powsign {EQUIVNONINF}

Nonconditional power of the UMPU sign test for equivalence and its nonrandomized counterpart

Description

The program computes for each possible value of the number n_0 of zero observations the power conditional on N_0 = n_0 and averages these conditional power values with respect to the distribution of N_0. Equivalence is defined in terms of the logarithm of the ratio p_+/p_-, where p_+ and p_- denotes the probability of obtaining a positive and negative sign, respectively.

Usage

 powsign(alpha,n,eps1,eps2,poa)

Arguments

alpha

significance level

n

sample size

eps1

absolute value of the lower limit of the hypothetical equivalence range for \log(p_+/p_-).

eps2

upper limit of the hypothetical equivalence range for \log(p_+/p_-).

poa

probability of a tie under the alternative of interest

Value

alpha

significance level

n

sample size

eps1

absolute value of the lower limit of the hypothetical equivalence range for \log(p_+/p_-).

eps2

upper limit of the hypothetical equivalence range for \log(p_+/p_-).

poa

probability of a tie under the alternative of interest

POWNONRD

power of the nonrandomized version of the test against the alternative p_+ = p_- = (1-p_\circ)/2

POW

power of the randomized UMPU test against the alternative p_+ = p_- = (1-p_\circ)/2

Note

A special case of the test whose power is computed by this program, is the exact conditional equivalence test for the McNemar setting (cf. Wellek 2010, pp. 76-77).

Author(s)

Stefan Wellek <[e-mail]>
Peter Ziegler <[e-mail]>

References

Wellek S: Testing statistical hypotheses of equivalence and noninferiority. Second edition. Boca Raton: Chapman & Hall/CRC Press, 2010, \S 5.1.

Examples

    
powsign(0.06580,50,0.847298,0.847298,0.26)