Abstract

We study structural equation modeling with factors and composites in the framework of the basic design. We propose two estimation methods. The first one is a new non-iterative approach based on singular value decomposition calculations (SVD-SEM). This new approach produces consistent and asymptotically normal estimators of the parameters. Then, we describe the restricted maximum likelihood approach in the framework of the basic design. SVD-SEM and restricted ML-SEM are compared in a Monte Carlo simulation on a nonrecursive model with factors and composites. SVD-SEM and ML-SEM are compared on a case study on a recursive model with factors.