The aim of this paper was to provide a comprehensive introduction of the Bayesian approach for SEM estimation. Despite receiving a strong attention across other related fields, the use of the Bayesian approach is still highly limited in the tourism literature. We highlighted in this paper the power of the Bayesian approach and discussed its distinctive difference from the traditional covariance-based approach to SEM estimation.Overall, we believe there are five main reasons why tourism researchers might select the Bayesian approach for SEM estimation.First, some complicated models such as the ones discussed in the previous section are harder to converge with traditional methods (e.g. mixture models; non-normal models, etc.), and some models are not even possible to estimate. Bayesian statistics can also help in model identification and result in more accurate parameter estimates (Depaoli, 2013; 2014).Second, “many scholars prefer Bayesian statistics because they believe population parametersshould be viewed as random” (Depaoli & van de Schoot, 2015, p. 3).Third, with the Bayesian approach one can prior information into the estimation. Fourth, as highlighted several times above, the Bayesian statistics is not based on large samples. This was also reinforced by the results of our Monte Carlo simulation. Fifth, and finally, the Bayesian approach offers more accurate and less sensitive fit statistics and model comparison tools.Despite all these advantages, the main goal should not be understood as encouraging some naïve applications of the Bayesian approach, or even using the Bayesian approach in the interest of“mathematistry”. We understand that most researchers in tourism are usually more comfortable using the frequentist approach for SEM estimation. As indicated by Depaoli and van de Schoot (2015), using the Bayesian approach without good knowledge of the method can be dangerous, particularly in terms of interpreting the Bayesian features and/or results. The Bayesian approach can also be sensitive to the selection of appropriate priors ebut this is an empirical matter. From here, conducting sensitivity analysis to check whether the results are stable across prior choices becomes essential (Assaf et al., 2016). There are also other important steps that should be checked when using the Bayesian approach-we refer the reader to the study of Depaoli and van de Schoot (2015) for more details.