Cart is empty 
PLS-SEM is a variance-based approach to structural equation modelling (SEM) that differs from other SEM methods in several aspects. Some of the main differences are:
- PLS-SEM focuses on maximising the explained variance of the endogenous latent variables, while other SEM methods (such as CB-SEM) focus on minimising the discrepancy between the observed and implied covariance matrices.
- PLS-SEM does not require data to be normally distributed or to have a large sample size, while other SEM methods (such as CB-SEM) have more stringent assumptions and requirements for data quality and quantity.
- PLS-SEM can handle both reflective and formative measurement models, while other SEM methods (such as CB-SEM) are more suitable for reflective measurement models.
- PLS-SEM is more suitable for exploratory and predictive research, while other SEM methods (such as CB-SEM) are more suitable for confirmatory and explanatory research.
For more details on the differences between PLS-SEM and other SEM methods, you can refer to the following sources:
- Which SEM to Use and What to Report? A Comparison of CB-SEM and PLS-SEM (Usakli & Rasoolimanesh, 2023)
- An Introduction to Structural Equation Modelling (Hair et al., 2021)
- PLS-SEM Compared with CB-SEM
How can I choose between PLS-SEM and CB-SEM for my research?
Choosing between PLS-SEM and CB-SEM for your research depends on several factors, such as your research goal, your data characteristics, your measurement model, and your model complexity. Here are some general guidelines that may help you decide:
- If your research goal is to predict key target constructs or identify key driver constructs, you may want to use PLS-SEM, as it focuses on maximising the explained variance of the endogenous latent variables.
- If your research goal is to test or confirm a theory or compare alternative theories, you may want to use CB-SEM, as it focuses on minimising the discrepancy between the observed and implied covariance matrices.
- If your data is not normally distributed or has a small sample size, you may want to use PLS-SEM, as it does not require these assumptions and can handle non-normal and small data.
- If your data are normally distributed and have a large sample size, you may want to use CB-SEM, as it can provide more reliable and valid estimates of the model parameters and fit indices.
- If your measurement model includes formative indicators, you may want to use PLS-SEM, as it can handle both reflective and formative measurement models.
- If your measurement model includes only reflective indicators, you may want to use CB-SEM, as it is more suitable for reflective measurement models.
- If your structural model is complex (many constructs and many indicators), you may want to use PLS-SEM, as it can handle complex models without losing much statistical power.
- If your structural model is simple (few constructs and few indicators), you may want to use CB-SEM, as it can test more sophisticated models, such as those with higher-order factors, latent interactions, latent growth curves, or multilevel structures.
Of course, these are not absolute rules but rather suggestions based on the literature. You should also consider the specific context and purpose of your research when choosing between PLS-SEM and CB-SEM.
References
Usakli, A., & Rasoolimanesh, S. M. (2023). Which SEM to Use and What to Report? A Comparison of CB-SEM and PLS-SEM. In Cutting Edge Research Methods in Hospitality and Tourism (pp. 5-28). Emerald Publishing Limited.
Hair, J. F., Hult, G. T. M., Ringle, C. M., Sarstedt, M., Danks, N. P., Ray, S., … & Ray, S. (2021). An introduction to structural equation modeling. Partial least squares structural equation modelling (PLS-SEM) using R: a workbook, 1-29.
