XI Conference – European Congress of Methodology

The Structural-after-Measurement (SAM) approach: updates and extensions.

23 Jul 2025, 08:45

15m

Faculty of Social Sciences and Communication. (The Pyramid)/13 - Room (Faculty of Social Sciences and Communication. (The Pyramid))

Oral Presentation Statistical analyses Session 6 : "Structural models and complex data analysis"

Speaker

Yves Rosseel (Ghent University)

Abstract

In the Structural-after-Measurement (SAM) framework for Structural Equation Modeling (SEM), parameters of the measurement model are estimated first, followed by the estimation of the structural model parameters. This presentation focuses on the 'local' SAM (LSAM) approach, where summary statistics (mean, covariance matrix) of latent variables are derived in the first step. In this presentation, I will present recent LSAM developments, including: 1) incorporating binary or ordinal indicators in the measurement model, 2) integrating multiple interaction and quadratic terms into the structural model, and 3) applying the infinitesimal jackknife technique to obtain local two-step standard errors. All these extensions have been incorporated in the sam() function, which is part of the R package lavaan.

Presentation materials

rosseel_EAM2025.pdf