Detecting Latent Classes in Tourism Data Through Response-Based Unit Segmentation (REBUS) in Pls-Sem

Abstract

This research note describes Response-Based Unit Segmentation (REBUS), a “latent class detection” technique used in partial least squares–structural equation modeling (PLS-SEM) to examine data heterogeneity. The research note is presented in two parts: Part A presents an overview of REBUS, including its development, algorithm, and its primary functions. Part B demonstrates the application of REBUS in examining a validated tourism model of destination image, satisfaction, and destination loyalty. The example shows how REBUS is used to examine variances in a structural equation model, to detect “classes,” and to profile and understand the heterogeneous groups in an SEM context. REBUS is powerful in uncovering variances and possible moderators in structural models, especially when the data are cross-sectional, heterogeneous, and multivariate nonnormal. Finally, the research note demonstrates how REBUS detects classes in models with higher order (multidimensional) constructs, which are often the case in tourism research.