vignettes/web_only/ggwithinstats.Rmd
ggwithinstats.Rmd
The function ggstatsplot::ggwithinstats
is designed to facilitate data exploration, and for making highly customizable publication-ready plots, with relevant statistical details included in the plot itself if desired. We will see examples of how to use this function in this vignette.
To begin with, here are some instances where you would want to use ggwithinstats
-
to check if a continuous variable differs across multiple groups/conditions
to compare distributions visually and check for outliers
Note: This vignette uses the pipe operator (%>%
), if you are not familiar with this operator, here is a good explanation: http://r4ds.had.co.nz/pipes.html
ggwithistats
To illustrate how this function can be used, we will use the bugs
dataset throughout this vignette. This data set, “Bugs”, provides the extent to which men and women want to kill arthropods that vary in freighteningness (low, high) and disgustingness (low, high). Each participant rates their attitudes towards all anthropods. Subset of the data reported by Ryan et al. (2013). Note that this is a repeated measures design because the same participant gave four different ratings across four different conditions (LDLF, LDHF, HDLF, HDHF).
Suppose the first thing we want to inspect is the distribution of desire to kill across all conditions (disregarding the factorial structure of the experiment). We also want to know if the mean differences in this desire across conditions is statistically significant.
The simplest form of the function call is-
# since the confidence intervals for the effect sizes are computed using
# bootstrapping, important to set a seed for reproducibility
set.seed(123)
library(ggstatsplot)
# function call
ggstatsplot::ggwithinstats(
data = bugs_long,
x = condition,
y = desire
)
Note:
The function automatically decides whether a dependent samples test is preferred (for 2 groups) or an ANOVA (3 or more groups). based on the number of levels in the grouping variable.
The output of the function is a ggplot
object which means that it can be further modified with ggplot2
functions.
As can be seen from the plot, the function by default returns Bayes Factor for the test. If the null hypothesis can’t be rejected with the null hypothesis significance testing (NHST) approach, the Bayesian approach can help index evidence in favor of the null hypothesis (i.e., \(BF_{01}\)).
By default, natural logarithms are shown because Bayes Factor values can sometimes be pretty large. Having values on logarithmic scale also makes it easy to compare evidence in favor alternative (\(BF_{10}\)) versus null (\(BF_{01}\)) hypotheses (since \(log_{e}(BF_{01}) = - log_{e}(BF_{10})\)).
We can make the output much more aesthetically pleasing as well as informative by making use of the many optional parameters in ggwithinstats
. We’ll add a title and caption, better x
and y
axis labels, and tag and label the outliers in the data. We can and will change the overall theme as well as the color palette in use.
# for reproducibility
set.seed(123)
library(ggstatsplot)
library(firatheme)
# plot
ggstatsplot::ggwithinstats(
data = bugs_long,
x = condition,
y = desire,
type = "nonparametric", # type of statistical test
xlab = "Condition", # label for the x-axis
ylab = "Desire to kill an artrhopod", # label for the y-axis
effsize.type = "biased", # type of effect size
sphericity.correction = FALSE, # don't display sphericity corrected dfs and p-values
pairwise.comparisons = TRUE, # display pairwise comparisons
outlier.tagging = TRUE, # whether outliers should be flagged
outlier.coef = 1.5, # coefficient for Tukey's rule
outlier.label = region, # label to attach to outlier values
outlier.label.color = "red", # outlier point label color
mean.plotting = TRUE, # whether the mean is to be displayed
mean.color = "darkblue", # color for mean
ggtheme = firatheme::theme_fira(), # a different theme
ggstatsplot.layer = FALSE, # turn off default modification of the used theme
package = "yarrr", # package from which color palette is to be taken
palette = "info2", # choosing a different color palette
title = "Comparison of desire to kill bugs",
caption = "Source: Ryan et al., 2013"
) + # modifying the plot further
ggplot2::scale_y_continuous(
limits = c(0, 10),
breaks = seq(from = 0, to = 10, by = 1)
)
As can be appreciated from the effect size (partial eta squared) of 0.18, there are small differences in the mean desire to kill across conditions. Importantly, this plot also helps us appreciate the distributions within any given condition.
