[讨论]Manova in SPSS by Programming

Example

GLM DEP BY A B

/ RANDOM = B

/ DESIGN = A, B, A* B.

In the example, effects B and A* B are considered as random effects. Notice that if only effect B is specified in the RANDOM subcommand, A* B is automatically considered as a random effect.

The hypothesis testing for each effect ( A, B, and A* B) in the design will be carried out using the appropriate error term, which is calculated automatically.

[此贴子已经被作者于2006-5-21 23:02:26编辑过]

Example

GLM OUTCOME BY TREATMNT

/ REGWGT WT.

The procedure performs a weighted least- squares analysis. The variable WT is used as the weight variable.

Example GLM DEP BY A B WITH COV

/ PRINT= DESCRIPTIVE, TEST( LMATRIX), PARAMETER

/ DESIGN.

Since the design in the DESIGN subcommand is not specified, the default design is used. In this case, the design includes the intercept term, the covariate COV, and the full factorial terms of A and B, which are A, B, and A* B.

For each combination of levels of A and B, SPSS displays the descriptive statistics of DEP.

The set of L matrices that generates the sums of squares for testing each effect in the design is displayed.

The parameter estimates, their standard errors, t tests, confidence intervals, and the observed power for each test are displayed.

Example GLM DEP BY A B

/ PLOT = SPREADLEVEL PROFILE( A A* B A* B* C)

/ DESIGN. Assume each of the factors A, B, and C has three levels.

Spread- versus- level plots are produced showing observed cell means versus standard deviations and observed cell means versus variances.

Five profile plots are produced. For factor A, a line plot of estimated means at each level of A is produced ( one plot). For the two- way crossed factor combination A* B, a multipleline plot of estimated means at each level of A, with a separate line for each level of B, is produced ( one plot). For the three- way crossed factor combination A* B* C, a multiple- line plot of estimated means at each level of A, with a separate line for each level of B, is produced for each of the three levels of C ( three plots).

Example

GLM DEP BY A B

/ LMATRIX = “ Effect A” A 1 0 - 1; A 1 - 1 0

/ LMATRIX = “ Effect B” B 1 0 - 1; B 1 - 1 0

/ KMATRIX = 0; 0

/ DESIGN = A B.

In this example, assume that factors A and B each have three levels.

There are two LMATRIX subcommands; both have two rows.

The first LMATRIX subcommand tests whether the effect of A is 0, while the second LMATRIX subcommand tests whether the effect of B is 0.

The KMATRIX subcommand specifies that the K matrix also has two rows, each with value 0.

Example GLM DEP BY FAC

/ CONTRAST( FAC)= DIFFERENCE

/ DESIGN.

Suppose that the factor FAC has five categories and therefore four degrees of freedom.

CONTRAST requests DIFFERENCE contrasts, which compare each level ( except the first) with the mean of the previous levels.

Example

GLM DEP BY A B

/ EMMEANS = TABLES( A* B) COMPARE( A)

/ DESIGN.

The output of this analysis includes a pairwise comparisons table for the dependent variable DEP.

Assume that A has three levels and B has two levels. The first level of A is compared with the second and third levels, the second level with the first and third levels, and the third level with the first and second levels. The pairwise comparison is repeated for the two levels of B.

Example

GLM Y BY A B C WITH X

/ DESIGN A B( A) X* A.

In this example, the design consists of a main effect A, a nested effect B within A, and an interaction effect of a covariate X with a factor A.

MANOVA x1 x2 x3 WITH y1 y2 y3

/DISCRIM ALL ALPHA(1)

/PRINT=SIG(EIGEN DIM).

MANOVA x1 x2 x3 WITH y1 y2 y3

/DISCRIM ALL ALPHA(1)

/PRINT=SIG(EIGEN DIM).