Mann-Whitney U Test with Plain-English Interpretation

Usage

mannwhitney_interpret(
  x,
  y,
  conf.level = 0.95,
  var_name = "Variable",
  context = NULL
)

Arguments

x: A numeric vector (group 1)
y: A numeric vector (group 2)
conf.level: Confidence level. Default 0.95.
var_name: Optional label for the report. Default "Variable"
context: Optional description of the study design or sampling method, echoed back in the printed report. Default NULL.

Value

An object of class statease_mannwhitney containing test results and interpretation. Use print() to display the formatted report.

Examples

x <- c(23, 45, 12, 67, 34, 89, 56)
y <- c(19, 38, 22, 51, 29, 74, 44)
result <- mannwhitney_interpret(x, y)
#> Warning: Sample size is small (n < 10). Interpret results with caution.
print(result)
#> 
#> -- statease Mann-Whitney U Test Report --------------------------
#>   Variable     : Variable
#>   Group 1      : n = 7  |  Median = 45.00
#>   Group 2      : n = 7  |  Median = 38.00
#> -----------------------------------------------------------------
#>   W statistic  : 29.000
#>   p-value      : 0.6200
#>   95% CI      : [-21.000, 38.000]
#>   Effect size  : 0.133 (small)
#> -----------------------------------------------------------------
#>   Assumption Checks:
#>     Sample size (Group 1)   : NOTE     (n = 7)
#>     Sample size (Group 2)   : NOTE     (n = 7)
#>     Test method used        : NOTE     (Wilcoxon rank sum exact test)
#>     Independence            : NOTE     (assumed from study design, not testable from data)
#> 
#>   NOTE: These are contextual notes rather than pass/fail checks.
#>   They describe aspects of the data and test method relevant to
#>   interpretation, but are not automatically verifiable by the
#>   package.
#> -----------------------------------------------------------------
#>   Interpretation:
#>   The result is not statistically significant (p = 0.6200 > alpha 0.05).
#>   Values in Group 1 appear stochastically greater than values in Group 2. (Reported medians: Group 1 = 45.00, Group 2 = 38.00)
#>   Effect size is small (r = 0.133).
#>   Note: Mann-Whitney tests stochastic superiority,
#>   not differences in medians.
#> -----------------------------------------------------------------
#>