AN OVERVIEW OF STRUCTURAL EQUATION MODELING IN COGNITIVE AGING RESEARCH

Richard N. Jones, Sc.D
Associate Director, Social and Health Policy Research, Institute for Aging Research,
Assistant Professor of Medicine, Hebrew SeniorLife
Harvard University Medical School, Boston, MA, USA

Research in cognitive aging involves behaviors and traits that are not measured directly. Examples include general and specific areas of neuropsychological performance. Research questions often concern how persons change over time or with aging. Both of these aspects of cognitive aging research questions can be approached with methods involving latent variables. Factor analysis and item response theory models can be used to measure traits and behaviors not directly observed. Structural equation models can be used to relate latent variables to other latent variables, to background characteristics, and describe changes over time and with age. Structural equation models therefore represent a concise and unified analytic framework for addressing measurement and change in cognitive aging research, and can be used to address substantive, confounding, and methodological challenges often encountered in applied research settings. This workshop presents an overview of some applications of general latent variable and structural equation modeling in applied cognitive aging research. Emphasis is on a broad survey of applications, results, challenges, and opportunities for future research. As a result of attending this workshop: (1) participants will be introduced to the concepts and terminology relevant to structural equation modeling as applied to cognitive aging research, (2) participants will understand the range of successful applications and the challenges and future opportunities for structural equation modeling in cognitive aging research.