3 Factorial designs with more than two factors
Jan A. van den Brakel
Previous | Next
The results
developed for factorial designs are extended to designs with
more than two factors. A more general notation for the treatment factors is
introduced first. Let denote
the treatment factor in the experiment with levels
. In the general case there are factors
included in the experiment. The population parameters observed under the treatment combinations are collected in the
vector . The index for the levels of a factor runs
within each level of its preceding factor. Thus index runs
from within
each level of . Hypotheses about the main effects and
interactions are, as motivated in section 2.2, formulated about in
expectation over the measurement error model.
The
contrast matrices for the main effects and interactions in (2.13) are developed
for the general case of a factorial
design. Let denote
the set of labels for the factors and . The following three functions are defined
first;
The main
effect of factor is
defined as the contrasts between the levels,
averaged over the levels of the other factors
and is given by:
Postmultiplication
of by sums
over the levels of the factors that are
nested within each level of .Subsequently, defines
the contrasts between the levels of that are
nested within each combination of the levels of . Premultiplication of by adds the
contrast matrices that are
nested within all combinations of the levels of
The
interaction between and is
defined as the contrasts of factor between
the contrasts of averaged
over the levels of the other factors
and is given by:
Postmultiplication of by adds the
levels of the factors that are
nested within each level of . defines
the contrasts of the main effect of factor which
are nested within each combination of the levels of . Postmultiplication of by sums the
contrast matrices over the
levels of that are
nested within each combination of the levels of . Premultiplication of with defines
the contrasts of the interactions between and , within each combination of the levels of . Finally, premultiplication of by sums the
contrasts of the interactions between and over the
levels of
The interaction between , and is
defined as the contrasts of
factor between
the interactions of and , averaged over the levels of the other factors.
This process expands in a similar way to higher order interactions, which
results in the following definitions of the higher order interactions:
The number
of rows of each contrast matrix coincides with the number of contrasts that
define the various main effects and interactions. The number of columns of
these matrices equals .
These contrast
matrices are inserted in (2.13) to define the various hypotheses about the main
effects and interactions between the G treatment factors. The sampling units
in the initial sample are randomly divided over all possible treatment
combinations according to a CRD or an RBD, resulting in
different subsamples. Let denote
the number of sampling units assigned to treatment combination in subsample and the size
of the initial sample. In the case of a CRD, the first order inclusion
probabilities for the units in subsample are now
given by . In the case of an RBD, the first order
inclusion probabilities for the units in subsample are
given by where denotes
the number of sampling units assigned to treatment combination in block
and the
total number of sampling units in block
Now denotes
the GREG estimator for based on
the observations obtained in subsample and is defined
analogously to expression (2.18). These GREG
estimators are collected in the vector and is
an approximately design-unbiased estimator for and . Design-based estimators for the covariance
matrices of the contrasts between the elements of are
defined by (2.25), where the diagonal elements of are
defined analogously to expression (2.26) in the case of a CRD or (2.27) in the
case of an RBD.
Finally
hypotheses about main effects and interactions are tested with the Wald
statistic (2.28), which is asymptotically distributed as a chi-squared random
variable where the number of degrees of freedom equals the number of contrasts
specified in the various hypotheses. As an example, the contrast matrices of
the main effects and interactions in a factorial design with four factors are
given in Table 3.1.
Table 3.1
Contrasts in a
factorial design
Table summary
This table displays contrasts in an factorial design. The information is grouped by Contrast matrix and Number of contrasts (appearing as column headers).
Contrast matrix |
Number of contrasts (degrees of freedom) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Previous | Next