Ansari-Bradley Test with Arbitrarily Missing Data.

abwm

This package includes one function abwm.test(), which performs the Ansari-Bradley test (Ansari & Bradley, 1960) with missing data proposed in (Zeng et al., 2025). This method is a theoretical extension of the Ansari-Bradley test, with controlled Type I error regardless of values of missing data.

Bounds of the Ansari-Bradley test statistic and its p-value will be computed in the presence of missing data. The p-value of the test method proposed in (Zeng et al., 2025) is then returned as the maximum possible p-value of the Ansari-Bradley test.

Installation

You can install the development version of abwm from GitHub with:

# install.packages("devtools")
devtools::install_github("Yijin-Zeng/Ansari-Bradley-Test-with-Missing-data")

Example

This is a basic example which shows you how to perform the test with missing data:

library(abwm)

#### Two-sided test
X <- c(6.2, 3.5, NA, 7.6, 9.2)
Y <- c(NA, 1.3, -0.5, -1.7)
abwm.test(X, Y, alternative = 'two.sided')
#> $p.value
#> [1] 1
#> 
#> $bounds.statistic
#> [1]  7 14
#> 
#> $bounds.pvalue
#> [1] 0.04795881 1.00000000
#> 
#> $alternative
#> [1] "two.sided"
#> 
#> $data.name
#> [1] "X and Y"

## One-sided test
X <- c(6.2, 3.5, NA, 7.6, 9.2)
Y <- c(NA, 1.3, -0.5, -1.7)
abwm.test(X, Y, alternative = 'less')
#> $p.value
#> [1] 0.9176989
#> 
#> $bounds.statistic
#> [1]  7 14
#> 
#> $bounds.pvalue
#> [1] 0.02397941 0.91769888
#> 
#> $alternative
#> [1] "less"
#> 
#> $data.name
#> [1] "X and Y"

See Also

The R function stats::ansari.test() performs the Ansari-Bradley test when data are completely observed.

References

Y. Zeng, N. M. Adams, D. A. Bodenham. Scale two-sample testing with arbitrarily missing data. arXiv preprint arXiv:2509.20332. 2025.

A. R. Ansari, R. A. Bradley. Rank-Sum Tests for Dispersions. The Annals of Mathematical Statistics, pages 1174–1189, 1960.