# permTestPaired: Permutation test for paired data. In CarletonStats: Functions for Statistics Classes at Carleton College

## Description

Permutation test for paired data.

## Usage

  1 2 3 4 5 6 7 8 9 10 11 permTestPaired(x, ...) ## Default S3 method: permTestPaired(x, y, B = 9999, alternative = "two.sided", plot.hist = TRUE, legend.loc = "topright", plot.qq = FALSE, x.name = deparse(substitute(x)), y.name = deparse(substitute(y)), ...) ## S3 method for class 'formula' permTestPaired(formula, data, subset, ...) 

## Arguments

 x a numeric vector. ... further arguments to be passed to or from methods. y a numeric vector. B the number of resamples. alternative the alternative hypothesis. Options are "two.sided", "less" and "greater". plot.hist a logical value. If TRUE, create a histogram displaying the permutation distribution of the statistic. legend.loc a logical value. If TRUE, include a legend on the histogram. plot.qq a logical value. If TRUE, include a quantile-normal plot of the permuation distribution. x.name Label for x variable y.name Label for y variable formula a formula of the form y ~ x, where x, y are both numeric variables. data an optional data frame containing the variables in the formula. By default the variables are taken from environment(formula). subset an optional vector specifying a subset of observations to be used.

## Details

For two paired numeric variables with n rows, randomly select k of the n rows (k also is randm) and switch the entries (a, b) \to (b, a) and then compute the mean of the difference of the two variables (y-x).

Observations with missing values are removed.

## Value

Returns invisibly a vector of the replicates of the test statistic (ex. mean of the difference of the resampled variables).

## Methods (by class)

• default: Permutation test for paired data.

• formula: Permutation test for paired data.

Laura Chihara

## References

Tim Hesterberg's website: http://www.timhesterberg.net/bootstrap

## Examples

 1 2 3 4 5 6 #Does chocolate ice cream have more calories than vanilla ice cream, on average? #H0: mean number of calories is the same #HA: mean number of calories is greater in chocolate ice cream permTestPaired(Icecream$VanillaCalories, Icecream$ChocCalories, alternative = "less") permTestPaired(ChocCalories ~ VanillaCalories, data = Icecream, alternative = "greater") 

CarletonStats documentation built on Aug. 10, 2018, 1:14 a.m.