A multivariate analysis technique: Structural equation modeling

This research paper gives a short and non technical introduction about a multivariate analysis technique named Structural Equation Modeling (SEM). SEM is a group of statistical techniques used to analyze the relationships of observed and latent variables which also considers residual errors. We have given preliminary information of SEM for beginners. The set of rules developed by many researchers are also listed to determine the sample size before data collection. The model fit criteria have been discussed to decide cut off values.