diffMeanVar (version 0.0.6)

TrimMeanLeveneTest: Test for equality of variance based on trimmed-mean based Levene test

Description

Test for equality of variance based on trimmed-mean based Levene test.

Usage

TrimMeanLeveneTest(value, group, trim.alpha = 0.25)

Arguments

value

numeric. Measurements to be compared between two groups.

group

numeric. Subject's group membership. Must be binary (i.e., taking values 0 or 1).

trim.alpha

numeric. Indicating proportion of data points to be trimmed from both ends before calculating sample mean.

Value

A list with 2 elements:

stat

test statistic value

pval

pvalue of the score test

References

Brown MB and Forsythe AB (1974) Robust Tests for Equality of Variances. Journal of the American Statistical Association, 69, 364-367.

Examples

# NOT RUN {
    # generate simulated data set from t distribution
    set.seed(1234567)
    es.sim = genSimData.tDistr(nCpGs = 100, nCases = 20, nControls = 20,
      df0 = 10, ncp0 = 0, df1 = 6, ncp1 = 2.393, testPara = "var",
      eps = 1.0e-3, applier = lapply) 
    print(es.sim)
    print(exprs(es.sim)[1:2,1:3])

    # do AW score test for the first probe
    dat = exprs(es.sim)
    pDat = pData(es.sim)
    print(pDat[1:2,])

    res = TrimMeanLeveneTest(value = dat[1,], group = pDat$memSubj)
    print(names(res))
    print(res)
# }