Help System (web edition)

Repeated measurements design involve one or more treatments on the
same subject. Usually we cannot asumme the observational errors to
be independent of each other. Using multivariate approach we can
avoid the restrictive assumptions of the usual univariate mixed
model. Example, two within subjects factors and one between subjects
(grouping) factor:
      ANOVA TIMM,20
      DEPENDENT=Y11,Y12,Y13,Y21,Y22,Y23,Y31,Y32,Y33
      GROUPING=Group   Group=1,2
      WITHIN=C,B  WLEVELS=3,3
where the values of the variate measured on several conditions
(combination of the two treatments) are given as dependent variables.
The order of a dependent variable defines the treatment combination.
First variable referred in WITHIN and WLEVELS indicates the treatment
which is changed fastests. In the example, the first number in the
name tells the treatment B and the latter the treatment C (which
values change fastest).
The parameter HFORM has the same effect on between factors as
in usual MANOVA model. When an interaction consisting only of one
type of factors is tested, we can either assume that the effects
of within and between factors do not interact (PARALL=1) or
we can do tests without that assumption (PARALL=0). When we testing
an interaction involving both types (e.g. CG in the example) then
all higher order interactions of within factors are assumed to be
zero (BCG in the example). Setting PARALL=1 is the same that we
assume that profiles for each group have the same shape.
At this moment you must have at least one between (grouping) factor,
if you have more than one within factors. The overall parallelism
(in the example H:GC & GB & GCB = 0) may be tested as follows:
       ANOVA TIMM,20
       DEPENDENT=Y11,Y12,Y13,Y21,Y22,Y23,Y31,Y32,Y33
       GROUPING=Group  WITHIN=Combined WLEVELS=9
In multivariate repeated measurements situation several variates are
observed repeatedly over several experimental conditions. E. g.
       ANOVA MANDIBLE,20
       DEPENDENT=SOR1,ANS1,PAL1,SOR2,ANS2,PAL2,SOR3,ANS3,PAL3
       GROUPING=Group WITHIN=Activatr WLEVELS=3
Remark that the values of variates measured on the same treatment
combination must be given in sequence and in the same order for each
treatment combination.

Note: No covariates accepted in repeated measurements yet!
Further information:
  M = Multivariate analysis of (co)variance 
  A = More on ANOVA operations