Confirmatory Factor Analysis Intro
Factor Analysis Exploratory Confirmatory Principle components Rotations Confirmatory Split sample Structural equations
Structural Equation Approach Structural equation or covariance structure models
Components Latent variables (endogenous) Manifest variables (exogenous) Residual variables Covariances Influences
Path Diagrams (components) Observed Variable E1 Residual or Error Latent Variable Influence Path Covariance between exogenous variables or errors
Path Diagram for Multiple Regression y = a0 + a1. x1 +a2. x2 + a3 Path Diagram for Multiple Regression y = a0 + a1*x1 +a2*x2 + a3*x3 + a4*x4 + e1 X1 X2 Y E1 X3 X4
Regression All variables are manifest One error term All covariances allowed among independent variables
Two Factor Confirmatory Path Model V1 V2 V3 V4 V5 V6 E1 E1 E1 E1 E1 E1
Confirmatory Model F1 and F2 correlated (oblique) Components of F1 and F2 are separate indicator variables
Example Y = v + e1 X = u + e2 X’ = u + e3 X, Y & X’ are manifest U, V are latent e1, e2, e3 are residual/errors e1, e2, e3 independent with mean = 0 e2, e3, u uncorrelated e1, v uncorrelated
Example Covariance Y X X’ Var(Y)= Var(v) + Var(e1) Cov(XY) = Cov(uv) Var(X) = Var(u) + Var(e2) Cov(X’Y) = Cov(X/X) = Var(u) Var(X’) = Var(u) + Var(e3)
FACTOR Model Specification You can specify the FACTOR statement to compute factor loadings F and unique variances U of an exploratory or confirmatory first-order factor (or component) analysis. By default, the factor correlation matrix P is an identity matrix. C = FF’ + U, U = diag C= data covariance matrix
First-order Confirmatory Factor Analysis For a first-order confirmatory factor analysis, you can use MATRIX statements to define elements in the matrices F, P, and U of the more general model C = FPF' + U, P = P' , U = diag factor loadings F unique variances U factor correlation matrix P data covariance matrix C
PROC FACTOR RESIDUALS / RES displays the residual correlation matrix and the associated partial correlation matrix. The diagonal elements of the residual correlation matrix are the unique variances.