How to Validate Formative Model in SMART-PLS
Before Discussing How to Validate Formative Model, To Review What is Formative Construct and How does it differ from Reflective Construct, Click Here
Assessment of formative measurement model for collinearity issues
Since the indicators are not essentially inter-changeable, high correlations are not expected between indicators in formative measurement models. In fact, high correlations between two formative indicators are known as collinearity (Hair et al. 2014).
High levels of collinearity between formative indicators are problematic because they have an impact on the estimation of weights and their statistical significance.
In order to assess the level of collinearity in PLS-SEM, Variance Inflation Factor (VIF) is looked at. There are two widely accepted rules of thumbs:
- If VIF 5 or higher, it indicates a potential issue with collinearity problem (Hair, Ringle & Sarstedt, 2011).
- If VIF 3.3 or higher, it indicates a potential issue with collinearity problem (Diamantopoulos & Siguaw, 2006).
Source: Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM) Book (Hair et al., 2017)