Heads-up: ggstatsplot 1.0.0 stable release with breaking changes

[IndrajeetPatil]

We are preparing to submit {ggstatsplot} 1.0.0 to CRAN — the first stable release of this package. We wanted to give you a heads-up since your package lists {ggstatsplot} as a dependency.

We apologize in advance if any of these changes break things for your package. The good news is that after this release, the API will be stable and the package is entering maintenance-only mode — no further breaking changes are expected.

We'd appreciate it if you could check your package against the development version and adjust as needed:

pak::pak("IndrajeetPatil/ggstatsplot")

Changelog for the upcoming release

NEW FEATURES

• ggbarstats() (and grouped_ggbarstats()) now supports one-sample goodness-of-fit tests when only x is specified (with y = NULL as the new default).

• ggpiestats() and ggbarstats() now compute pairwise contingency table analyses (Fisher's exact tests via pairwise_contingency_table()) when x has more than two levels.

• ggscatterstats() now sets a default number of bins for marginal histograms, suppressing the stat_bin() message about picking a better binwidth.

• ggscatterstats() gains xsidehistogram.scale and ysidehistogram.scale parameters to control the scale of the marginal distribution histograms.

• All top-level plotting functions now include an alternative argument for specifying the alternative hypothesis for effect size confidence intervals.

• ggbetweenstats() and ggwithinstats() now include a pairwise.alpha argument to control the alpha cutoff for filtering displayed pairwise comparisons.

BREAKING CHANGES

• The following expert-level statistical parameters have been removed from all function signatures because their defaults are the universally recommended values: var.equal, nboot, sampling.plan, fixed.margin, prior.concentration, effsize.type. Users who need non-default values should call {statsExpressions} directly.

• ggwithinstats() gains a subject.id parameter for correctly pairing observations. Plots for unsorted repeated-measures data may differ once subject.id is supplied.

• The package argument has been removed from all plotting functions. The palette argument now accepts a single "package::palette" string (e.g., palette = "ggthemes::gdoc").

• The default palette has changed from "RColorBrewer::Dark2" (8 colors) to "ggthemes::gdoc" (24 colors).

• When the chosen palette does not have enough colors, an error is now thrown (previously a warning was issued).

• pairwise_comparisons() and pairwise_contingency_table() are no longer re-exported from {ggstatsplot}. Use statsExpressions::pairwise_comparisons() and statsExpressions::pairwise_contingency_table() directly instead.

BUG FIXES

• Grouped plot functions now preserve the order of groups as they appear in the data rather than sorting them alphabetically.

• combine_plots() now renders the overall annotation title in bold by default.

• grouped_ggbarstats() and grouped_ggpiestats() now display a single unified legend across different groups.

• ggbetweenstats() and ggwithinstats() now correctly display sample size labels on the x-axis even when centrality.plotting = FALSE.

• ggcoefstats() now preserves the model term order and no longer draws empty stats.labels boxes for model terms whose label expression is absent.

Thank you for using {ggstatsplot}! If you run into any issues, please don't hesitate to file an issue at https://github.com/IndrajeetPatil/ggstatsplot/issues.