A combination of box and violin plots along with jittered data points for between-subjects designs with statistical details included in the plot as a subtitle.
ggbetweenstats(data, x, y, plot.type = "boxviolin",
type = "parametric", pairwise.comparisons = FALSE,
pairwise.annotation = "asterisk", pairwise.display = "significant",
p.adjust.method = "holm", effsize.type = "unbiased",
partial = TRUE, effsize.noncentral = FALSE, bf.prior = 0.707,
bf.message = FALSE, results.subtitle = TRUE, xlab = NULL,
ylab = NULL, caption = NULL, title = NULL, subtitle = NULL,
sample.size.label = TRUE, k = 2, var.equal = FALSE,
conf.level = 0.95, nboot = 100, tr = 0.1, mean.label.size = 3,
mean.label.fontface = "bold", mean.label.color = "black",
notch = FALSE, notchwidth = 0.5, linetype = "solid",
outlier.tagging = FALSE, outlier.shape = 19, outlier.label = NULL,
outlier.label.color = "black", outlier.color = "black",
outlier.coef = 1.5, mean.plotting = TRUE, mean.ci = FALSE,
mean.size = 5, mean.color = "darkred", point.jitter.width = NULL,
point.jitter.height = 0.1, point.dodge.width = 0.6,
ggtheme = ggplot2::theme_bw(), ggstatsplot.layer = TRUE,
package = "RColorBrewer", palette = "Dark2", direction = 1,
ggplot.component = NULL, messages = TRUE)
A dataframe from which variables specified are preferentially to be taken.
The grouping variable from the dataframe data
.
The response (a.k.a. outcome or dependent) variable from the
dataframe data
.
Character describing the type of plot. Currently supported
plots are "box"
(for pure boxplots), "violin"
(for pure violin plots),
and "boxviolin"
(for a combination of box and violin plots; default).
Type of statistic expected ("parametric"
or "nonparametric"
or "robust"
or "bayes"
).Corresponding abbreviations are also accepted:
"p"
(for parametric), "np"
(nonparametric), "r"
(robust), or
"bf"
resp.
Logical that decides whether pairwise comparisons
are to be displayed. Only significant comparisons will be shown by
default. (default: FALSE
). To change this behavior, select appropriate
option with pairwise.display
argument.
Character that decides the annotations to use for
pairwise comparisons. Either "p.value"
or "asterisk"
(default).
Decides which pairwise comparisons to display.
Available options are "significant"
(abbreviation accepted: "s"
) or
"non-significant"
(abbreviation accepted: "ns"
) or
"everything"
/"all"
. The default is "significant"
. You can use this
argument to make sure that your plot is not uber-cluttered when you have
multiple groups being compared and scores of pairwise comparisons being
displayed.
Adjustment method for p-values for multiple
comparisons. Possible methods are: "holm"
(default), "hochberg"
,
"hommel"
, "bonferroni"
, "BH"
, "BY"
, "fdr"
, "none"
.
Type of effect size needed for parametric tests. The
argument can be "biased"
("d"
for Cohen's d for t-test;
"partial_eta"
for partial eta-squared for anova) or "unbiased"
("g"
Hedge's g for t-test; "partial_omega"
for partial
omega-squared for anova)).
Logical that decides if partial eta-squared or omega-squared
are returned (Default: TRUE
). If FALSE
, eta-squared or omega-squared
will be returned. Valid only for objects of class lm
, aov
, anova
, or
aovlist
.
Logical indicating whether to use non-central
t-distributions for computing the confidence interval for Cohen's d
or Hedge's g (Default: FALSE
).
A number between 0.5 and 2 (default 0.707
), the prior width
to use in calculating Bayes factors.
Logical that decides whether to display Bayes Factor in
favor of the null hypothesis for parametric test (Default: FALSE
).
Decides whether the results of statistical tests are
to be displayed as a subtitle (Default: TRUE
). If set to FALSE
, only
the plot will be returned.
Label for x
axis variable.
Label for y
axis variable.
The text for the plot caption.
The text for the plot title.
The text for the plot subtitle. Will work only if
results.subtitle = FALSE
.
Logical that decides whether sample size information
should be displayed for each level of the grouping variable x
(Default:
TRUE
).
Number of digits after decimal point (should be an integer)
(Default: k = 2
).
a logical variable indicating whether to treat the
variances in the samples as equal. If TRUE
, then a simple F
test for the equality of means in a one-way analysis of variance is
performed. If FALSE
, an approximate method of Welch (1951)
is used, which generalizes the commonly known 2-sample Welch test to
the case of arbitrarily many samples.
A scalar value between 0 and 1. If unspecified, the
default is to return 95%
lower and upper confidence intervals (0.95
).
Number of bootstrap samples for computing effect size (Default:
100
).
Trim level for the mean when carrying out robust
tests. If you
get error stating "Standard error cannot be computed because of Winsorized
variance of 0 (e.g., due to ties). Try to decrease the trimming level.",
try to play around with the value of tr
, which is by default set to
0.1
. Lowering the value might help.
