Perform normality test on a variable

Conducts normality tests for a specified variable, optionally by group. Supports automatic testing and interactive visualization.

normal_test(data = NULL, var = NULL, group = NULL, norm = "auto")

Arguments

data

A data frame containing the variables to be tested.

var

A character string specifying the numeric variable in data to test.

group

A character string specifying the grouping variable in data. If NULL, treated as one group.

norm

Control parameter for test behavior. Accepts:

  • 'auto': Automatically decide based on p-values, but the same as 'ask' when n > 1000, default

  • 'ask': Show p-values, plots QQ plots and prompts for decision

  • TRUE/'true': Always returns TRUE

  • FALSE/'false': Always returns FALSE

Value

A logical value:

  • TRUE: data are normally distributed

  • FALSE: data are not normally distributed

Methodology for p-values

Automatically selects test based on sample size per group:

  • n < 3: Too small, assuming non-normal

  • 3 ≤ n ≤ 50: Shapiro-Wilk test

  • 50 < n<= 1000: D'Agostino Chi2 test, instead of Kolmogorov-Smirnov test

  • n > 1000: Show p-values, plots QQ plots and prompts for decision references: 1.https://www.graphpad.com/guides/prism/latest/statistics/stat_choosing_a_normality_test.htm 2.D’Agostino, R.B. (1971) An Omnibus Test of Normality for Moderate and Large Size Samples. Biometrika, 58, 341-348.

Examples

normal_test(iris, "Sepal.Length", "Species", norm = "auto")
#> [1] TRUE
normal_test(iris, "Sepal.Length", "Species", norm = TRUE)
#> [1] TRUE