Katerina Marcoulides is rethinking large datasets
Working with big datasets is increasingly challenging for researchers, especially in psychology, where studies rely on complex, detailed data. A recent study, “A Bayesian Synthesis Approach to Structural Equation Modeling Data Fusion for the Analysis of Massive Data,” by Katerina Marcoulides, faculty, with two Quantitative & Psychometric Methods graduate students, Xinyu Liu and Hannah Hamling, has been published by Structural Equation Modeling: A Multidisciplinary Journal and explores a more practical approach to this problem.
This research addresses computational and estimation challenges of applying structural equation modeling to massive datasets by proposing a Bayesian Synthesis data-fusion approach combined with a modified divide-and-conquer strategy. Illustrated using both simulated and empirical data, this method improves both computational feasibility and estimation performance for complex structural equation models with massive datasets.
Katerina Marcoulides, PhD, Associate Professor in the Department of Psychology, College of Liberal Arts, University of Minnesota; faculty and Program Director of Quantitative and Psychometric Methods (QPM), and director of the Data Analytics & Visualization Lab.
Composed by Nguyen Kiet Pham, communications assistant.
