Add Anova Test P-values to a GGPlot

Source: R/stat_anova_test.R

Adds automatically one-way and two-way ANOVA test p-values to a ggplot, such as box blots, dot plots and stripcharts.

stat_anova_test(
  mapping = NULL,
  data = NULL,
  method = c("one_way", "one_way_repeated", "two_way", "two_way_repeated",
    "two_way_mixed"),
  wid = NULL,
  group.by = NULL,
  type = NULL,
  effect.size = "ges",
  error = NULL,
  correction = c("auto", "GG", "HF", "none"),
  label = "{method}, p = {p.format}",
  label.x.npc = "left",
  label.y.npc = "top",
  label.x = NULL,
  label.y = NULL,
  step.increase = 0.1,
  p.adjust.method = "holm",
  significance = list(),
  geom = "text",
  position = "identity",
  na.rm = FALSE,
  show.legend = FALSE,
  inherit.aes = TRUE,
  parse = FALSE,
  ...
)

Arguments

mapping

Set of aesthetic mappings created by aes(). If specified and inherit.aes = TRUE (the default), it is combined with the default mapping at the top level of the plot. You must supply mapping if there is no plot mapping.

data

The data to be displayed in this layer. There are three options:

If NULL, the default, the data is inherited from the plot data as specified in the call to ggplot().

A data.frame, or other object, will override the plot data. All objects will be fortified to produce a data frame. See fortify() for which variables will be created.

A function will be called with a single argument, the plot data. The return value must be a data.frame, and will be used as the layer data. A function can be created from a formula (e.g. ~ head(.x, 10)).

method

ANOVA test methods. Possible values are one of c("one_way", "one_way_repeated", "two_way", "two_way_repeated", "two_way_mixed").

wid

(factor) column name containing individuals/subjects identifier. Should be unique per individual. Required only for repeated measure tests ("one_way_repeated", "two_way_repeated", "friedman_test", etc).

group.by

(optional) character vector specifying the grouping variable; it should be used only for grouped plots. Possible values are :

  • "x.var": Group by the x-axis variable and perform the test between legend groups. In other words, the p-value is compute between legend groups at each x position

  • "legend.var": Group by the legend variable and perform the test between x-axis groups. In other words, the test is performed between the x-groups for each legend level.

type

the type of sums of squares for ANOVA. Allowed values are either 1, 2 or 3. type = 2 is the default because this will yield identical ANOVA results as type = 1 when data are balanced but type = 2 will additionally yield various assumption tests where appropriate. When the data are unbalanced the type = 3 is used by popular commercial softwares including SPSS.

effect.size

the effect size to compute and to show in the ANOVA results. Allowed values can be either "ges" (generalized eta squared) or "pes" (partial eta squared) or both. Default is "ges".

error

(optional) for a linear model, an lm model object from which the overall error sum of squares and degrees of freedom are to be calculated. Read more in Anova() documentation.

correction

character. Used only in repeated measures ANOVA test to specify which correction of the degrees of freedom should be reported for the within-subject factors. Possible values are:

  • "GG": applies Greenhouse-Geisser correction to all within-subjects factors even if the assumption of sphericity is met (i.e., Mauchly's test is not significant, p > 0.05).

  • "HF": applies Hyunh-Feldt correction to all within-subjects factors even if the assumption of sphericity is met,

  • "none": returns the ANOVA table without any correction and

  • "auto": apply automatically GG correction to only within-subjects factors violating the sphericity assumption (i.e., Mauchly's test p-value is significant, p <= 0.05).

label

character string specifying label. Can be:

  • the column containing the label (e.g.: label = "p" or label = "p.adj"), where p is the p-value. Other possible values are "p.signif", "p.adj.signif", "p.format", "p.adj.format".

  • an expression that can be formatted by the glue() package. For example, when specifying label = "Anova, p = \{p\}", the expression {p} will be replaced by its value.

  • a combination of plotmath expressions and glue expressions. You may want some of the statistical parameter in italic; for example:label = "Anova, italic(p) = {p}".

  • a constant: label = "as_italic": display statistical parameters in italic; label = "as_detailed": detailed plain text; label = "as_detailed_expression" or label = "as_detailed_italic": detailed plotmath expression. Statistical parameters will be displayed in italic.

.

label.x.npc, label.y.npc

can be numeric or character vector of the same length as the number of groups and/or panels. If too short they will be recycled.

  • If numeric, value should be between 0 and 1. Coordinates to be used for positioning the label, expressed in "normalized parent coordinates".

  • If character, allowed values include: i) one of c('right', 'left', 'center', 'centre', 'middle') for x-axis; ii) and one of c( 'bottom', 'top', 'center', 'centre', 'middle') for y-axis.

label.x, label.y

numeric Coordinates (in data units) to be used for absolute positioning of the label. If too short they will be recycled.

step.increase

numeric value in with the increase in fraction of total height for every additional comparison to minimize overlap. The step value can be negative to reverse the order of groups.

p.adjust.method

method for adjusting p values (see p.adjust). Has impact only in a situation, where multiple pairwise tests are performed; or when there are multiple grouping variables. Allowed values include "holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none". If you don't want to adjust the p value (not recommended), use p.adjust.method = "none".

significance

a list of arguments specifying the signifcance cutpoints and symbols. For example, significance <- list(cutpoints = c(0, 0.0001, 0.001, 0.01, 0.05, Inf), symbols = c("****", "***", "**", "*", "ns")).

