Structural Equation Model Fit Report
Each time you click Run in the Model Specification report, a Structural Equation Model report for the specified model appears. By default, this report contains a Summary of Fit report, a Parameter Estimates report, and a Path Diagram. If the MIIV Two-Stage Least Squares estimation method is selected in the launch window, an Equation Details report is also included in the Structural Equation Model report.
Note: When you specify a Groups variable in the launch window, there is a separate model fit report for each level of the Groups variable. You can navigate between these reports using the group tabs above the Structural Equation Model fit report outline box.
Summary of Fit
Table of information about the model fit, including the convergence status and the estimation method. When you specify a Groups variable in the launch window, this table contains a second column for statistics that refer only to the level of the Groups variable in the current model fit tab. The following statistics are reported in this table:
Sample Size
The number of observations (rows) used to fit the model.
Rows with Missing
The number of observations (rows) that contained at least one missing value. All missing values are handled using full information maximum likelihood (Finkbeiner 1979).
Total Number of Estimated Parameters
(Available only when the MIIV Two-Stage Least Squares estimation method is selected.) The total number of freely estimated parameters in the model.
MIIV-2SLS Parameters
(Available only when the MIIV Two-Stage Least Squares estimation method is selected.) The number of freely estimated parameters in the model that are estimated by the MIIV Two-Stage Least Squares estimation method.
Maximum Likelihood Parameters
(Available only when the MIIV Two-Stage Least Squares estimation method is selected.) The number of freely estimated parameters in the model that are estimated by the Maximum Likelihood estimation method.
Number of Equations
(Available only when the MIIV Two-Stage Least Squares estimation method is selected.) The number of equations that are estimated in the model after a latent-to-observed transformation has been conducted. Each equation represents an outcome variable that is predicted by one or more specified predictors using model-implied instrumental variables. For more information about the latent-to-observed transformation, see Bollen (1996).
-2 Log Likelihood
The log-likelihood of the fitted model multiplied by -2. This value can be used to compare nested models; the difference between two models’ -2 Log Likelihood values is chi-square distributed with degrees of freedom equal to the difference of degrees of freedom between the models. See “Likelihood, AICc, and BIC” in Fitting Linear Models.
Iterations
The number of iterations used to fit the model.
Number of Parameters
The number of freely estimated parameters in the model.
AICc
The corrected Akaike information criterion. This value can be used to compare models, where a smaller number indicates a better model fit. See AICc, BIC, and BICu.
BICu
The BIC relative to the unrestricted model (BICu) is a reformulation of the Bayesian information criterion. The BICu is defined as a comparison to the unrestricted model. A negative value supports the fitted model, and a positive value supports the unrestricted model. Similar to other information criteria, this value can be used to compare models, where a smaller number between two models indicates the better fitting model. See AICc, BIC, and BICu.
ChiSquare
The chi-square statistic for the model.
DF
The degrees of freedom for the chi-square test for model fit.
Prob>ChiSq
The p-value of the chi-square statistic for the model.
CFI
The Bentler’s comparative fit index (CFI) provides additional guidance for determining model fit. The CFI is bounded between 0 and 1. Values greater than 0.90 are preferred (Browne and Cudeck 1993; Hu and Bentler 1999). See CFI.
RMSEA
The root mean square error of approximation (RMSEA) provides additional guidance for determining model fit. The RMSEA is bounded between 0 and 1. Values less than 0.10 are preferred (Browne and Cudeck 1993; Hu and Bentler 1999). See RMSEA.
Lower 90%
The 90% lower confidence limit for the RMSEA. See RMSEA.
Upper 90%
The 90% upper confidence limit for the RMSEA. See RMSEA.
Equation Details
(Available only when the MIIV Two-Stage Least Squares estimation method is selected.) Tables of key details for each equation in the model, as well as the model-implied instruments that were used for estimation. The first table contains the outcome variable (Outcome), the set of specified predictor variables (Predictors), and the related model fit statistics for each equation. The fit statistics include the Sargan test statistic (Sargan), degrees of freedom (DF), and corresponding p-value (Prob>ChiSq), which help assess potential misspecification at the equation level. When the Variance of the Error and Composite Error options are selected from the Equation Details red triangle menu, the table also shows the estimated residual variance (Var(Error)) and the composite error term (Composite Error) for each equation. The second table contains each outcome variable along with the set of model-implied instrumental variables (Instruments) that were used to estimate its predictors. The Equation Details red triangle menu also includes the option to Show All Equations, which is turned off by default. This option applies to both tables.
Parameter Estimates
Table of estimates for the model parameters. The table is organized in sections for Means/Intercepts, Loadings, Regressions, and Variances. For each estimate, a standard error (Std Error), Wald test statistic (Wald Z), and a corresponding p-value (Prob>|Z|) are given. When you specify a Groups variable in the launch window, this table contains parameter estimates only for the level of the Groups variable in the current model fit tab.
Tip: The Parameter Estimates table contains additional hidden columns. To show these columns, right-click the table and select the additional columns from the Columns submenu.
Path Diagram
Shows the path diagram representation of the fitted model. See Diagram Tab. When you specify a Groups variable in the launch window, this diagram represents only the level of the Groups variable in the current model fit tab.