I develop statistical methods for causal discovery and algorithmic fairness. I've applied these methods to quantify racial bias in peer review and ballot order bias in election administration. In addition, I have worked in computational algebraic geometry (see publications sidebar), MCMC sampling algorithms (undergraduate thesis), machine learning in genomics (internship at Microsoft Research), and medicine (UW statistical consulting program).

Publications

• Refinement: Measuring informativeness of ratings in the absence of a gold standard (published in the British Journal of Mathematical and Statistical Psychology, March 2022) [SSRN open-access preprint]. With Marina Meila, Elena Erosheva, and Carole Lee.
• Causality, Fairness, and Information in Peer Review (dissertation, published by ProQuest, October 2021) [Google Scholar open-access version].
• Alternative grant models might perpetuate Black–White funding gaps (published in The Lancet, October 2020). With Carole Lee and Elena Erosheva.
• NIH peer review: Criterion scores completely account for racial disparities in overall impact scores (published in Science Advances, June 2020). With Elena Erosheva, Mei-Ching Chen, Mark Lindner, Richard Nakamura, and Carole Lee.
• Testing for bias in order assignment, with an application to Texas election ballots (published in the Journal of Statistical Planning and Inference, May 2020) [SSRN open-access preprint]. With Michael Perlman and Darren Grant.
• Polynomial-Time Amoeba Neighborhood Membership and Faster Localized Solving (published in Topological and Statistical Methods for Complex Data; Springer, 2014). With Eleanor Anthony, Peter Gritzmann, and J. Maurice Rojas.

In Progress

Software and Data

If you're looking for code samples:

• For a reproducible data analysis script, see reproduce_Erosheva-et-al.Rmd from the NIH project's OSF repository (below).
• For production-type code, see targetedordertest, a basic R package implementing the targeted tests for uniform randomness of orderings from Grant et al. (2020).

Open Science Foundation repository for NIH peer review data and Erosheva et al. (2020) reproduction code