approaches have superior performance among the designs included in this study.

Another area of research that I am working on is statistical causal inference on mediation. In statistics, causal effects are functions of counterfactual quantities that would be observed if intervention were set to some specific values. Since an observational unit can be exposed to only one level of intervention, there are missing data problems that require several assumptions to identify casual relations. 

How would you describe your area of study/specific research to your grandmother?

I am studying experimental designs for clinical trials and causal inference based on statistical methods. Research on these topics will help people have the opportunity to receive advanced therapy sooner and help people understand how to prevent fatal diseases by changing their lifestyle.

Fun Facts

- A talent you have always wanted: A talent in singing.

- Favorite book: Le Petit Prince (The Little Prince)

- Favorite quote: “It shall also come to pass.”

- Role model: My parents

- Favorite vacation destination: Banff and Jasper National Parks, Canada

- If you weren’t in grad school, what would you do? Study at medical school or work in industry

Michael Mahometa Launches MOOC on Foundations of D...
Statistics Online Support (SOS) Resource Platform ...

Related Posts