Dear Professor Chin,

I am experiencing a problem regarding how to depict formative constructs in the PLS. I would highly appreciate it if you could kindly give me a hand in this matter.

For constructs with reflective indicators, I can draw a circle for the construct and depict its indicators with squares. What should I do to depict a construct (e.g. readiness) when it is measured through two subconstructs (e.g. IT sophistication and partner readiness), each of which is measured with reflective indicators?

Answer:

It seems what  you're talking about is creating a second order factor. If the number of indicators for each of your two constructs are approximately equal, you can use the method of repeated manifest variables.

I believe a brief pictorial representation is provided in my slides given at the 2000 ICIS conference and available at my website at

http//disc-nt.cba.uh.edu/chin/icis2000plstalk.pdf

This method is also mentioned  in the appendix of my ICIS 1996 conference paper on interaction effects available at:

http//disc-nt.cba.uh.edu/chin/icis96.pdf

Essentially, your overall factor that represents the two first order constructs is created by using all the indicators used for the two first order constructs.  Thus is the IT sophistication and partner readiness each had 3 indicators, you'd create the overall factor using the 6 indicators.

If you model the relationship from the second order to the first order construct, you have a traditional "molecular model" discussed in

Chin, W. W. & Gopal, A. (1995).  Adoption intention in GSS: Relative importance of beliefs.  The Data Base for Advances in Information Systems, 26(2&3), pp. 42-64.

If you model the paths from the first order to the overall second order construct, you would have specified a "molar model."  In this situation, the R-square for the second order construct will, of course, end up as 1.0.  But the reason for this model is to examine the relative path weights as this "molar construct" is used to predict other constructs in your model. The Chin & Gopal paper provides a reasonable discussion on interpretation of these models.