design
(typically) represent data from a designed
experiment.
These objects are generated by the following functions:
design
,
fac.design
,
oa.design
.
Generic functions that have methods for designs include:
alias
,
plot
,
print
.
Objects of class
"design"
inherit from
"data.frame"
.
Objects of class
"design"
are just like data frames -- each one is
a list all of whose components have the same length, and have attributes
row.names
and
names
.
Designs need have no attributes other than those for data frames.
The assumption is that a
"design"
object starts life with all of its
variables being factors;
then one or more quantitative variables are added.
Unless told otherwise, methods for designs tend to assume that the
first quantitative variable found in the design should be the response.
They may have special information indicating that they were produced
as fractional factorial designs or as orthogonal array designs
(see
fac.design
or
oa.design
).
The important function
factor.names
returns or sets the names
for the factors and for their levels.
It does not use an explicit attribute of this name, however.
Instead, it uses the names of the design object and the
levels of each factor variable.