Cengiz Zopluoglu
Cengiz Zopluoglu, Associate Professor, joined the faculty in the Department of Educational Methodology, Policy, and Leadership at the University of Oregon in June 2020 after serving seven years in the Research, Measurement, and Evaluation program at the University of Miami. Dr. Zopluoglu received his Ph.D. in Quantitative Methods in Education program from the Department of Educational Psychology at the University of Minnesota. He was teaching math in a middle school in Turkey and moved to the U.S. in 2007 after he was awarded a highly competitive scholarship from the Turkish Ministry of National Education for graduate study in the U.S. in educational measurement.
Prior to his faculty experience, Dr. Zopluoglu held research assistant positions with the Quantitative Methods in Education program and Office of Research and Consultation Services at the University of Minnesota, and also had the privilege of doing psychometric research at the Minneapolis Public Schools in Minnesota and ACT, Inc. in Iowa. Dr. Zopluoglu has taught advanced-level graduate courses on psychometrics and statistical methodology including General Linear Models, Item Response Theory, Measurement and Psychometric Theory, Categorical Data Analysis, and Data Analysis with R in Educational and Behavioral Research. Dr. Zopluoglu has substantial experience in providing statistics and measurement consultation to faculty and students on their projects.
His methodological publications have appeared in leading measurement and statistics journals including Applied Psychological Measurement, Behavior Research Methods, Educational and Psychological Measurement, Multivariate Behavioral Research, Psychometrika, Psychological Methods, and Structural Equation Modeling as well as in journals focusing on applied research.
His most recent work and personal blog can be accessed from his personal website at cengiz.me
BRT Researcher Captures Grant Prize in 2023 NCES Challenge
Summer 2023
BRT researcher Cengiz Zopluoglu, Ph.D. recently won Grand Prize in the 2023 NAEP Automated Scoring Challenge for NAEP Mathematics items
Math Scoring Data Challenge PR Newswire
McGrath explained that the NCES, challenge is part of their efforts to modernize their system by incorporating data science and machine learning into the National Center on Education Statistics’ operational activities. He wrote, that NCES “began work on this challenge with a question: ‘Can someone do this work at a level that would allow NAEP to maintain its status as the gold standard?’ ... Thanks to your hard work, we can be confident that the work is doable, and we can point to two models of excellence as evidence.”
In describing his work, Zopluoglu explained that NCES opened this competition in the spring of 2023 and released some datasets, inviting researchers across the country to submit predictions for machine scoring of ten open-ended constructed response math items where the kids provided a textual response and explained their reasoning for each math problem. More than 250,000 individual responses were originally scored by human raters for these ten math items. Zopluoglu said that he trained some models using the open-source advanced NLP models and the models performed great, yielding scores very similar to those obtained by human raters.
Zopluoglu further noted his appreciation for the support he received on this project from BRT and the University of Oregon’s Research Advanced Computing Services and more specifically the UO’s computing cluster Talapas. “Without these supports,” he explained, “There is no way I could train so many models that required specialized software in a short amount of time. Talapas made it possible. Without the support from BRT and Talapas, this wouldn't happen.”
Zopluoglu’s ground-breaking data science / machine learning work is part of his ongoing research with real-world applications. We look forwards to sharing similar machine-scoring innovations through the CBMSkills and WriteRightNow! applications in the near future.