A comparative study of covariance and partial least squares based structural equation modelling in hospitality and tourism research

Purpose Structural equation modelling (SEM) has increasingly been used by hospitality and tourism researchers to examine complex relationships. This paper aims to highlight the benefits and limitations of SEM for hospitality and tourism research and compare its two main approaches, i.e. covariance-based SEM (CB-SEM) and partial least squares-SEM (PLS-SEM). Design/methodology/approach By using a comparative approach, this study parallels SEM’s two main approaches, i.e. CB-SEM and PLS-SEM, using three different examples from hospitality and tourism industry. Both the approaches are compared side by side in terms of assumptions, validity and reliability of measurement models, item retention and loadings, strength and significance of path relationships and coefficient of determinations. Findings The findings show that even though both methods analyse measurement theory and structural path models, there are relatively higher advantages for hospitality and tourism researchers in applying PLS-SEM. Research limitations/implications Because of the limitations of only using three examples, the results and trends generated in this study may not be generalized to all research in hospitality and tourism discipline. Moreover, the Likert scale has been used to measure the constructs in both the studies, which may have biased the results. Originality value This study is the first to compare the usage of both the SEM approaches in hospitality and tourism research. The findings of this study provide significant implications and directions for hospitality and tourism researchers to apply PLS-SEM in the future.