Monte Carlo power for binary two-sample PPI++ test

Monte Carlo power for binary two-sample PPI++ test

Usage

simulate_ppi_ttest_binary(
  R,
  p_A,
  p_B,
  sens_A,
  spec_A,
  n_A,
  n_B,
  N_A,
  N_B,
  sens_B = sens_A,
  spec_B = spec_A,
  alpha = 0.05,
  delta = NULL,
  seed = 1
)

Arguments

R

Number of Monte Carlo replicates.

p_A, p_B

Group-specific prevalences \(P(Y=1)\).

sens_A, spec_A

Sensitivity/specificity for group A predictions.

n_A, n_B

Labeled sample sizes.

N_A, N_B

Unlabeled sample sizes.

sens_B, spec_B

Sensitivity/specificity for group B predictions (defaults match group A).

alpha

Two-sided significance level.

delta

Optional effect size (defaults to p_A - p_B).

seed

RNG seed (set to NULL to avoid seeding).

Value

List with empirical and analytical power plus simulation details.