Katherine Delno · Statistician
Statistical research and quantitative reasoning.
I am a statistician with an M.S. in Statistics from the University of Washington and a B.S. in Mathematics from the University of Nevada, Reno. My graduate work encompassed statistical inference, modeling, causal inference, and statistical learning, with extensive computational work in R and Python.
My interests lie in the foundations of statistical methods: the assumptions on which they depend, how they behave in practice, and what their conclusions can reasonably establish. Since completing my graduate studies in 2025, I have maintained an independent mathematics and statistics teaching practice. I also accept selected statistical and quantitative projects.
Research
Selected Statistical Work
My graduate work included independent methodological research and applied projects in causal inference and machine learning. The selections below reflect an interest in statistical methods as objects of study in their own right, including the assumptions they require, the circumstances in which they perform well, and the limits of their interpretation.
RandomForestSpecCheck
Developed a permutation-based random forest diagnostic for detecting mean-structure misspecification in linear mixed models. The procedure was evaluated through simulation across varying data structures and applied to the Framingham Heart Study, with particular attention to its operating characteristics and limitations.
Causal Analysis of Food Insecurity and Type 2 Diabetes
Examined the relationship between food insecurity and type 2 diabetes using pooled national survey data, augmented inverse probability weighting, and g-computation. The analysis emphasized causal identification, the assumptions required for estimation, and the limitations of observational data.
Read project: Causal Analysis of Food Insecurity and Type 2 Diabetes
Conformal Prediction for Multi-Label Classification
Fine-tuned a DistilBERT model for multi-label movie genre classification and applied conformal prediction to construct prediction sets. The study examined empirical coverage, prediction-set size, and the practical limitations of uncertainty quantification in multi-label classification.
Read project: Conformal Prediction for Multi-Label Classification
Professional Work
Statistical and Quantitative Services
I accept a limited number of projects involving statistical analysis, methodological review, research support, and quantitative content. My approach begins with the structure of the problem: the research question, the available evidence, and the assumptions necessary to connect one to the other.
I place particular importance on reproducible analysis, transparent methodological choices, and interpretations that remain within the limits of the data.
Teaching
Private Mathematics and Statistics Instruction
I maintain a small independent teaching practice in AP Statistics, AP Calculus AB/BC, and AP Precalculus. Instruction is individual and typically extends across an academic course, allowing for close attention to a student's reasoning, developing understanding, and recurring difficulties.
My approach to teaching is informed by the same interests that shape my statistical work: the structure behind a method, the assumptions that justify its use, and the distinction between carrying out a procedure and understanding why it works.
Background & Inquiries
Further details of my academic training, research interests, and teaching history are available on the About page. For inquiries concerning statistical projects or private instruction, I can be reached by email.