forum.smartpls.com

Hello,

I just wondering why some people tried to re-examine the construct form (formative - reflective).
What is the justification to use formative model of a construct that traditionally has been treated as reflective?

For example, relationship quality that consists of trust, commitment, satisfaction. In traditional way, or in CB-SEM, directions are from the construct pointing to dimensions. But in one or two research, I saw that they used the same construct and dimensions, but in formative. Beside of its statistical proof (no multicollinearity, etc), what kind of theoretical justification can be addressed in our research to start evaluating or questioning form of the construct?

Thank you!

Often, old research has not though about the possibility of modeling constructs formatively. Thus, it might be warranted to re-evaluate the constructs measurement in some cases. Just because someone has done it in the past might not be a good reason to do it again.

However, as you said, it should always be guided by theory and not only by pure empirical observations. Whether it is justified to re-evaluate the nature of a constructs must always be answered on a case-by-case decision by looking at empirical proof in conjunction with theoretical arguments.

Several studies have shown an error specification of direction indicator constructs. So that re-testing is needed.

You may find this article useful:
Gudergan, S. P., Ringle, C. M., Wende, S., & Will, A. 2008. Confirmatory Tetrad Analysis in PLS Path Modeling. Journal of Business Research, 61(12): 1238-1249.

CTA-PLS has been implemented into SmartPLS 3. For the theoretical underpinnings of the construct type, you may want to take a look at this article:
Sarstedt, M., Hair, J. F., Ringle, C. M., Thiele, K. O., & Gudergan, S. P. 2016. Estimation Issues with PLS and CBSEM: Where the Bias Lies! Journal of Business Research, 69(10): 3998-4010.

Best
Christian