Steven Culpepper, University of Illinois Urbana-Champaign
Refining Hidden Markov Models for Diagnostic Research
Diagnostic models provide a framework for providing researchers and educators with fine-grained information for understanding student performance on assessments. Recent research developed exploratory methods for inferring the underlying attributes or skills and extended the diagnostic modeling framework to longitudinal settings. We review advances in longitudinal diagnostic models and discuss current challenges. Many longitudinal diagnostic models rely on the first-order hidden Markov modeling (HMM) framework. However, HMMs require two identification assumptions, which may be too stringent for educational applications. Specifically, HMMs assume an irreducible transition model and homogeneous emission probabilities, which may be untenable for learning applications that involve absorbing states or different items over time. We review the challenges and discuss several solutions.
about the speaker
Dr. Steven Andrew Culpepper is a Professor and Data Science Founder Professorial Scholar in the Department of Statistics at the University of Illinois at Urbana-Champaign. He earned a Ph.D. in Educational Psychology from the University of Minnesota Twin Cities. Dr. Culpepper’s research examines issues related to measurement, item response theory, latent class models, and Bayesian computation for psychometric models. His recent research focuses on restricted latent class models for applications involving diagnostic assessments and the fine-grained evaluation of learning interventions. To date, he coauthored over 60 publications and he received support for several psychometric projects from the Spencer Foundation, the American Educational Research Association, and the National Science Foundation. Professor Culpepper is Editor-in-Chief of the Journal of Educational and Behavioral Statistics and Associate Editor of Psychometrika. His professional service includes membership on the Board of Trustees of the Psychometric Society (2018-2022), the Grants Program Governing Board of the American Educational Research Association, the Technical Review Panel of the National Indian Education Study, and the Design and Analysis Committee for the National Assessment of Educational Progress. Professor Culpepper collaborates on interdisciplinary teams of scholars, advises undergraduate and doctoral students, and works to broaden participation of under-represented and first-generation college students in the mathematical and data sciences.