harbin.test: Harbin test (EXPERIMENTAL)

In tpepler/harbin: Harbin quantification for RT-qPCR data

Description

Performs the Harbin test for two groups. (PLEASE NOTE that this method is experimental and still undergoing testing.)

Usage

harbin.test(x, y, reps = 1000)

Arguments

x	First dataset, as a vector.
y	Second dataset, as a vector.
reps	Number of bootstrap replications to use.

Details

The Harbin test is a non-parametric test for two-sample location-scale-shape problem, testing the hypothesis that two datasets originated from population distributions which can be described by the same probability distribution function. The alternative hypothesis is that the location, variability and/or shapes of the two population distributions differ.

The test statistic is calculated from quantiles of the pooled datasets, and compared to a bootstrap distribution of the statistic under the null hypothesis.

Value

Returns a list with the following components:

statistic	Test statistic (the proportion of samples in the first dataset for which the labels have changed when the second dataset was added.
crit.val	95% critical value for the test.
p.value	P-value for the test.

Author(s)

Theo Pepler

See Also

harbin.quant

Examples

# Example 1: Difference in location
data1 <- rnorm(n = 30, mean = 0, sd = 1)
data2 <- rnorm(n = 20, mean = 0.5, sd = 1)
harbin.test(x = data1, y = data2)

# Example 2: Difference in location and variability
data1 <- rnorm(n = 30, mean = 0, sd = 1)
data2 <- rnorm(n = 20, mean = 0.2, sd = 1.5)
harbin.test(x = data1, y = data2)

# Example 3: Difference in location and shape
data1 <- rnorm(n = 30, mean = 0, sd = 1)
data2 <- runif(n = 20, min = -2.5, max = 3.5)
harbin.test(x = data1, y = data2)

tpepler/harbin documentation built on May 31, 2019, 6:47 p.m.