On the Use of Formative Measurement Specifications in Structural Equation Modeling: A Monte Carlo Simulation Study to Compare Covariance-Based and Partial Least Squares Model Estimation Methodologies
The broader goal of this paper is to provide social researchers with some analytical guidelines when investigating structural equation models (SEM) with predominantly a formative specification. This research is the first to investigate the robustness and precision of parameter estimates of a formative SEM specification. Two distinctive scenarios (normal and non-normal data scenarios) are compared with the aid of a Monte Carlo simulation study for various covariance-based structural equation modeling (CBSEM) estimators and various partial least squares path modeling (PLS-PM) weighting schemes. Thus, this research is also one of the first to compare CBSEM and PLS-PM within the same simulation study. We establish that the maximum likelihood (ML) covariance-based discrepancy function provides accurate and robust parameter estimates for the formative SEM model under investigation when the methodological assumptions are met (e.g., adequate sample size, distributional assumptions, etc.). Under these conditions, ML-CBSEM outperforms PLS-PM. We also demonstrate that the accuracy and robustness of CBSEM decreases considerably when methodological requirements are violated, whereas PLS-PM results remain comparatively robust, e.g. irrespective of the data distribution. These findings are important for researchers and practitioners when having to choose between CBSEM and PLS-PM methodologies to estimate formative SEM in their particular research situation.
|Date of creation:||2009|
|Date of revision:|
|Contact details of provider:|| Postal: |
Web page: https://mpra.ub.uni-muenchen.de
More information through EDIRC
Please report citation or reference errors to , or , if you are the registered author of the cited work, log in to your RePEc Author Service profile, click on "citations" and make appropriate adjustments.:
- Albert Satorra & Peter M. Bentler, 1999.
"A scaled difference chi-square test statistic for moment structure analysis,"
Economics Working Papers
412, Department of Economics and Business, Universitat Pompeu Fabra.
- Albert Satorra & Peter Bentler, 2001. "A scaled difference chi-square test statistic for moment structure analysis," Psychometrika, Springer, vol. 66(4), pages 507-514, December.
- Tenenhaus, Michel & Vinzi, Vincenzo Esposito & Chatelin, Yves-Marie & Lauro, Carlo, 2005. "PLS path modeling," Computational Statistics & Data Analysis, Elsevier, vol. 48(1), pages 159-205, January.
- Allen Fleishman, 1978. "A method for simulating non-normal distributions," Psychometrika, Springer, vol. 43(4), pages 521-532, December.
- Jarvis, Cheryl Burke & MacKenzie, Scott B & Podsakoff, Philip M, 2003. " A Critical Review of Construct Indicators and Measurement Model Misspecification in Marketing and Consumer Research," Journal of Consumer Research, University of Chicago Press, vol. 30(2), pages 199-218, September.
- Albert Satorra, 1990. "Robustness issues in structural equation modeling: a review of recent developments," Quality & Quantity: International Journal of Methodology, Springer, vol. 24(4), pages 367-386, November.
- Claes Cassel & Peter Hackl & Anders Westlund, 1999. "Robustness of partial least-squares method for estimating latent variable quality structures," Journal of Applied Statistics, Taylor & Francis Journals, vol. 26(4), pages 435-446.
- Wold, Herman, 1974. "Causal flows with latent variables : Partings of the ways in the light of NIPALS modelling," European Economic Review, Elsevier, vol. 5(1), pages 67-86, June.
When requesting a correction, please mention this item's handle: RePEc:pra:mprapa:15390. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Ekkehart Schlicht)
If references are entirely missing, you can add them using this form.