There are four items and standardized estimates of all items are above .3. The output does not give the values of fit indices.
A saturated model is one in which the number of observed variables is equal to the number of parameters.
SEM tries to reproduce a covariance matrix of observed variables, and the degree to which the SEM model is able to reproduce the covariance matrix of observed variables is what determines how well the model fits the data. (What is actually done is that a discrepancy function between an implied and observed covariance matrix is minimized.) Typically when we do an SEM, we have more observed variables than we do parameters to estimate, because we introduce some constraints on the parameters. This is called over-identification. If you have more observed variables than parameters, then you have positive degrees of freedom. Measures of model fit are based in part on the number of degrees of freedom (and require this to be positive). The saturated model has zero degrees of freedom, which is why measures of model fit cannot be applied. It perfectly reproduces the observed covariance matrix, because it is has zero degrees of freedom.
In short, a saturated model does not tell you anything. Try adding at least one more constraint to the model so that it will be over-identified.
A saturated model is one in which the number of observed variables is equal to the number of parameters.
SEM tries to reproduce a covariance matrix of observed variables, and the degree to which the SEM model is able to reproduce the covariance matrix of observed variables is what determines how well the model fits the data. (What is actually done is that a discrepancy function between an implied and observed covariance matrix is minimized.) Typically when we do an SEM, we have more observed variables than we do parameters to estimate, because we introduce some constraints on the parameters. This is called over-identification. If you have more observed variables than parameters, then you have positive degrees of freedom. Measures of model fit are based in part on the number of degrees of freedom (and require this to be positive). The saturated model has zero degrees of freedom, which is why measures of model fit cannot be applied. It perfectly reproduces the observed covariance matrix, because it is has zero degrees of freedom.
In short, a saturated model does not tell you anything. Try adding at least one more constraint to the model so that it will be over-identified.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages