Friedman Rank Sum Test

FRIEDMAN_TEST() tests whether treatment effects differ across k >= 2 related conditions measured on the same block (e.g. repeated measures on the same subject, or a randomized block design). It is the rank-based, distribution-free analogue to a two-way repeated-measures ANOVA without an interaction term.

Usage

FRIEDMAN_TEST(.var_id = NULL, .data = NULL, ...)

Arguments

.var_id

A variable mapper <var_id>. Currently supports x_by_b(), mapping a continuous response, a treatment/grouping variable, and a blocking variable.

.data

A data frame. Only used on the standalone path.

...

Accepted for pipeline consistency. See the Supported variable mapper section — the current implementation takes no additional arguments.

Value

A cld_exec object, or a test_spec when .var_id = NULL. cld_exec@data is a class_friedman_test object.

Details

H0: there is no difference in treatment effects across blocks. H1: at least one treatment differs from another within blocks.

Supported variable mapper <var_id>s

• x_by_b(): blocked Friedman test, single grouping variable only. See details from friedman-xby.

See also

friedman-xby, class_friedman_test

Examples

set.seed(123)
x = rnorm(30)
g = rep(letters[1:3], 10)
b = rep(1:10, each = 3)

# Eager Form
FRIEDMAN_TEST(x_by_b(x, g, b))
#> -- Summary ---------------------------------------------------------------------
#> 
#> ──────────────────────────
#>   statistic  df  p_value  
#> ──────────────────────────
#>     1.400    2    0.497   
#> ──────────────────────────
#> 
#> 

# Piped/Grammar Syntax Form
out =
    define_model(x_by_b(x, g, b)) |>
    prepare_test(FRIEDMAN_TEST) |>     # Or just `prepare()`
    conclude()
#> Error in conclude(prepare_test(define_model(x_by_b(x, g, b)), FRIEDMAN_TEST)): could not find function "conclude"

print(out)
#> Error: object 'out' not found
tidy(out)
#> Error in tidy(out): could not find function "tidy"