seqdesign
operating.char(x, theta=<<see below>>, how.many=50, range.theta=<<see below>>, power.range=c(.01,.99), upper=!(hypothesis(x)$type=="less"), lower=!(hypothesis(x)$type=="greater"))
"seqdesign"
or a list of group sequential design objects.
theta
is initialized as
how.many
equally spaced values from
range.theta[1]
to
range.theta[2]
.
theta
; only used if
theta
not provided.
theta
; only used if
theta
not provided.
By default,
range.theta=inv.power(x,
power.range,
upper=upper,
lower=lower)
;
i.e.,
theta
spans values which achieve the maximum and minimum power
as specified by
power.range
.
power.range
is used to determine
range.theta
(see above).
Only used if
range.theta
not provided.
TRUE
, then the upper power is used to determine
range.theta
.
TRUE
, then the lower power is used to determine
range.theta
.
x
is a
"seqdesign"
object, then
the function returns an object of class
"operating.char.seqdesign"
.
Otherwise,
if
x
is a list of
"seqdesign"
objects,
then function returns a list of
"operating.char.seqdesign"
objects.
A
"operating.char.seqdesign"
is a matrix with
length(theta)
rows
and the following columns
The C routine
Sseq_OperatingChar
is called.
des <- seqdesign(sample.size=c(60,90,120), P=.5, not.stop="") ## Compute operating characteristics for five values of ## theta spanning the lower and upper alternative hypothesis; ## this is the default behaviour of the summary function theta.min <- min(theta.alt(des)) theta.null <- theta0(des) theta.max <- max(theta.alt(des)) operating.char(des, theta=c(theta.min, mean(theta.min+theta.null), theta.null, mean(theta.max+theta.null), theta.max)