pwrFDR (version 1.95)

CDF.Pval.HA: CDF of p-values for test statistics distribted under HA.

Description

Computes the CDF of p-values for test statistics distribted under HA.

Usage

CDF.Pval.HA(u, groups = 2, r.1, effect.size, n.sample, control)

Arguments

u

Argument of the CDF. Result will be Pr( P_i <= u )

groups

The number of experimental groups to compare. Default value is 2.

r.1

The proportion of all test statistics that are distributed under HA.

effect.size

The effect size (mean over standard deviation) for test statistics having non-zero means. Assumed to be a constant (in magnitude) over non-zero mean test statistics.

n.sample

The number of experimental replicates.

control

Optionally, a list with components with the following components: 'groups', used when distop=3 (F-dist), specifying number of groups. 'version', used only in the 'JL' method, choice 0 gives the 'JL' version as published, whereas choice 1 replaces the FDR with r.0*FDR resulting in the infinite simultaneous tests limiting average power, which is the 'Iz' version, but this is redundant because you can specify the 'Iz' method to use this option. 'tol' is a convergence criterion used in iterative methods which is set to 1e-8 by default 'max.iter' is an iteration limit, set to 1000 by default 'distop', specifying the distribution family of the central and non-centrally located sub-populations. =1 gives normal (2 groups) =2 gives t- (2 groups) and =3 gives F- (2+ groups) 'CS', correlation structure, for use only with 'method="simulation"' which will simulate m simulatenous tests with correlations 'rho' in blocks of size 'n.WC'. Specify as list CS = list(rho=0.80, n.WC=50) for example

Value

A list with components

call

The call which produced the result

u

The argument that was passed to the function

CDF.Pval.HA

The value of the CDF

Details

Computes the CDF of p-values for test statistics distribted under HA. If Fc_0 is the cCDF of a test statistic under H0 and Fc_A is the cCDF of a test statistic under HA then the CDF of a P-value for a test statistic distributed under HA is

G_A(u) = Fc_A(Fc_0^-1(u))

The limiting true positive fraction is the infinite simultaneous tests average power,

lim_m S_m/M_m = average.power (a.s.),

which is used to approximate the average power for finite 'm', is G_1 at gamma f:

G_1( gamma f) = average.pwer

where f is the FDR and gamma = lim_m J_m/m (a.s.) is the limiting positive call fraction.

References

Genovese, C. and L. Wasserman. (2004) A stochastic process approach to false discovery control. Annals of Statistics. 32 (3), 1035-1061.

Izmirlian G. (2017) Average Power and $\lambda$-power in Multiple Testing Scenarios when the Benjamini-Hochberg False Discovery Rate Procedure is Used. arXiv:1801.03989

See Also

CDF.Pval

Examples

# NOT RUN {
  ## First calculate an average power for a given set of parameters
  rslt.avgp <- pwrFDR(effect.size=0.79, n.sample=46, r.1=2000/54675, FDR=0.15)

  ## Now verify that G_A( gamma f ) = average.power

  gamma <- rslt.avgp$gamma
  f <- rslt.avgp$call$FDR

  GA.gma.f <- CDF.Pval.HA(u=gamma*f, r.1=2000/54675, effect.size=0.79, n.sample=46)

  c(G.A.of.gamma.f=GA.gma.f$CDF.Pval.HA, average.power=rslt.avgp$average.power)
# }