oa.design(levels, factor.names, min.resid.df=<<see below>>)
0
,
unless the number of levels in the factors are all equal
in which case the default is
3
.
"design"
which is a design for the factors specified.
Three attributes are special to orthogonal array designs:
generating.oa gives the name of the object
that contains the complete orthogonal array design
from which the result was generated;
selected.columns says which columns of this object
were used to produce the result;
residual.df gives the number of residual degrees of freedom
in the design,
when only the main effects are fitted
(you may want to check this value to see
how many more residual degrees of freedom than are needed
in your application are available).
oa.design
may not be able to find a design
as requested.
f so, an error stop is made.
The design is generated by selecting some of the columns
from one of a stored catalog of orthogonal array designs.
An error occurs if
oa.design
can not
find a design as specified.
If factor names are not given, they default to
std.factor.names(length(levels))
,
that is,
"A"
,
"B"
, etc.
If a factor's levels are not named, the levels are set to
the factor name (possibly abbreviated) followed by level numbers.
Orthogonal Array designs are (typically) highly fractional designs that are used with main-effects-only models.
# Produce an 18 run design with 6 degrees of freedom # for error assuming only main effects are fit: oa <- oa.design(c(2, 3, 3, 3, 3, 3))