Having Your Cake and Eating It Too: Multiple Dimensions and a Composite

Abstract: This presentation will focus on multiple psychometric modeling alternatives for multidimensional measurement situations where both the individual dimension outcomes, and the summative combination of those multiple dimensions is of interest. We will briefly present the current approaches in the literature such as (a) the uni- and multidimensional models, (b) the bi-factor model, and (c) the higher-order model. We will then describe the fourth approach, the composite approach, that we see as having certain advantages over the others. We discuss its estimation, the variety of weighting schemes that might be used in creating the composite, and the calculation of reliability for the composite. We will present empirical example with different solutions and discuss their qualities.

Perman Gochyyev is currently a research psychometrician at the University of California, Berkeley at the Berkeley Evaluation and Assessment Research (BEAR) Center. Perman received his Ph.D. in Quantitative Methods and Evaluation from UC Berkeley in 2015. His research focuses on latent variable and multilevel modeling, multidimensional and ordinal IRT models, latent class models, and issues related to causal inference in behavioral statistics.

Mark Wilson's interests focus on measurement and applied statistics. His work spans a range of issues in measurement and assessment from the development of new statistical models for analyzing measurement data, to the development of new assessments in subject matter areas such as science education, patient-reported outcomes and child development, to policy issues in the use of assessment data in accountability systems.

Tuesday, October 16, 2018 - 2:00pm
2121 Berkeley Way
PDF icon composite_model_BEAR_Oct162018.pdf836.62 KB