Introduction to boxTest


Introduction

The boxTest package provides a simple workflow for comparing two groups using boxplots and statistical tests. It automatically:

  1. Checks normality via Shapiro–Wilk test.

  2. Applies the appropriate test:

    • Independent 2-sample t-test (if both groups are normal)
    • Mann–Whitney U test (if at least one group is non-normal)
  3. Returns a publication-ready boxplot with optional jittered points.

Example

We will use the built-in mtcars dataset to compare mpg between automatic (am = 0) and manual (am = 1) cars.

# Load package
library(boxTest)

# Compare mpg between automatic and manual cars
res <- compare_two_groups(mtcars, "mpg", "am")

# Display the boxplot
res$plot

Normality Check

The function also returns the Shapiro-Wilk normality test results for each group.

res$normality
#>   group shapiro_W   p_value normal
#> 0     0 0.9767743 0.8987358   TRUE
#> 1     1 0.9458037 0.5362729   TRUE

Test Summary

Finally, the package provides a summary of the statistical test applied, including test statistic, degrees of freedom, and p-value.

res$test_summary
#>                        method statistic df      p_value
#> 1 Independent 2-sample t-test -4.106127 30 0.0002850207

Conclusion

The boxTest package streamlines exploratory analysis and significance testing for two-group comparisons in R. It is particularly useful for beginners and provides a clear, publication-ready output.