Researchers' using both composite-based as well as co-variance based techniques for analyzing their data. Still, a lot of reviewers have a merely look at the composite-based (SmartPLS and R) techniques and sometimes commented as insufficient. What is the reason for reviewers to suggest AMOS / covariance-based over composite-based technique though we all know when and how to use both the techniques? Thanks in advance.
Theory testing vs. prediction...
Does it like old adults (AMOS) vs. new adults (R and PLS)...?
This is really a critical as well as extant question, especially for those who primarily use so-called second-generation statistical tool - structural equation modeling. As you correctly pointed out, some early career researchers struggle with this issue to "persuade" some reviewers. Some editorial boards even have open and strong reservation in accepting manuscript using PLS approach. If you really want to understand deeper on the differences between these two alternatives, especially the flaws of PLS method (as some scholars raised), I would suggest you to read this paper (attached).
I hope this would help you understand the reason why covariance-based approach is preferred to composite-based approach.
Best,
Khan Taufique
The difference between CB-SEM (Mplus, AMOS, Lavaan in R, sem in Stata) and PLS-SEM (SmartPLS, ADANCO, WarpPLS, cSEM in R, and plssem in Stata) is similar to the difference between common factor analysis and principal component analysis as explained by Chin in the following paper: Article Partial least squares is to LISREL as principal components a...
Article Estimation issues with PLS and CBSEM: Where the bias lies!
Cristian Ramos-Vera, Mehmet Mehmetoglu Khan Taufique Md. Aftab Uddin Thanks for your sharing and comments. As both methods are suitable and appropriate to some extent then why reviewers biased their judgement to a specific one? Thanks
When PLS-SEM is used instead of CB-SEM, it produces biased parameters. This comes as a result of the fact that PLS-SEM ignores measurement error in its estimation. There has thus been established skepticism about PLS-SEM among reviewers/editors. However, a recent version of PLS-SEM called consistent PLS-SEM can be used instead of CB-SEM whereas traditional PLS-SEM should be used only for prediction purposes.
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy Of Science
- Reasoning