Conditional type I and type II error probabilities given current value of the test statistic

Share:

Description

Computes conditional type I and type II error probabilities given current value of the test statistic for monitoring based upon stochastic curtailment. This is now obsolete and included in the functionality of “GrpSeqBnds” and is here for instructional purposes only.

Usage

1
CondPower(Z, frac, drift, drift.end, err.I, sided = 1)

Arguments

Z

Current value of test statistic standardized to unit variance.

frac

Current value of the information fraction (variance fraction).

drift

Current value of the drift, i.e. the expected value of the test statistic normalized to have variance equal to the information fraction. Required if you want to compute conditional type II error, otherwise enter 0.

drift.end

Projected value of the drift at the end of the trial.

err.I

Overall (total) type I error probability

sided

Enter 1 or 2 for sided-ness of the test.

Value

A named numeric vector containing the two components “Pr.cond.typeIerr” and “Pr.cond.typeIIerr”

Author(s)

Grant Izmirlian <izmirlian@nih.gov>

References

A General Theory on Stochastic Curtailment for Censored Survival Data D. Y. Lin, Q. Yao, Zhiliang Ying Journal of the American Statistical Association, Vol. 94, No. 446 (Jun., 1999), pp. 510-521

See Also

GrpSeqBnds

Examples

1
## None as yet

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.