The boxTest package provides a simple way to compare two groups using boxplots and statistical tests. It automatically checks normality (via Shapiro–Wilk test) and then applies the appropriate test:

• Independent 2-sample t-test (if both groups are normally distributed)
• Mann–Whitney U test (if at least one group is non-normal)

It also generates a publication-ready boxplot with jittered points.

Installation

From GitHub (development version)

# install.packages("remotes")  # if not already installed
remotes::install_github("arka1985/boxTest")

From CRAN (once accepted)

install.packages("boxTest")

From tar.gz release

Download the .tar.gz file from Releases and install locally:

install.packages("path/to/boxTest_0.1.0.tar.gz", repos = NULL, type = "source")

Example

library(boxTest)
# Create a sample dataset: Male and Female BMI
df <- data.frame(
  gender = rep(c("Male", "Female"), each = 10),
  BMI = c(22, 24, 25, 23, 26, 28, 24, 23, 27, 25,
          21, 22, 23, 20, 24, 22, 21, 23, 22, 21)
)
# Compare BMI between Male and Female
res <- compare_two_groups(df, "BMI", "gender")
# Show the boxplot
print(res$plot)
# Normality test results
print(res$normality)
# Test summary (t-test or Mann–Whitney U)
print(res$test_summary)

Output

• res$plot → A ggplot object showing the boxplot with jittered points
• res$normality → Shapiro–Wilk test results for each group
• res$test_summary → Test used, statistic, df (if applicable), and p-value

Authors

• Arkaprabha Sau (Author & Maintainer)
• Santanu Phadikar (Author)
• Ishita Bhakta (Author)