Skip to contents

Evaluates exact joint power of the two one-sided tests (TOST) by integrating the conditional normal acceptance probability over the independent chi-square variance estimate (Phillips 1990). Uses balanced even totals and reports an error if the target power cannot be reached within 10000 participants. Matches sampleN.TOST in the PowerTOST package exactly.

Usage

calculate_bioequivalence(
  cv,
  alpha,
  power,
  theta0 = 1,
  margin_low = 0.8,
  margin_high = 1.25
)

Arguments

cv

Intra-subject coefficient of variation

alpha

Type I error (one-sided per test, default 0.05)

power

Target power (default 0.80)

theta0

Anticipated T/R geometric mean ratio (default 1.0)

margin_low

Lower equivalence margin (default 0.80)

margin_high

Upper equivalence margin (default 1.25)

Value

List with sample size results consumed by as_sample_size_result()