ggstatsplot 0.12.3

MAJOR CHANGES

• The cryptic but very useful parameter k has been renamed to digits to
  improve its discoverability.

• To be consistent with other functions, ggpiestats() and ggbarstats() now
  default to two-sided alternative hypothesis.

ggstatsplot 0.12.0

BREAKING CHANGES

• To be internally consistent, the plot.type argument has been removed from
  ggbetweenstats(), since no such argument exists for ggwithinstats(). This
  argument was also redundant. Since removing a specific geom is straightforward
  using *.args arguments. Examples for these two functions illustrate how.

• ggbetweenstats() and ggwithinstats() retire pairwise.comparisons
  argument since it was redundant. In order to turn off showing pairwise
  comparisons, you can now use pairwise.display = "none".

NEW FEATURES

• ggbetweenstats() gets boxplot.args argument to pass additional arguments
  to the underlying geom function. This also fixes regression introduced in
  0.11.1 release where outlier points were displayed along with box plot.

ggstatsplot 0.11.1

BREAKING CHANGES

• The outlier tagging functionality in ggbetweenstats() and ggwithinstats()
  has been removed. It was too crude to be useful or reliable, and users should
  instead prefer more informative methods (e.g. performance::check_outliers()).

MINOR CHANGES

• Fix failures due to changes in {parameters}.

ggstatsplot 0.11.0

BREAKING CHANGES

• The minimum needed R version is now bumped to R 4.1 because a crucial
  dependency ({pbkrtest}) requires this R version.

MINOR CHANGES

• Maintenance release to catch up with {ggplot2} and {easystats} updates.

ggstatsplot 0.10.0

BREAKING CHANGES

• The output parameter for all functions has been removed. All functions now
  return only the plot, which itself contains all necessary details that were
  previously extracted using the output argument. You can extract all the
  necessary details (including expressions containing statistical details)
  from a plot using extract_stats() function. There are two additional helpers
  to get expressions: extract_subtitle() and extract_caption().

MAJOR CHANGES

• xfill and yfill arguments for ggscatterstats() have been removed. You
  can specify all aesthetic modifications for side histograms in scatter plot
  using xsidehistogram.args and ysidehistogram.args arguments.

• Updates to changes made in the latest {ggplot2} release (3.4.0).

ggstatsplot 0.9.5

MAJOR CHANGES

• Due to changes to the underlying API of {parameters}, the effsize argument
  has been renamed to effectsize.type.

• Removes unnecessary re-exports of {tidyverse} operators.

MINOR CHANGES

• Fixes tests for changes in dependencies.

ggstatsplot 0.9.4

MINOR CHANGES

• Internal housekeeping to adjust to changes in upstream dependencies.

ggstatsplot 0.9.3

MINOR CHANGES

• Hot fix release to correct a failing example in CRAN daily checks.

ggstatsplot 0.9.2

MAJOR CHANGES

• The pairwise_comparions() function implementation now lives in
  {statsExpressions} package, although it will continue to be exported from
  {ggstatsplot} package.

• The details about pairwise test for ggbetweenstats() and ggwithinstats()
  functions are now displayed as a label for the secondary axis. Previously,
  this information was displayed in the caption. Given that caption already
  contained Bayesian test details, it was becoming difficult to stack
  different expressions on top of each other. To avoid unnecessary code
  complexity and also to avoid crowded caption, this decision was made.
  Additionally, the pairwise test label has been slightly abbreviated, and so
  is the label for significance bars. This is done to not let the text
  overwhelm the numeric values, the latter being more important.

ggstatsplot 0.9.1

MAJOR CHANGES

• Moves {PMCMRplus} package from Imports to Suggests. So, if, as a user, you
  wish to use pairwise comparisons in ggbetweenstats() and
  ggwithinstats(), you will need to download this package.

MINOR CHANGES

• To keep the documentation maintainable, a number of vignettes have either
  been removed or they are no longer evaluated and only code is reported.