sbias: Bias of the sample proportion

In arsenionhacolo/InferenceBEAGSD: Inference for Binary Endpoint Adaptive Group-Sequential Designs

Description

Calculates bias due to using sample proportion as estimator of the true response rate.

Usage

sbias(n1, r1, n, p)

Arguments

n1	Stage 1 sample size.
r1	Stage 1 critical value (trial is stopped at stage 1 if the number of successes is at most r1).
n	Total sample size.
p	True success probability.

Details

For fixed designs the sample propotion is an unbiased (maximum likelihood) estimator of the response rate, but in group sequential designs (e.g., Simon's) it is biased.

Value

Bias.

Author(s)

Arsenio Nhacolo

References

Porcher, R. and Desseaux, K. What inference for two-stage phase II trials? BMC Med Res Methodol, 2012, 12, 117.

See Also

sfms and pg.

Examples

sbias(19, 4, 54, 0.4)

arsenionhacolo/InferenceBEAGSD documentation built on May 9, 2019, 4:10 a.m.