Measurement Invariance in Longitudinal Data
Description2010-2011 Methodology Applications Series
In longitudinal data, the interpretation of change over time is often ambiguous due to a co-dependence between changes in the construct over time and changes in the measurement tool and its properties. Because the primary goal in longitudinal data analysis is to articulate changes over time on the latent constructs, measurement of those constructs must be constant, or invariant. This presentation will (a) discuss the psychometric properties that should be assessed to achieve measurement invariance based on item response theory, and (b) introduce statistical models that can be used to make inferences for longitudinal data based on meeting an assumption of longitudinal measurement invariance.
Ji Hoon Ryoo received his PhD in Quantitative Methods in Education from the University of Minnesota in 2010. He is currently a postdoctoral fellow in CYFS working with Dr. James Bovaird on the area of statistical approaches and research methods for the educational and social-behavioral sciences. His research interests include longitudinal and multilevel modeling, educational measurement and program evaluation.