Generally, structural equation modeling (SEM) is similar to simultaneous equation modeling (SIM). Econometricians and statisticians mostly use SIM than SEM, while marketers, psychologists, sociologists and educationists, among others, use more of SEM than SIM because of, perhaps, their belief in the existence of latent variables.
Both are the same in term of analyzing the inter-relationships among the constructs simultaneously. However, Structural Equation Modeling can include latent constructs as well as observed variable into the model whereas the other one can take only observed variables. Its like several regression equations run simultaneously using Amos graphics.
In simultaneous equation modeling You have variables which simultaneously explain one another. For example variable X explains variable Y and variable Y explains variable x.
Generally, structural equation modeling (SEM) is similar to simultaneous equation modeling (SIM). Econometricians and statisticians mostly use SIM than SEM, while marketers, psychologists, sociologists and educationists, among others, use more of SEM than SIM because of, perhaps, their belief in the existence of latent variables.
Yes both are basically the same, except SEM may include latent variable(s) while Simultaneous Equation Model (some use its abbreviation SIM to avoid complication with SEM) does not include latent variable(s).
Both Simultaneous Equation Modelling and Structural Equation Modelling measure relationships between variables. Simultaneous Equation Modelling is used mostly by Econometricians where they try to correct for the bias that results when a variable is behaving as both an endogenous and exogenous variable in a same system of equations. Structural Equation Modelling on the other hand is mostly used by Statisticians and other social Scientist where they measure an immeasurable variable (Latent Variable) by observing other variables. For instance the level of smartness cannot be directly measured but what an individual does that constitute smartness can be measured and attributed to it. Is an umbrella of three areas of study. 1. Regression 2. Path Analysis 3. Factor Analysis.
Suppose I have only one equation. Then what is the difference with OLS? I would imagine some generalization of the variance-covariance structure on the RHS of the equation, but is there a 1-2 sentence explanation that would correctly capture the essence of it?