Three currents, one through-line.
The work moves between psychometric rigor and applied policy questions — with a consistent focus on the students who are easiest for measurement systems to overlook.
Accessibility & universal design in assessment
How can we tell whether a testing accommodation actually helps a student — or whether the test itself is the problem? This research line uses NAEP process data (the timestamped, keystroke-level record of what students do during a digital assessment) to distinguish universal-design features that benefit everyone from accommodations that benefit specific populations.
Recent papers in Educational Measurement: Issues and Practice and the International Electronic Journal of Elementary Education introduce process-data methods for identifying students who would benefit from extended time, and for evaluating whether universal-design changes to digital assessments produce equitable improvements. Earlier work in the Journal of Learning Disabilities examined accessibility supports for students with decoding difficulties on large-scale reading assessments.
Course-taking patterns & postsecondary transitions
What students take in high school shapes what comes after — but the way we measure "rigor" of coursework has been inconsistent across studies, leading to different conclusions about the same students. This line of work asks how different operationalizations of course-taking change what we conclude, and traces course-taking sequences forward into college enrollment, STEM major choice, and occupational expectations.
Recent papers in Educational Measurement: Issues and Practice (Ogut, Yee, Circi, & Dizdari, 2023; Ogut & Circi, 2023) and AERA Open (Bohrnstedt, Ogut, Yee, & Bai, 2023) document gaps in advanced coursework across student groups, the consequences of those gaps for postsecondary STEM outcomes, and the methodological choices that change the story.
AI & machine learning in education research
Data mining and machine learning give education researchers powerful new tools — and the responsibility to apply them in ways that don't reproduce the biases of the data they were trained on. This work combines NLP and supervised learning approaches with traditional psychometric methods to study problem-solving processes in mathematics, predict NAEP item difficulty from passage features, and explore log-file data from large-scale assessments.
The Computers in the Schools paper (Ogut, Webb, Hicks, Circi, & Yin, 2024) applies process mining to NAEP mathematics log data; the founding of the Inclusive AI Institute extends this research orientation into practice, with a focus on AI tools that include learners with disabilities by design.
– present
The Educational Divide: Transition, Retention, and Course Selection in Digital and On-Campus Immersion Students
– 2023
Rethinking Accessibility Using NAEP Process Data: Exploring Universal Design and Accommodations
– 2022
The Relationship Between Course-Taking Patterns and Postsecondary Outcomes
– present
National Assessment of Educational Progress (NAEP) — College Preparedness Benchmarks & Related Studies
– 2016
Girls' Opportunities to Access Learning (GOAL & GOAL Plus), Liberia
Looking for a research partner?
I welcome inquiries about co-investigation, methods consultation, and commissioned research — particularly on assessment accessibility, course-taking, and applied AI in education.
Start a conversation