Package 'testequavar'

Title: Bootstrap Tests for Equality of 2, 3, or 4 Population Variances
Description: Tests the hypothesis that variances are homogeneous or not using bootstrap. The procedure uses a variance-based statistic, and is derived from a normal-theory test. The test equivalently expressed the hypothesis as a function of the log contrasts of the population variances. A box-type acceptance region is constructed to test the hypothesis. See Cahoy (2010) <doi:10.1016/j.csda.2010.04.012>.
Authors: Dexter Cahoy
Maintainer: Dexter Cahoy <[email protected]>
License: GPL (>= 3)
Version: 0.1.3
Built: 2024-10-25 05:47:12 UTC
Source: https://github.com/dcahoy/testequavar

Help Index

Bootstrap test for equality of two (2) population variances

Description

Testing equality of two (2) population variances against the alternative that both variances are not equal.

Usage

equa2vartest(x1, x2, a, B)

Arguments

x1

first sample vector of data or observations
x2

second sample vector of data or observations
a

significance level alpha
B

number of bootstrap samples. At least 500 is recommended.

Value

list consisting of a non-numeric decision whether to reject the null hypothesis or not, the significance level, the number of bootstrap samples used, and the bootstrap P-value calculated using the Euclidean distance.

References

Cahoy, DO (2010), A Bootstrap Test For Equality Of Variances, Computational Statistics & Data Analysis, 54(10), 2306-2316. <doi:10.1016/j.csda.2010.04.012>

Examples

x1=sqrt(10)*runif(8, -sqrt(3), sqrt(3) )
x2=sqrt(1)*runif(8, -sqrt(3), sqrt(3) )
equa2vartest(x1,x2,0.05, 1000)



x1=sqrt(1)*rexp(8)
x2=sqrt(1)*rexp(8)
equa2vartest(x1,x2,0.01, 1000)

Bootstrap test for equality of three (3) population variances

Description

Testing equality of three (3) population variances against the alternative that all variances are unequal.

Usage

equa3vartest(x1, x2, x3, a, B)

Arguments

x1

first sample vector of data or observations
x2

second sample vector of data or observations
x3

third sample vector of data or observations
a

significance level alpha
B

number of bootstrap samples. At least 500 is recommended.

Value

list consisting of a non-numeric decision whether to reject the null hypothesis or not, the significance level, the number of bootstrap samples used, and the bootstrap P-value calculated using the Euclidean distance.

References

Cahoy, DO (2010), A Bootstrap Test For Equality Of Variances, Computational Statistics & Data Analysis, 54(10), 2306-2316. <doi:10.1016/j.csda.2010.04.012>

Examples

x1=sqrt(10)*runif(10, -sqrt(3), sqrt(3) )
x2=sqrt(1)*runif(10, -sqrt(3), sqrt(3) )
x3=sqrt(1)*runif(10, -sqrt(3), sqrt(3) )
equa3vartest(x1,x2,x3, a=0.05, B=1000)


equa3vartest( rexp(10) ,rexp(10) ,rexp(10) ,  a=0.01, B=1000)

Bootstrap test for equality of four (4) population variances

Description

Testing equality of four (4) population variances against the alternative that all variances are not equal.

Usage

equa4vartest(x1, x2, x3, x4, a, B)

Arguments

x1

first sample vector of data or observations
x2

second sample vector of data or observations
x3

third sample vector of data or observations
x4

fourth sample vector of data or observations
a

significance level alpha
B

number of bootstrap samples. At least 500 is recommended.

Value

list consisting of a non-numeric decision whether to reject the null hypothesis or not, the significance level, the number of bootstrap samples used, and the bootstrap P-value calculated using the Euclidean distance.

References

Cahoy, DO (2010), A Bootstrap Test For Equality Of Variances, Computational Statistics & Data Analysis, 54(10), 2306-2316. <doi:10.1016/j.csda.2010.04.012>

Examples

x1=sqrt(10)*runif(10, -sqrt(3), sqrt(3) )
x2=sqrt(1)*runif(10, -sqrt(3), sqrt(3) )
x3=sqrt(1)*runif(10, -sqrt(3), sqrt(3) )
x4=sqrt(1)*runif(10, -sqrt(3), sqrt(3) )
equa4vartest(x1,x2,x3, x4, a=0.05, B=500)



equa4vartest(rexp(10) ,rexp(10) ,rexp(10) , rexp(10),  a=0.01, B=1000)

testequavar Package

Description

Tests the hypothesis that 2, 3, or 4 population variances are homogeneous or not using bootstrap.

Details

Reference:

Cahoy (2010) <doi:10.1016/j.csda.2010.04.012>

Author(s)

Dexter Cahoy [email protected]