A Comparative Analysis of Covariance-based,Consistent and Partial Least Squares Structural Equation Modelling for Examining Customer Loyalty in the Indian Hotel Industry

Abstract

The study evaluates the appropriateness and effectiveness of covariance-based, consistent and partial least squares structural equation modelling by comparing their outcomes on a proposed conceptual model while examining factors impacting customer loyalty in the Indian hotel industry. Using 382 responses collected through a structured survey, the model was evaluated using SmartPLS and AMOS software packages. The study underscores the superiority of partial least squares in handling non-normal data and small samples. At the same time, covariance-based method is more appropriate for larger samples and factor-based models. A consistent method demonstrates advantages in path coefficients and R² values while examining the model. Managerial implications emphasize tailoring different analytical methods to a specific research context. At the same time, researchers are advised to consider study-specific characteristics, such as sample size and model complexity, in their respective research problems. The study has limitations pertaining to the moderate sample size, the absence of control variables and dependency on a specific software. Future research could explore more complex models, incorporate formative constructs and address cross-industry applications. This study contributes to the methodological rigour of structural equation modelling approaches, providing valuable guidance to academics and practitioners in selecting the most appropriate analytical tools.