Structural equation models (SEM) are frequently used in information systems (IS) to analyze and test theoretical propositions. As IS researchers frequently reuse measurement instruments and adapt or extend theories, they frequently re-estimate regression relationships in their SEM that have been examined in previous studies. We advocate the use of Bayesian estimation of structural equation models as an aid to cumulative theory building; Bayesian statistics offer a statistically sound way to incorporate prior knowledge into SEM estimation, allowing researchers to keep a “running tally” of the best estimates of model parameters.
This tutorial on the application of Bayesian principles to SEM estimation discusses when and why the use of Bayesian estimation should be considered by IS researchers, presents an illustrative example using best practices, and makes recommendations to guide IS researchers in the application of Bayesian SEM.
Evermann, J., & Tate, M. (2014). Bayesian Structural Equation Models for Cumulative Theory Building in Information Systems―A Brief Tutorial Using BUGS and R. Communications of the Association for Information Systems, 34, pp-pp. https://doi.org/10.17705/1CAIS.03476