Power analysis in structural equation modeling: Principles and methods

: Structural equation modeling (SEM) is an important statistical tool in psychology, management, and sociology. However, the lack of analyses and reports of statistical power in many studies using SEM has reduced the probative power of the study results. There are three main types of statistical power analysis methods for SEM: the Satorra–Saris, MacCallum, and Monte Carlo methods. The Satorra–Saris method applies to cases where the alternative model is clear, analysis object is simple, and the statistic to be examined is based on the χ2 distribution. The MacCallum method is applicable to the case of χ2-based overall model fit tests with unknown alternative models. Further, the Monte Carlo method applies to cases where the analysis object is complex, or the test is performed using simulation or resampling methods. In practice, researchers should clarify the purposes and methods of the test as well as the availability of alternative models as a priority. The research method can then be determined on the basis of the above information.