Structural Equation Modeling (SEM) Discussion > simulation in structural equation modeling
Hey Abeer,
In a SEM model, observed variables are usually assumed to be error-free (unfortunately, because we know this is not true). The exception is if your observed variable is part of a latent construct, in which case it does have an error term. If that is the case in your simulation (it is part of a latent measurement), then you should indeed simulate the error term also.
February 15, 2012 |
Jeremy Taylor
Jeremy Taylor
i want to simulate observed variables for structural equation modeling. in real data it is assumed that observed variables are not error free variables. so should i also simulate error term for these variables or simply simulate observed variables??