Shenghai Dai
  • Home
  • Research
  • R Packages
  • Projects
  • Publications
  • Teaching
  • About

Projects

My research projects bring together educational measurement, psychometrics, quantitative methodology, large-scale data, and machine learning to address methodological and substantive questions in education, psychology, and related fields.

The projects below represent selected areas of my current and recent research. They illustrate how I connect methodological development with applications involving large-scale assessments, longitudinal data, digital assessment, and complex research datasets.

Research Programs

Large-Scale Assessment and NAEP

My work in large-scale assessment focuses on methodological and substantive questions involving NAEP and other large-scale educational assessments. This includes complex sample designs, variance estimation, psychometric methods, assessment data, and the use of large-scale data to investigate educational questions.

Current and recent work includes:

  • Analysis of NAEP data to examine educational outcomes and equity across student populations
  • Statistical methods for complex-sample variance estimation
  • Research using NAEP process data from digitally based assessments
  • Applications involving language status, disability, mathematics, reading, and science achievement
  • Training and capacity building for researchers using large-scale assessment data

Learn more →

Machine Learning in Education and Psychology

I investigate how machine learning can complement traditional statistical and psychometric approaches.

This work considers both methodological and substantive applications of machine learning, including:

  • Prediction and classification
  • Feature selection and regularization
  • Tree-based and ensemble methods
  • Neural networks and other computational approaches
  • Machine learning for complex and longitudinal data
  • Integration of theory-, measurement-, and data-driven approaches

A central goal is to use machine learning not simply as a predictive tool, but as part of a broader quantitative framework for addressing questions in education and psychology.

Longitudinal and Complex Data

This work addresses methodological challenges in longitudinal, multisource, and complex datasets, including advanced quantitative modeling, missing data, and psychometric harmonization.

Areas of interest include:

  • Longitudinal modeling
  • Complex and high-dimensional data
  • Missing data
  • Psychometric harmonization
  • Integration of multiple data sources
  • Computational approaches for large-scale longitudinal research

These projects span educational, psychological, and health-related research settings.

Digital Assessment and Process Data

I study digital assessment and process data to better understand how learners interact with digitally based assessments and how these data can provide additional evidence about engagement, behavior, and learning.

Psychometrics and Research Software

My methodological research is also reflected in the development of R packages and research software for psychometric and quantitative analysis. These tools translate methodological work into reproducible resources for researchers and students.

My R packages address topics including:

  • Subscore computation
  • Missing item response imputation
  • Differential item functioning
  • Diagnostic test evaluation

Explore my R packages →

Related Research

Many of my projects cross multiple methodological areas rather than fitting within a single category. For example, research involving large-scale assessment may simultaneously involve psychometrics, complex-sample methods, machine learning, and digital process data.

This interdisciplinary approach reflects my broader goal of developing and applying rigorous quantitative methods for complex data and challenging research questions.

View publications →

© 2026 Shenghai Dai

 
  • WSU Large-Scale Data Lab