tost_power: TOST power and sample size calculations.

In dennismalandro/tostpower: TOST Equivalence/Noninferiority Power Calculation

Description

Calculate the power to detect equivalence or non-inferiority for either paired or unpaired experiments.

Usage

tost_power(n = 10, sigma = 1, true_diff = 0, eqv_interval = c(-1, 1),
  alpha = 0.05, paired = FALSE, df = NULL)

Arguments

n	The number in each group for !paired (the default); the number of pairs for paired
sigma	The standard devation: total for !paired, within for paired.
true_diff	The true (unknown) difference of population means.
eqv_interval	The interval within which true_diff is _equivalant_ considered equivalent (as a proportion of the reference mean).
alpha	The significance level of the test.
paired	Is this a paired comparison?
df	Degress of freedom, (usually not set independent of n, but can be, see details).

Value

Power for both equivalence and non-inferiority alternatives

Author(s)

Dennis L. Malandro, [e-mail]

References

Kem F. Phillips (1990) Power of the Two One-Sided Tests Procedure in Bioequivalence, J. Pharmacokinetics and Biopharmaceutics Vol 18, No. 2

Donald J. Schuirmann (1987) A Comparison of the Two One-Sided Tests Procedure and the Power Approach for Assessing the Equivalence of Average Bioavailability, J. Pharmacokinetics and Biopharmaceutics Vol 15, No. 6

See Also

power.t.test

Examples

# Reproduce of n=9 part of power curve in Phillips, Fig. 1 on p. 139

delta_mu <- 0:20

sapply(delta_mu, function(x) {
  tost_power(true_diff = x,
    n = 9, sigma = 10,
    eqv_interval = c(-20, 20), df = 7)[['eq']]
  }
)

dennismalandro/tostpower documentation built on May 15, 2019, 3:25 a.m.