Skip to contents

Frequentist Test from Parast et al. (2024)

Source: R/frequentist_test.R
frequentist_test.Rd

This function performs a frequentist test for surrogate evaluation based on the rank-based methodology of Parast et al. (2024) . It calculates confidence intervals, the \(\varepsilon\) threshold used, and determines if the surrogate is valid. This function is generally not intended to be called directly by the user and is instead used internally within BSET_no_X and BSET_X.

Usage

frequentist_test(P_observed, Z, alpha = 0.05, beta = 0.2, delta_true = NULL)

Arguments

P_observed

Observed outcomes \((Y, S)\) corresponding to the assigned treatment \(Z\).

Z

Treatment assignment vector.

alpha

Numeric. Significance level for the confidence interval (default is 0.05).

beta

Numeric. Type II error rate (default is 0.2).

delta_true

Numeric (optional). The true value of \(\delta\), used to calculate frequentist coverage during simulations.

Value

A list containing:

  • CI: The calculated confidence interval for \(\delta\).

  • threshold: The \(\varepsilon\) threshold value used in the test.

  • coverage: Logical indicating if delta_true falls within the CI (if delta_true is provided).

  • power: Logical indicating if the upper bound of CI is below threshold, which indicates that the test identifies the surrogate as valid.

References

Parast L, Cai T, Tian L (2024). “A rank-based approach to evaluate a surrogate marker in a small sample setting.” Biometrics, 80(1), ujad035.