This is a portfolio of research questions I’ve answered and the ones I’m still working on. The work clusters into three interconnected strands: education policy and data science, writing assessment and AI, and cultural analytics. Methods are the membranes between strands — causal inference, NLP, and data science show up everywhere, so questions in one strand often connect to questions in another.
Education Policy & Data Science
As Director of Data Science at Accelerate and a Harvard Strategic Data Project Fellow, I work on understanding how pandemic-era education funding was spent, whether high-dosage tutoring programs are effective, and how states can build data infrastructure to answer these questions going forward.
Published
- How can states continue pandemic-era direct student services post-ESSER?
- If the money moves do the metrics move too?
- What have the states leading recovery learned after two years of tutoring implementation?
- How much has the US spent on high-dosage tutoring?
- How can we engineer data-informed tutoring at scale?
- Where does Accelerate have grantees?
- Where have ESSER funds been allocated to tutoring?
- Where have ESSER funds been allocated?
- How were ESSER funds spent?
- ESSER Data Snapshots
Open
- What would an open standard for tutoring data look like?
- What’s so high about tutoring anyway — dosage, quality, impact?
Writing, Assessment & AI
My PhD at University of Michigan centered on placement equity: whether standardized tests reliably sort students into writing courses, and what happens when students get a say. That thread extends into alumni outcomes, GenAI policy, and — most recently — NLP research on how AI development itself is characterized.
Published
- Are essay-based tests a reliable metric of student performance on college-level writing tasks?
- Do students select first-year class placements differently by race, sex, SES?
- How do academic advisors influence student writing course selection?
- What do student reflections reveal about SSP ecological impacts?
- How does social justice inform writing placement administration?
- How does course recommendation impact student outcomes?
- How do alumni describe the effects of college?
- How do alumni describe how their college writing experiences affected their careers?
- Do students learn during one-shot library sessions?
- Are video metrics useful for course assessment in online writing instruction?
- How should educators treat GenAI in the classroom?
- What are common patterns in mislabeling AI development?
- How does the narrator’s perspective inform popular interpretations of Northanger Abbey?
Stalled
- What are principles for an ethical framework for GenAI use in writing programs?
- What are best practices for using GenAI in writing placement?
Cultural Analytics
The same causal inference and data science toolkit I use for education policy also applies to entertainment industry data. Through the CPR/Film Substack, I study Oscar economics, franchise saturation, and other questions about how the movie business works.
Published
- Does an Oscar nomination boost box office revenue?
- How big was the 2024 Oscar nomination bump?
- Did the 2026 nominations follow the 2024 pattern?
- Is the Oscar nomination bump consistent across years?
- Does the Oscar calendar affect the nomination bump?
- Does the Oscar calendar affect the win bump?
- How much is an Oscar win worth at the box office?
- Who are the most accurate Oscar pundits?
- How saturated is the box office with franchise films?
- Who do my movie tastes overlap with?
- The Most Expensive Party Balloon in History?