On the nature of emotions: Defining a theoretical model based on data-mining through the Confirmatory Tetrad Analysis for PLS-SEM.

The psychological dimension of service experience is becoming increasingly important, and modern data science techniques allow its extraction even from atypical data, such as reviews provided by users on the web. Starting from these techniques and Google Maps data, an exploratory study developed a Customer Satisfaction model based on Plutchik’s taxonomy of emotions, which accounts for the final rating attributed to restaurant services. This study seeks to assess and determine its nature (formative vs. reflective) through a statistical criterion: the Confirmatory Tetrad Analysis (CTA-PLS). CTA-PLS protects against model misspecification and allows to empirically verify whether the theoretical specification is supported by the data. Results seem to confirm a formative nature, where the first-order constructs represent emotionally connoted, yet differentiated aspects of the general service experience, which are in line with Plutchik’s theory. CTA-PLS emerges as a useful statistical decision criterion for investigating the nature of measurement models.