Despite its power, SmartPLS is not a magic wand. It is a tool for exploration and prediction, but it cannot fix poor theory or bad data. Critics often point out that because PLS-SEM is so flexible, it can be misused to "fish for significance"—running models until a statistically significant result appears, regardless of whether the theory supports it. Furthermore, because it maximizes variance, the "fit indices" used to judge the quality of a model differ from those in CB-SEM, requiring researchers to be diligent in their interpretation of model fit.
When applying this framework in SmartPLS, researchers typically model the following six variables as latent constructs: smartpls
: Researchers define the PIECES variables as independent variables that affect a dependent variable, such as "User Satisfaction". Despite its power, SmartPLS is not a magic wand