Binomial-class {adoptr} | R Documentation |
Binomial data distribution
Description
Implements the normal approximation for a test on rates.
The reponse rate in the control group,
rC, has to be specified by
rate_control
.
The null hypothesis is:
rE ≤ rC,
where rE denotes the response rate
in the invervention group.
It is tested against the alternative
rE > rC.
The test statistic is given as
X1 = √n (rE - rC) / √(2 r0 (1-r0)),
where r0 denotes the mean between
rE and
rC in the two-armed case,
and rE in the one-armed case.#'
All priors have to be defined for the rate difference
rE - rC.
Usage
Binomial(rate_control, two_armed = TRUE)
## S4 method for signature 'Binomial'
quantile(x, probs, n, theta, ...)
## S4 method for signature 'Binomial,numeric'
simulate(object, nsim, n, theta, seed = NULL, ...)
Arguments
rate_control |
assumed response rate in control group |
two_armed |
logical indicating if a two-armed trial is regarded |
x |
outcome |
probs |
vector of probabilities |
n |
sample size |
theta |
distribution parameter |
... |
further optional arguments |
object |
object of class |
nsim |
number of simulation runs |
seed |
random seed |
Details
Note that simulate
for class Binomial
simulates the
normal approximation of the test statistic.
Slots
rate_control
cf. parameter 'rate_control'
See Also
see probability_density_function
and
cumulative_distribution_function
to evaluate the pdf
and the cdf, respectively.
Examples
datadist <- Binomial(rate_control = 0.2, two_armed = FALSE)