Added correlation-adjusted intervals to get_adjusted_ci() with adjustments = list(decorrelation = "CA"). Correlations are computed separately within each between-subject group after cell aggregation and incomplete-subject removal. Cousineau-Morey ("CM") remains the default.
get_adjusted_ci() no longer looks up SE.<name>/CI.<name> precision functions by name; for summary functions without a built-in formula, pass errorbar as a function.
Compacted interval-precision selection, ANOVA and regression formatting, and confidence-interval text formatting while preserving existing output conventions.
Simplified plot_pointrange() to use the same interval columns for between- and within-subject summaries, including margin summaries.
Made incomplete-cell diagnostics part of the warning or error so that suppressing a warning also suppresses its diagnostic table.
Fixed format_results(type = "plotmath") returning NULL when the input contains no comma.
Addressed CRAN review: removed quotes around APA in the package metadata, replaced T/F logical abbreviations, documented return classes and meanings (including deprecated functions), added package links, and corrected HTML rendering of column-name placeholders.
Made examples executable or marked long examples with \donttest{}, replaced the slow BayesFactor example with a small t-test, and added regression coverage for interval calculations, formatting, and plot summaries.
apastats2 1.0.6
get_adjusted_ci() now supports correlation-adjusted (decorrelation = "CA") intervals for within-subject and mixed designs, using the mean pairwise condition correlation separately within each between-subject group. The default remains Cousineau-Morey ("CM").
CA supports the existing aggregation, confidence-level, precision-function, missing-subject, and difference-purpose options. Undefined correlations produce informative errors. The implementation has no runtime dependency on superb; regression tests compare against it when available.
apastats2 1.0.5
Removed the internal summarySE(), summarySEwithin() and normDataWithin() helpers; plot_pointrange() now computes margins with get_adjusted_ci(), which also fixes add_margin = TRUE with within_subj = TRUE.
get_adjusted_ci() no longer warns that the drop argument is ignored when reporting incomplete cells.
Fixed typos in the documentation, added a spelling word list, and made the apa.BFBayesFactor() example faster.
Changed the maintainer e-mail address.
apastats2 1.0.4
Added the memory_noise example dataset derived from Chetverikov & Hansmann-Roth (2026), using unequal-noise trials and density-asymmetry bias scores; Exp. 1A vs. Exp. 1B is the between-subject factor and target relative noise is the within-subject factor.
Removed the legacy faces example dataset; maintained examples now use memory_noise or built-in datasets.
Added get_adjusted_ci() as the general interval engine for between-subject, within-subject, and mixed designs; plot_pointrange() now uses it for both between- and within-subject summaries.
Reimplemented Cousineau-Morey intervals internally, removing the superb/reshape2 dependency and the development-branch Remotes entry while documenting the implementation’s basis in the superb framework.
get_adjusted_ci() supports arbitrary scalar summary functions, with superb-compatible SE/CI formulas for common statistics and custom precision-function hooks. The old get_superb_ci() name remains as a deprecated compatibility wrapper.
Added apa() support for stats::aov() models with Error() terms (aovlist objects), fixing #9.