Theory-driven multi-group analyses via latent growth with PLS-SEM: A two-stage anchor-factorial approach

The article below shows how one can conduct theory-driven multi-group analyses via latent growth, in the context of structural equation modeling via partial least squares (PLS-SEM).

Kock, N. (2025). Theory-driven multi-group analyses via latent growth with PLS-SEM: A two-stage anchor-factorial approach. Data Analysis Perspectives Journal, 6(4), 1-8.

Link to full-text file for this and other DAPJ articles:

https://scriptwarp.com/dapj/#Published_Articles

Abstract:

Classic multi-group analyses (MGAs) are plagued by two primary flaws: they require splitting samples into smaller subsamples, which complicates parameter comparisons due to reduced power and varying subsample characteristics; and they are typically exploratory rather than theory-driven. This paper addresses these issues by proposing a theory-driven MGA via latent growth method, using a two-stage anchor-factorial approach. The methodology is implemented within the context of structural equation modeling with partial least squares (PLS-SEM) using the WarpPLS software. The first stage involves creating and analyzing a target SEM model, and then converting a categorical grouping variable (e.g., Country) into a numeric latent growth variable (LGV) using an anchor-factorial conversion with variation sharing. The LGV is anchored on the latent variables involved in the hypothesized effects. The second stage involves inspecting the LGV's scores, related path coefficients, and 3D graphs to confirm the theory-driven effects. This approach offers a robust and theoretically superior alternative to classic MGA for assessing how multi-group influences affect model parameters.

Video demonstrating the techniques employed in the article:

Best regards to all!