So far we have only used a classic parametric test, but we can also use other available options: The type
(of test) argument also accepts the following abbreviations: "p"
(for parametric), "np"
(for nonparametric), "r"
(for robust), "bf"
(for Bayes Factor).
Let’s use the combine_plots
function to make one plot from four separate plots that demonstrates all of these options. Let’s compare desire to kill bugs only for low versus high disgust conditions to see how much of a difference whether a bug is disgusting-looking or not makes to the desire to kill that bug. We will generate the plots one by one and then use combine_plots
to merge them into one plot with some common labeling. It is possible, but not necessarily recommended, to make each plot have different colors or themes.
For example,
# for reproducibility
set.seed(123)
library(ggstatsplot)
# selecting subset of the data
df_disgust <- dplyr::filter(.data = bugs_long, condition %in% c("LDHF", "HDHF"))
# parametric t-test
p1 <-
ggstatsplot::ggwithinstats(
data = df_disgust,
x = condition,
y = desire,
type = "p",
effsize.type = "d",
conf.level = 0.99,
title = "Parametric test",
package = "ggsci",
palette = "nrc_npg",
ggtheme = ggthemr::ggthemr(palette = "light")
)
# Mann-Whitney U test (nonparametric test)
p2 <-
ggstatsplot::ggwithinstats(
data = df_disgust,
x = condition,
y = desire,
xlab = "Condition",
ylab = "Desire to kill bugs",
type = "np",
conf.level = 0.99,
title = "Non-parametric Test",
package = "ggsci",
palette = "uniform_startrek",
ggtheme = ggthemes::theme_map(),
ggstatsplot.layer = FALSE
)
# robust t-test
p3 <-
ggstatsplot::ggwithinstats(
data = df_disgust,
x = condition,
y = desire,
xlab = "Condition",
ylab = "Desire to kill bugs",
type = "r",
conf.level = 0.99,
title = "Robust Test",
package = "wesanderson",
palette = "Royal2",
ggtheme = hrbrthemes::theme_ipsum_tw(),
ggstatsplot.layer = FALSE
)
# Bayes Factor for parametric t-test
p4 <-
ggstatsplot::ggwithinstats(
data = df_disgust,
x = condition,
y = desire,
xlab = "Condition",
ylab = "Desire to kill bugs",
type = "bf",
title = "Bayesian Test",
package = "ggsci",
palette = "nrc_npg",
ggtheme = ggthemes::theme_fivethirtyeight()
)
# combining the individual plots into a single plot
ggstatsplot::combine_plots(
p1, p2, p3, p4,
nrow = 2,
title.text = "Effect of disgust on desire to kill bugs ",
caption.text = "Source: Bugs dataset from `jmv` R package",
title.size = 14,
caption.size = 12
)
grouped_ggwithinstats
What if we want to carry out this same analysis but for each region (or gender)?
ggstatsplot
provides a special helper function for such instances: grouped_ggwithinstats
. This is merely a wrapper function around ggstatsplot::combine_plots
. It applies ggwithinstats
across all levels of a specified grouping variable and then combines list of individual plots into a single plot. Note that the grouping variable can be anything: conditions in a given study, groups in a study sample, different studies, etc.
Let’s focus on the two regions and for years: 1967, 1987, 2007. Also, let’s carry out pairwise comparisons to see if there differences between every pair of continents.
# for reproducibility
set.seed(123)
library(ggstatsplot)
ggstatsplot::grouped_ggwithinstats(
# arguments relevant for ggstatsplot::ggwithinstats
data = bugs_long,
x = condition,
y = desire,
grouping.var = gender,
xlab = "Continent",
ylab = "Desire to kill bugs",
type = "nonparametric", # type of test
pairwise.display = "significant", # display only significant pairwise comparisons
p.adjust.method = "BH", # adjust p-values for multiple tests using this method
ggtheme = ggthemes::theme_tufte(),
package = "ggsci",
palette = "default_jco",
outlier.tagging = TRUE,
ggstatsplot.layer = FALSE,
outlier.label = education,
k = 3,
title.prefix = "Year",
# arguments relevant for ggstatsplot::combine_plots
title.text = "Desire to kill bugs across genders",
plotgrid.args = list(ncol = 1)
)
ggwithinstats
+ purrr
Although this grouping function provides a quick way to explore the data, it leaves much to be desired. For example, the same type of test and theme is applied for all genders, but maybe we want to change this for different genders, or maybe we want to gave different effect sizes for different years. This type of customization for different levels of a grouping variable is not possible with grouped_ggwithinstats
, but this can be easily achieved using the purrr
package.
