!Lines preceded by exclamation marks are comments
MANOVA with 3 DVs, 1 IV,
NGroups=1
Data NInput=4 NObservation=50
!Ninput represents the number of variables.
!Nobservation the number of variables.
!Cmatrix gives the population covariance matrix. All variances are given as 1, to
!simplify the calculations. The correlations between the y-variables are all equal to
!0.2. The effect of the x-variable on each of the y's is 0.5.
CMatrix
1.0
0.5 1.0
0.5 0.2 1.0
0.5 0.2 0.2 1.0
Labels x y1 y2 y3
!Begin matrices is used to tell Mx the matrices that we use for the analysis. RAM
!specification of SEM uses 4 matrices, labelled as S, A, F and I.
!The first part of the command gives the matrix name, the next part is its form (Symm
!is a symmetrical matrix, Full is a full matrix, and Iden is an identity matrix.
!Finally, the rows and columns of the matrix are given.
Begin Matrices;
S Symm 4 4
A Full 4 4
F Full 4 4
I Iden 4 4
End Matrices;
!Matrix S represents the correlations between the variables. The correlations between
!the x variable and each of the y-variables is fixed to zero (necessarily for !identification).
!The correlations between the y-variables are set to 0.2 - their population values.
Matrix S
1.0
0.0 1.0
0.0 0.2 1.0
0.0 0.2 0.2 1.0
!In this section, we specify the parameters to be estimated by providing them with a
!number. Parameters which are fixed to zero are not estimated, parameters with !unique numbers are estimated separately, parameters with the same number are
!constrained to equality.
!The variances of each of the variables is estimated, and the correlations between the
!y-variables are estimated. The correlations between each of the y-variables and the
!x-variable are not estimated, but remain at the values that were given under Matrix S
!(in this case, 0).
Specify S
1
0 2
0 3 4
0 5 6 7
Label Row S
X Y1 Y2 Y3
Label Col S
X Y1 Y2 Y3
!matrix A contains regression estimates
!As specified below, starting values for y2 and y3 are set to
!population values. Value for regression estimate for y1 is fixed
!to zero. This is a univariate test. For multivariate, see
!below.
Matrix A
0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0
0.5 0.0 0.0 0.0
0.5 0.0 0.0 0.0
!Again, unique numbers provide unique estimates.
Specify A
0 0 0 0
0 0 0 0
8 0 0 0
9 0 0 0
!Specification given below for A provides multivariate test
!Matrix A
! 0 0 0 0
! 0 0 0 0
! 0 0 0 0
! 0 0 0 0
!Specify A
! 0 0 0 0
! 0 0 0 0
! 0 0 0 0
! 0 0 0 0
Label Row A
X Y1 Y2 Y3
Label Col A
X Y1 Y2 Y3
Matrix F
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
Label Row F
X Y1 Y2 Y3
Label Col F
X Y1 Y2 Y3
Covariance F & ((I - A)~ & S);
Options Rsidual
! Power option is power=alpha, df. Alpha is usually set to 0.05, by convention.
!df is equal to 1, because we are testing one parameter (a univariate regression) for
!significance. For the multivariate test, df would be equal to 3.
Options power=0.05, 1
End Group;