These classes of objects are returned from
fac.aov
.
An object of class
"fac.aov"
represents the fit for a fractional
factorial analysis of variance.
The class
"fac.aov"
has methods for the following generic functions:
pareto
,
plot
,
predict
,
print
,
print.summary
,
qqnorm
,
summary
.
Class
"fac.aov"
inherits from class
"aov"
.
The
fac.aov
object is implemented as a list
with the following components in addition to an
aov
object:
fcoefficients
or half-effects). For unbalanced
or nonorthogonal designs, the
feffects
are calculated from the
fcoefficients
using the value of
adj.coef
to make all of the
standard errors equal. (See below.)
feffects
. However if a
fac.aov
object has been
refitted with
update
, forming a non-saturated model, the saturated
effects are stored here, and the reduced model effects in
feffects
feffects
=
fcoefficients*adj.coef
so that
feffects
have a constant standard
error, a property that is required for the computation methods used
for
pse
,
tse
and
ase
.
aov
.