Create or Generate the Model Implied Correlation or Covariance Matrices
It creates or generates the model implied correlation or covariance matrices based on the RAM model specification.
impliedR(Amatrix, Smatrix, Fmatrix, corr=TRUE, labels, ...)
rimpliedR(Amatrix, Smatrix, Fmatrix, AmatrixSD, k=1, corr=TRUE,
nonPD.pop=c("replace", "nearPD", "accept"))Amatrix |
An asymmetric matrix in the RAM specification with
|
Smatrix |
A symmetric matrix in the RAM specification with
|
Fmatrix |
A filter matrix in the RAM specification with
|
AmatrixSD |
Standard deviations (SD) of the elements in the
|
k |
Number of studies. |
corr |
Logical. The output is either the model implied correlation matrix or the covariance matrix. |
labels |
A character vector of the observed and latent
variables with the same dimensions as that in the |
nonPD.pop |
If it is |
... |
Not used. |
This function can be used to generate the model implied
correlation matrix for the standardized parameters with the
corr=TRUE argument. Suppose we want to calculate the
population correlation matrix for a mediation model with x, m, and
y. We only need to specify the population path coefficients among x,
m, and y in the Amatrix. We do not need to specify the
population error variances of m and y. We treat the error variances as
unknown parameters by giving them starting values in the
Smatrix matrix. When the covariance matrix is requested by
specifying corr=FALSE, it simply calculates the
population model covariance matrix by treating the values in
Smatrix as the population values.
A list of RAM matrices, the model implied correlation or
covariance matrix of the observed variables (SigmaObs), of both
observed and latent variables (SigmaAll), the minimum fit
(minFit) which should be zero, and the status code of the
optimization (status) which should also be zero when the
optimization is fine. The last object is mx.fit which is the output after running the
model. It can be used in the diagnosis.
It is important to ensure that all the population values in
Amatrix must be set as fixed parameters; otherwise, these
values may be altered with the corr=TRUE argument. When
there is an error or warning message about the status code, there is a
high chance that some of the values in Amatrix are incorrectly
set as free parameters.
Mike W.-L. Cheung <mikewlcheung@nus.edu.sg>
set.seed(100)
## A simple mediation model
## All are population parameters in the A matrix
A1 <- matrix(c(0, 0, 0,
0.3, 0, 0,
0.2, 0.4, 0), nrow=3, ncol=3, byrow=TRUE,
dimnames=list(c("x", "m", "y"), c("x", "m", "y")))
A1
## Variance of x is fixed at 1 while the other variances are free.
S1 <- matrix(c(1, 0, 0,
0, "0.1*ErrVarM",0,
0, 0, "0.1*ErrVarY"), nrow=3, ncol=3,
dimnames=list(c("x", "m", "y"), c("x", "m", "y")))
S1
impliedR(Amatrix=A1, Smatrix=S1)
## SD of A1
A1SD <- matrix(c(0, 0, 0,
0.1, 0, 0,
0.1, 0.1, 0), nrow=3, ncol=3, byrow=TRUE,
dimnames=list(c("x", "m", "y"), c("x", "m", "y")))
A1SD
rimpliedR(Amatrix=A1, Smatrix=S1, AmatrixSD=A1SD, k=2)
## A CFA model
A2 <- matrix(c(0, 0, 0, 0.3,
0, 0, 0, 0.4,
0, 0, 0, 0.5,
0, 0, 0, 0), nrow=4, ncol=4, byrow=TRUE,
dimnames=list(c("x1", "x2", "x3", "f"),
c("x1", "x2", "x3", "f")))
A2
## Variance of f is fixed at 1 while the other variances are free.
S2 <- matrix(c("0.1*Err1", 0, 0, 0,
0, "0.1*Err2", 0, 0,
0, 0, "0.1*Err3", 0,
0, 0, 0, 1), nrow=4, ncol=4,
dimnames=list(c("x1", "x2", "x3", "f"), c("x1", "x2", "x3", "f")))
S2
F2 <- create.Fmatrix(c(1,1,1,0), as.mxMatrix=FALSE)
F2
impliedR(Amatrix=A2, Smatrix=S2, Fmatrix=F2)
## SD of A2
A2SD <- matrix(c(0, 0, 0, 0.1,
0, 0, 0, 0.1,
0, 0, 0, 0.1,
0, 0, 0, 0), nrow=4, ncol=4, byrow=TRUE,
dimnames=list(c("x1", "x2", "x3", "f"),
c("x1", "x2", "x3", "f")))
A2SD
rimpliedR(Amatrix=A2, Smatrix=S2, Fmatrix=F2, AmatrixSD=A2SD, k=2)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.