Page 1 of 1

categorical independent variable in reflective-formative second order construct

Posted: Tue May 31, 2022 2:50 pm
by evaking04
Hi friends! I am struggling with the right second-order reflective-formative model set up in SmartPLS3. I am currently researching the impact AI in the strategic sourcing process (AI in SS as my independent variable, categorical measured with Yes or No use) on 5 capabilities (Sourcing Flexibility, SC Responsiveness, Efficient use of resources, Strong SRM, and SC Visibility - all measured on a 7-point Likert-scale) and how it impacts supply chain resilience (SCRES also measured on a 7-point likert-scale).

I believe my issues are with my "AIinSS" variable which is showing 0.000 path coefficients for all 5 capabilities and with my SCRES variable as it is also showing 0.000 for all. Would mean all of my hypothesis are rejected as there isn't a hint of a relationship. Cannot be true as I tested the same model with fictional data which should 100% show significant results - still nothing. And if check path coefficients in R, it gives me signifcant results for 9/10 hypotheses.

I created the second order construct by pulling each of the items reflective to the first-order construct into the formative second-order item as well. The yes and no in AIinSS is coded as either 0 or 1. Any advice is highly appreciated!

I am also happy to jump on a quick zoom call and reward the time spent via paypal!

Thank you so much, Eva