In other words, we use the following convention for symbols indicating statistical significance:

  • ns: p > 0.05

  • *: p <= 0.05

  • **: p <= 0.01

  • ***: p <= 0.001

  • ****: p <= 0.0001

geom

The geometric object to use to display the data, either as a ggproto Geom subclass or as a string naming the geom stripped of the geom_ prefix (e.g. "point" rather than "geom_point")

position

Position adjustment, either as a string naming the adjustment (e.g. "jitter" to use position_jitter), or the result of a call to a position adjustment function. Use the latter if you need to change the settings of the adjustment.

na.rm

If FALSE (the default), removes missing values with a warning. If TRUE silently removes missing values.

show.legend

logical. Should this layer be included in the legends? NA, the default, includes if any aesthetics are mapped. FALSE never includes, and TRUE always includes. It can also be a named logical vector to finely select the aesthetics to display.

inherit.aes

If FALSE, overrides the default aesthetics, rather than combining with them. This is most useful for helper functions that define both data and aesthetics and shouldn't inherit behaviour from the default plot specification, e.g. borders().

parse

If TRUE, the labels will be parsed into expressions and displayed as described in ?plotmath.

...

other arguments to pass to geom_text, such as:

  • hjust: horizontal justification of the text. Move the text left or right and

  • vjust: vertical justification of the text. Move the text up or down.

Computed variables

  • DFn: Degrees of Freedom in the numerator (i.e. DF effect).

  • DFd: Degrees of Freedom in the denominator (i.e., DF error).

  • ges: Generalized Eta-Squared measure of effect size. Computed only when the option effect.size = "ges".

  • pes: Partial Eta-Squared measure of effect size. Computed only when the option effect.size = "pes".

  • F: F-value.

  • p: p-value.

  • p.adj: Adjusted p-values.

  • p.signif: P-value significance.

  • p.adj.signif: Adjusted p-value significance.

  • p.format: Formated p-value.

  • p.adj.format: Formated adjusted p-value.

  • n: number of samples.

Examples

# Data preparation
#%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
# Transform `dose` into factor variable
df <- ToothGrowth
df$dose <- as.factor(df$dose)
# Add individuals id
df$id <- rep(1:10, 6)
# Add a random grouping variable
set.seed(123)
df$group <- sample(factor(rep(c("grp1", "grp2", "grp3"), 20)))
df$len <- ifelse(df$group == "grp2", df$len+2, df$len)
df$len <- ifelse(df$group == "grp3", df$len+7, df$len)
head(df, 3)
#>    len supp dose id group
#> 1  4.2   VC  0.5  1  grp1
#> 2 18.5   VC  0.5  2  grp3
#> 3 14.3   VC  0.5  3  grp3


# Basic boxplot
#%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
# Create a basic boxplot
# Add 5% and 10% space to the plot bottom and the top, respectively
bxp <- ggboxplot(df, x = "dose", y = "len") +
  scale_y_continuous(expand = expansion(mult = c(0.05, 0.1)))

# Add the p-value to the boxplot
bxp + stat_anova_test()


if (FALSE) {
# Change the label position
# Using coordinates in data units
bxp + stat_anova_test(label.x = "1", label.y = 10, hjust = 0)
}

# Format the p-value differently
custom_p_format <- function(p) {
  rstatix::p_format(p, accuracy = 0.0001, digits = 3, leading.zero = FALSE)
}
bxp + stat_anova_test(
  label = "Anova, italic(p) = {custom_p_format(p)}{p.signif}"
)
#> Warning: Computation failed in `stat_compare_multiple_means()`
#> Caused by error in `mutate()`:
#> ! Problem while computing `label = glue(label)`.
#> Caused by error in `custom_p_format()`:
#> ! could not find function "custom_p_format"


# Show a detailed label in italic
bxp + stat_anova_test(label = "as_detailed_italic")



# Faceted plots
#%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
# Create a ggplot facet
bxp <- ggboxplot(df, x = "dose", y = "len", facet.by = "supp") +
  scale_y_continuous(expand = expansion(mult = c(0.05, 0.1)))
# Add p-values
bxp + stat_anova_test()



# Grouped plots
#%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
bxp2 <- ggboxplot(df, x = "group", y = "len", color = "dose", palette = "npg")

# For each x-position, computes tests between legend groups
bxp2 + stat_anova_test(aes(group = dose), label = "p = {p.format}{p.signif}")


#  For each legend group, computes tests between x variable groups
bxp2 + stat_anova_test(aes(group = dose, color = dose), group.by = "legend.var")



if (FALSE) {
# Two-way ANOVA: Independent measures
#%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
# Visualization: box plots with p-values
# Two-way interaction p-values between x and legend (group) variables
bxp3 <- ggboxplot(
  df, x = "supp", y = "len",
  color = "dose", palette = "jco"
)
bxp3 + stat_anova_test(aes(group = dose),  method = "two_way")

# One-way repeatead measures ANOVA
#%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
df$id <- as.factor(c(rep(1:10, 3), rep(11:20, 3)))
ggboxplot(df, x = "dose", y = "len") +
  stat_anova_test(method = "one_way_repeated", wid = "id")

# Two-way repeatead measures ANOVA
#%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
df$id <- as.factor(rep(1:10, 6))
ggboxplot(df, x = "dose", y = "len", color = "supp", palette = "jco") +
  stat_anova_test(aes(group = supp), method = "two_way_repeated", wid = "id")

# Grouped one-way repeated measures ANOVA
ggboxplot(df, x = "dose", y = "len", color = "supp", palette = "jco") +
  stat_anova_test(aes(group = supp, color = supp),
  method = "one_way_repeated", wid = "id", group.by = "legend.var")
 }