Mentors and Project Descriptions
winter 2026 Projects
Ethan Ancell: Pitch your own project!
Do you have a topic in statistics that you’ve been wanting to explore, but there isn’t a specific class offered for it? In this DRP, you’ll get the chance to pick a topic of your choosing, and I am happy to give direction and supervision for your self study.
If you are interested in this opportunity, in addition to filling out the usual DRP application, please also email me at ancell@uw.edu from your UW email with the following information:
- Proposed topic title
- Why you are interested in this particular topic
- A rough sketch of what resources or papers you plan to read to conduct your self-study
- Whether you would like this project to be more of a “reading” (self-study) project or more of a “research” project.
Preference will be given to applicants who have a specific and compelling plan for their self-study. Due to the fact that the other Winter 2026 DRP project offers are heavily skewed towards junior and senior undergraduates, I will be giving preference to applicants who are freshman or sophomores, and are interested in applying to the statistics major.
Ethan Ancell: Asymptotics without the sample size going to infinity
Asymptotic results (including the Central Limit Theorem) are commonly presented as requiring the “sample size going to infinity.” In this DRP, we will focus on this paper from Charles Geyer which explores the idea of asymptotic results without any requirement of the sample size (or anything) “going to infinity.”
This DRP topic is inherently a more difficult and technical topic than most DRPs I run. To be upfront, I have zero perfomance expectations from anyone who works with me on this project, and I only care that you’re enthusiastic and ready to ask lots of questions. This project is best for someone who has a little bit more time in the winter quarter and would enjoy self-studying a very mathy topic in statistics in a low-pressure environment with a senior PhD student supervising your study.
Jasper Yang: Statistical Decision Theory
This project will explore statistical decision theory as a framework for making principled choices under uncertainty. Using the textbook “Decision Theory: Principles and Approaches” as a rough guide, we will build our way from the foundations of utility theory through the seminal works of Abraham Wald before taking a deeper dive into a modern topic tailored to the mentee’s specific interests. Along the way, we will compare frequentist and Bayesian perspectives.
Kayla Irish: Stratified Randomization in Clinical Trials
Stratified randomization is widely used in clinical trials, yet its theoretical motivations and statistical consequences are often not fully explained. This project takes a mathematical perspective on stratified randomization, asking why it is used, what it accomplishes, and how it affects balance, variance, and inference. We will study common forms of stratified and covariate-adaptive randomization and discuss whether stratification remains beneficial in modern trials given increasingly sophisticated covariate adjustment methods. The project will focus on reading and understanding methodological papers, with optional simulation work for interested students.
Qian Meng: Introduction to Causal Inference
Causal inference provides a framework for estimating the effects of interventions or treatments on outcomes. This project introduces basic concepts and methods in causal inference. We will start by distinguishing association and causation, and then focus on key ideas in randomized experiments (e.g. Fisher’s test and Neyman inference) and observational studies without unmeasured confounding (e.g. ignorability and propensity score). If possible, we will also explore some useful matching algorithms such as optimal pair matching and full matching. This project will be primarily based on the first three chapters of the textbook “A First Course in Causal Inference” by Peng Ding.
Tony Lei: Statistical Foundations of Machine Translation and Large Language Models
Machine translation can be understood as a problem of probabilistic modeling. This project explores the foundational statistical ideas behind machine translation and large language models, including conditional probability modeling of text, maximum likelihood estimation, and Bayesian perspectives on sequence generation. We will study classical statistical machine translation models alongside modern LLM-based approaches to understand how statistical methodologies influence translation systems. The focus will be on developing a deeper theoretical understanding of these methods rather than on model implementation.