See the associated vignette here: https://indrajeetpatil.github.io/ggstatsplot/articles/web_only/purrr_examples.html
For independent measures designs, ggbetweenstats
function can be used: https://indrajeetpatil.github.io/ggstatsplot/articles/web_only/ggbetweenstats.html
The central tendency measure displayed will depend on the statistics:
Type | Measure | Function used |
---|---|---|
Parametric | mean | parameters::describe_distribution |
Non-parametric | median | parameters::describe_distribution |
Robust | trimmed mean | parameters::describe_distribution |
Bayesian | MAP estimate | parameters::describe_distribution |
MAP: maximum a posteriori probability
Following (within-subjects) tests are carried out for each type of analyses-
Type | No. of groups | Test |
---|---|---|
Parametric | > 2 | One-way repeated measures ANOVA |
Non-parametric | > 2 | Friedman rank sum test |
Robust | > 2 | Heteroscedastic one-way repeated measures ANOVA for trimmed means |
Bayes Factor | > 2 | One-way repeated measures ANOVA |
Parametric | 2 | Student’s t-test |
Non-parametric | 2 | Wilcoxon signed-rank test |
Robust | 2 | Yuen’s test on trimmed means for dependent samples |
Bayesian | 2 | Student’s t-test |
Following effect sizes (and confidence intervals/CI) are available for each type of test-
Type | No. of groups | Effect size | CI? | Function used |
---|---|---|---|---|
Parametric | > 2 | \(\eta_{p}^2\), \(\omega_{p}^2\) | Yes |
effectsize::omega_squared , effectsize::eta_squared
|
Non-parametric | > 2 | \(W_{Kendall}\) (Kendall’s coefficient of concordance) | Yes | effectsize::kendalls_w |
Robust | > 2 | \(\delta_{R-avg}^{AKP}\) Yes | Algina-Keselman-Penfield robust standardized difference average |
wAKP.avg (custom) |
Bayes Factor | > 2 | \(R_{posterior}^2\) | Yes | performance::r2_bayes |
Parametric | 2 | Cohen’s d, Hedge’s g | Yes |
effectsize::cohens_d , effectsize::hedges_g
|
Non-parametric | 2 | r (rank-biserial correlation) | Yes | effectsize::rank_biserial |
Robust | 2 | \(\delta_{R}^{AKP}\) (Algina-Keselman-Penfield robust standardized difference) | Yes | WRS2::dep.effect |
Bayesian | 2 | \(\delta_{posterior}\) | Yes | bayestestR::describe_posterior |
Here is a summary of multiple pairwise comparison tests supported in ggwithinstats-
Type | Test | p-value adjustment? | Function used |
---|---|---|---|
Parametric | Student’s t-test | Yes | stats::pairwise.t.test |
Non-parametric | Durbin-Conover test | Yes | PMCMRplus::durbinAllPairsTest |
Robust | Yuen’s trimmed means test | Yes | WRS2::rmmcp |
Bayesian | Student’s t-test | NA |
BayesFactor::ttestBF |
To see how the effect sizes displayed in these tests can be interpreted, see: https://indrajeetpatil.github.io/ggstatsplot/articles/web_only/effsize_interpretation.html
If you find any bugs or have any suggestions/remarks, please file an issue on GitHub
: https://github.com/IndrajeetPatil/ggstatsplot/issues
For details, see- https://indrajeetpatil.github.io/ggstatsplot/articles/web_only/session_info.html