The choice of structural equation modeling technique matters: A commentary on Dash and Paul (2021)

Ganesh Dash and Justin Paul authored an article titled “CB-SEM vs. PLS-SEM methods for research in social science and technological forecasting” in a special issue of Technological Forecasting and Social Change, co-edited by Justin Paul. Unfortunately, the article’s central conclusion – “CB or PLS o...

Full description

Saved in:

Bibliographic Details
Published in:	Technological forecasting & social change Vol. 194; p. 122665
Main Authors:	Schuberth, Florian, Hubona, Geoffrey, Roemer, Ellen, Zaza, Sam, Schamberger, Tamara, Chuah, Francis, Cepeda-Carrión, Gabriel, Henseler, Jörg
Format:	Journal Article
Language:	English
Published:	Elsevier Inc 01.09.2023
Subjects:	Consistent partial least squares (PLSc) Covariance-based structural equation modeling (CB-SEM) Partial least squares based structural equation modeling (PLS-SEM) Structural equation modeling (SEM) Consistent partial least squares (PLSc) Covariance-based structural equation modeling (CB-SEM) Partial least squares based structural equation modeling (PLS-SEM) Structural equation modeling (SEM)
ISSN:	0040-1625, 1873-5509
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Ganesh Dash and Justin Paul authored an article titled “CB-SEM vs. PLS-SEM methods for research in social science and technological forecasting” in a special issue of Technological Forecasting and Social Change, co-edited by Justin Paul. Unfortunately, the article’s central conclusion – “CB or PLS or PLSc do not matter” – is misleading and at odds with practically all extant conceptual and empirical research on this subject. This commentary identifies an unsuitable research design to be the major cause of the erroneous conclusion and aims to set the record straight. A Monte Carlo simulation demonstrates that the choice of the approach to structural equation modeling can have a substantial impact on the results and their validity. In general, analysts should choose a structural equation modeling approach that fits their conceptual model. •A Monte Carlo simulation compares CB-SEM, PLSc, and PLS.•CB-SEM and PLSc (but not PLS) can be used for reflective measurement models with latent variables.•CB-SEM with the H–O specification and PLS Mode B can be used for composite models.•The choice of structural equation modeling technique matters.
ISSN:	0040-1625 1873-5509
DOI:	10.1016/j.techfore.2023.122665