Kathleen Gates, PhD, assistant professor of quantitative psychology in the Department of Psychology at the University of North Carolina at Chapel Hill, “Clustering Heterogeneous Samples During Model Selection,” at 11 a.m. on Thursday, April 23, in Rooms 407A/B BAUM, 5607 Baum Blvd., The Offices at Baum.
Abstract: Individuals often differ in their processes across time. For instance, functional MRI studies have revealed that individuals vary in the models describing the temporal brain processes. This heterogeneity occurs even within predefined, arbitrary, categories (such as those based on diagnoses or gender), indicating that these groups by be further classified or perhaps are better classified along a different dimension. These results suggest a need for approaches which can accommodate individual-level heterogeneity as well as classify individuals based on similarities in processes. In this presentation I’ll introduce a new method for clustering individuals based on their dynamic processes. The approach builds from unified structural equation modeling (also referred to as structural vector autoregression), which is a statistical method for estimating effects among variables across time. Here, unsupervised classification of individuals occurs during a data-driven model selection procedure for arriving at unified structural equation models for each individual. It has several advantages over existing methods for classifying individuals. In particular, it places no assumption on homogeneity of predefined subgroups within the sample and utilizes individual-level parameters that have been shown to be more reliable than some competing approaches. This flexible statistical technique will be illustrated with a simulation study and empirical functional MRI data.
Biography: Dr. Gates obtained a BS degree in psychology from Michigan State University, an MA in Forensic Psychology from the City University of New York, and a PhD in human development and family studies from Penn State University. She currently is an assistant professor in the quantitative psychology program at the University of North Carolina at Chapel Hill. Her current work focuses on the development, testing, and dissemination of analytic methods for use with intensive longitudinal data such as psychophysiological (e.g., functional MRI and electrocardiogram), daily diary, and dyadic interaction (e.g., mother-infant). An overarching goal of these pursuits is to arrive at reliable inferences that generalize to the population while also gleaning meaningful information regarding nuances in processes for individuals and subgroups.