Higher order level to simple model

Small talk on SmartPLS that does not correspond to the other forums!
Post Reply
JH Tan
PLS Junior User
Posts: 4
Joined: Tue Sep 12, 2017 2:11 pm
Real name and title: JH Tan

Higher order level to simple model

Post by JH Tan » Tue Sep 12, 2017 2:34 pm

i have a higher order model, 7 latent variable namely Kanban, SMED, JIT, TPM, 5S, Kaizen and VSM. all these connect to a second order latent variable named "lean tool adoption level".

just wondering if i can just make a simple by get the mean of first order latent variable" Kanban, SMED, JIT, TPM, 5S, Kaizen and VSM" and make it become indicator of "Lean tool adoption level"(latent variable) to form a simple PLS-SEM model? as attached
forum1.png (9.31 KiB) Viewed 2670 times

User avatar
SmartPLS Developer
Posts: 812
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)

Re: Higher order level to simple model

Post by cringle » Fri Sep 29, 2017 4:15 pm


You could run a factor analysis for each lower order construct. If the reliability is high (e.g., Cronbach's alpha > 0.8), this approach may work.

Alternatively, you can use the repeated indictor approach or 2-stage approach in PLS-SEM to analyze second-order models. You find a chapter on this topic in the advanced PLS-SEM book: https://www.smartpls.com/documentation/ ... sem-issues

Best regards

Post Reply