Aesthetics for
the label displaying mean. Defaults: 3
, "bold"
,"black"
, respectively.
A logical. If FALSE
(default), a standard box plot will be
displayed. If TRUE
, a notched box plot will be used. Notches are used to
compare groups; if the notches of two boxes do not overlap, this suggests
that the medians are significantly different. In a notched box plot, the
notches extend 1.58 * IQR / sqrt(n)
. This gives a roughly 95%
confidence interval for comparing medians. IQR: Inter-Quartile Range.
For a notched box plot, width of the notch relative to the
body (default 0.5
).
Character strings ("blank"
, "solid"
, "dashed"
,
"dotted"
, "dotdash"
, "longdash"
, and "twodash"
) specifying the type
of line to draw box plots (Default: "solid"
). Alternatively, the numbers
0
to 6
can be used (0
for "blank", 1
for "solid", etc.).
Decides whether outliers should be tagged (Default:
FALSE
).
Hiding the outliers can be achieved by setting outlier.shape = NA. Importantly, this does not remove the outliers, it only hides them, so the range calculated for the y-axis will be the same with outliers shown and outliers hidden.
Label to put on the outliers that have been tagged.
Color for the label to to put on the outliers that
have been tagged (Default: "black"
).
Default aesthetics for outliers (Default: "black"
).
Coefficient for outlier detection using Tukey's method.
With Tukey's method, outliers are below (1st Quartile) or above (3rd
Quartile) outlier.coef
times the Inter-Quartile Range (IQR) (Default:
1.5
).
Logical that decides whether mean is to be highlighted
and its value to be displayed (Default: TRUE
).
Logical that decides whether 95
is to be displayed (Default: FALSE
).
Point size for the data point corresponding to mean
(Default: 5
).
Color for the data point corresponding to mean (Default:
"darkred"
).
Numeric specifying the degree of jitter in x
direction. Defaults to 40%
of the resolution of the data.
Numeric specifying the degree of jitter in y
direction. Defaults to 0.1
.
Numeric specifying the amount to dodge in the x
direction. Defaults to 0.60
.
A function, ggplot2
theme name. Default value is
ggplot2::theme_bw()
. Any of the ggplot2
themes, or themes from
extension packages are allowed (e.g., ggthemes::theme_economist()
,
hrbrthemes::theme_ipsum_ps()
, ggthemes::theme_fivethirtyeight()
, etc.).
Logical that decides whether theme_ggstatsplot
theme elements are to be displayed along with the selected ggtheme
(Default: TRUE
).
Name of package from which the palette is desired as string or symbol.
If a character string (e.g., "Set1"
), will use that named
palette. If a number, will index into the list of palettes of appropriate
type. Default palette is "Dark2"
.
Either 1
or -1
. If -1
the palette will be reversed.
A ggplot
component to be added to the plot prepared
by ggstatsplot
. This argument is primarily helpful for grouped_
variant
of the current function. Default is NULL
. The argument should be entered
as a function.
Decides whether messages references, notes, and warnings are
to be displayed (Default: TRUE
).
For parametric tests, Welch's ANOVA/t-test are used as a default (i.e.,
var.equal = FALSE
).
References:
ANOVA: Delacre, Leys, Mora, & Lakens, PsyArXiv, 2018
t-test: Delacre, Lakens, & Leys, International Review of Social Psychology, 2017
If robust tests are selected, following tests are used is .
ANOVA: one-way ANOVA on trimmed means (see ?WRS2::t1way
)
t-test: Yuen's test for trimmed means (see ?WRS2::yuen
)
Variant of this function ggwithinstats
is currently in progress. You can
still use this function just to prepare the plot for exploratory data
analysis, but the statistical details displayed in the subtitle will be
incorrect. You can remove them by adding + ggplot2::labs(subtitle = NULL)
.
https://indrajeetpatil.github.io/ggstatsplot/articles/web_only/ggbetweenstats.html
# NOT RUN {
# to get reproducible results from bootstrapping
set.seed(123)
# simple function call with the defaults
ggstatsplot::ggbetweenstats(
data = mtcars,
x = am,
y = mpg,
title = "Fuel efficiency by type of car transmission",
caption = "Transmission (0 = automatic, 1 = manual)",
bf.message = TRUE
)
# more detailed function call
ggstatsplot::ggbetweenstats(
data = datasets::morley,
x = Expt,
y = Speed,
plot.type = "box",
conf.level = 0.99,
xlab = "The experiment number",
ylab = "Speed-of-light measurement",
pairwise.comparisons = TRUE,
pairwise.annotation = "p.value",
p.adjust.method = "fdr",
outlier.tagging = TRUE,
outlier.label = Run,
nboot = 10,
ggtheme = ggthemes::theme_few(),
ggstatsplot.layer = FALSE,
bf.message = TRUE
)
# }
Run the code above in your browser using DataCamp Workspace