Research Projects
Methodological research programs grounded in applied quantitative work
My research is methods-first and applications-grounded. Methodological programs form the core of my work, while applied studies in education and social research generate substantive questions, difficult data structures, and recurring problems that motivate new quantitative research.
Developing and evaluating quantitative methods
Variance priors for Bayesian multilevel models
Informative prior specification for variance components when higher-level sample sizes are limited, including simulation-based evaluation and practical calibration guidance.
Selecting, translating, and reporting variance priors
A practical framework for choosing variance priors, translating them across parameterizations and software, conducting sensitivity analyses, and reporting prior decisions transparently.
Sparse ordinal CFA and model-fit decisions
How sparse response categories and category collapsing change fit-index behavior, model-fit decisions, and the calibration of practical decision rules.
Calibrating informative priors across models and software
Developing and evaluating informative prior calibration across Bayesian multilevel and item-response models, including work on borrowing strength, parameterization, and sensitivity to limited information.
Dyadic multilevel modeling with limited information
A connected research program addressing role parameterization, small numbers of dyads, random effects, estimands, and post-estimation in longitudinal dyadic models.
Educational measurement, learning, and evaluation
Applications of measurement, causal inference, and explainable machine learning to mathematics achievement, large-scale assessment, educational evaluation, and computational learning experiences.
Human trafficking measurement and evaluation
Quantitative work on child trafficking prevalence, correlates, latent constructs, complex surveys, cross-context measurement, and intervention or program evaluation.
Sampling, comparability, and measurement across settings
Applied studies using respondent-driven sampling, link-tracing designs, measurement validation, and cross-context comparisons to evaluate what quantitative conclusions are supportable across populations and settings.