easy_ttest() now defaults to var.equal = TRUE (the classical
pooled-variance t-test). Pass var.equal = FALSE to use Welch's t-test
instead.
easy_ttest() accepts a categorical/numeric pair in any order
(e.g. easy_ttest(group, outcome) or easy_ttest(outcome, group)) and
reshapes it into the standard outcome ~ group comparison automatically.
Fixed a bug where easy_ttest() crashed on a one-sample test (x numeric,
no y) because broom::tidy() returns estimate rather than estimate1
for one-sample t.test() objects.
Test statistics (t, z, F, χ²) are now reported as absolute values
across easy_ttest(), easy_ztest(), easy_ftest(), easy_anova(),
easy_chisq(), and easy_correlation(), since their sign is an artifact
of argument order rather than meaningful direction.
Mean differences in easy_ttest() and easy_ztest() are now reported as
larger − smaller and are always non-negative; the accompanying confidence
interval is recomputed in the same direction so its bounds no longer
appear sign-flipped relative to the point estimate.
easy_ttest() now reports both the degrees of freedom actually used (Welch
or pooled, depending on var.equal) and the equal-variance degrees of
freedom (n1 + n2 - 2) side by side, along with explicit n1/n2 sample
sizes and a "Variance Assumption" label.
easy_ztest() two-sample table now also reports n1/n2.
Variable names inferred from expressions like mtcars$mpg or
data[["mpg"]] are now cleaned to just mpg in result tables and plot
labels (easy_ttest(), easy_ztest(), easy_ftest(), easy_correlation(),
easy_chisq(), easy_wilcox(), easy_describe(), and the easy_* plot
functions).
Added easy_logistic_regression() for binary logistic regression with odds
ratios, approximate odds-ratio confidence intervals, likelihood-ratio model
tests, McFadden pseudo-R2, and automated narrative reporting.
Added richer supporting tables:
Added easy_wilcox() and easy_kruskal() for common non-parametric
workflows.
Added regression figures: easy_regression_diagnostics() and
easy_odds_ratio_plot().
Added comprehensive descriptive statistics: easy_describe() and
easy_group_summary() with 21 statistics per variable (mean, median, mode,
SD, variance, skewness, kurtosis, normality test, CI, and more).
Extended inferential test suite with four new functions:
easy_chisq(): Chi-square test of independence and goodness-of-fit,
with Cramér's V effect size.easy_ztest(): One- and two-sample z-tests with Cohen's d.easy_ftest(): Levene-style F-test for equality of variances via
stats::var.test(), with variance ratio CI.easy_correlation(): Pearson/Spearman/Kendall bivariate and pairwise
matrix correlation with confidence intervals.New visualization functions (all return easystat_result objects with
plain-language interpretations):
easy_histogram(): distribution histogram with normal-curve overlay.easy_boxplot(): grouped box-and-whisker plots with outlier marking.easy_scatter(): scatter plot with regression line and R-squared label.easy_barplot(): count or mean (+/- SE) bar charts with group support.easy_qqplot(): Q-Q normality plot.easy_density(): kernel density curve, optionally grouped.easy_correlation_heatmap(): annotated pairwise correlation heatmap.easy_autoplot(): smart dispatcher that picks the right plot type
for a given easystat_result.New theme: theme_easystat() — a professional ggplot2 theme and color
system applied uniformly across all visualization functions.
Word export: export_to_word() creates a fully formatted .docx report
(via flextable and officer) with title, authors, narrative, and tables.
easystat_result S3 object with consistent
fields: test_type, formula_str, raw_model, coefficients_table,
model_fit_table, and explanation.print.easystat_result) redesigned with improved
typography and responsive layout.easy_regression(), easy_ttest(), easy_anova(),
and basic print() / summary() S3 methods.