New apa_anova() makes APA-style ANOVA source tables for aov and lm models (Type 1 sums of squares, or Types 2 and 3 via car::Anova()), repeated-measures aov models with Error() strata, afex models, and tables from stats::anova() and car::Anova(). Effect sizes (partial, non-partial, or generalized η^2^, ω^2^, or ε^2^) and their two-sided confidence intervals come from effectsize.
New apa_note() writes the text of a table note. For apa_parameters() tables of regression models (including lists of models and lme4 models), it says which robust standard errors were used, how mixed-model degrees of freedom were approximated, and that ordinal thresholds are not tested, and it defines the standardized coefficient, exponentiated coefficient (e.g., OR), confidence interval, and R^2^ change columns. For apa_loadings() and apa_factor_cor() tables, it gives the extraction method and rotation, says that oblique loadings are pattern coefficients, and explains omitted and bolded loadings and the h^2^, u^2^, and Complexity columns. For apa_anova() tables, it gives the type of sums of squares, which effects have sphericity-corrected degrees of freedom (afex models), what the effect size and MSE stand for, and what stars or bold values mean. With apaquarto, use it as #| apa-note: !expr apa7::apa_note(tbl). Without apaquarto, apa_flextable(note = TRUE) places the note below the table.
apa_parameters() now returns the ANOVA table for aov, repeated-measures aov, and afex models. Before, aov models were treated as lm models and gave a coefficients table.
apa_loadings() has been reworked:
apa7_tbl, so it renders as a flextable in Quarto and R Markdown. It supports psych::principal() as well as psych::fa(), and output from parameters::factor_analysis() and parameters::principal_components() (including unrotated components from stats::prcomp()).bold_max argument bolds the largest loading in each row.factor_names argument relabels the factors.variance argument adds rows for the sum of squared loadings and the proportion and cumulative proportion of variance explained, with a rule above them.communality, digits, and format_variable_names arguments. apa_flextable() no longer reformats item text in the Variable column.NA.column_formats argument has been removed.New apa_factor_cor() makes a table of factor correlations from obliquely rotated psych::fa(), psych::principal(), and parameters::factor_analysis() solutions.
column_format() gains an abbreviation slot that says what a header stands for (e.g., "standardized coefficient" for β). apa_note() defines every abbreviated header in a table from these slots, for any table type, so setting the slot in column_formats adds or changes a definition. The default column formats define common abbreviations (e.g., CI, AIC, BIC, RMSEA, effect sizes, and R^2^ variants), and column_formats() gains a get_abbreviations getter. New column formats h2, u2, and Complexity set the headers of those apa_loadings() columns.
apa_flextable() gains a note argument: TRUE adds the note from apa_note() below the table, and a character string adds a custom note. The note does not affect column widths.
apa_flextable() leaves alone any columns named in a "no_format_columns" attribute of the data, and draws a rule above the row named in a "rule_above" attribute.
Partial η^2^ columns (Eta2_partial) are now headed η^2^~p~. New column formats: Sum_Squares, Mean_Square, Eta2_generalized, Omega2, Omega2_partial, Epsilon2, and Epsilon2_partial.
make_apaquarto() and install_apaquarto() now list missing packages separately and ask a short "Install them now?" question, so the prompt is readable in the console (e.g., R on macOS). New ask argument installs missing packages without asking when FALSE. If installation is declined, the function stops and prints the install.packages() command to run.
make_apaquarto shiny app locally.align_chr so that setting format_numeric_character = FALSE works.list_column gains a unite parameter that joins list and text columns.add_star_column no longer add stars if the column value is missing.apa_htest() runs t tests (Welch's or Student's, two-group or one-sample), Wilcoxon rank-sum and signed-rank tests, correlation tests (Pearson, Spearman, or Kendall), and chi-square tests of independence on the variables in a data frame, and makes a table with one row per outcome (e.g., apa_htest(d, mpg + hp ~ am)). It shows group means and standard deviations (or medians) under group spanners, the test statistic, df, p, and an effect size with a two-sided confidence interval: Cohen's d or Hedges' g, the rank-biserial correlation, or Cramér's V, φ, or Cohen's w. Effect sizes are computed from the data, so they do not depend on where the function is called, and Cohen's d for Welch's test is exact rather than approximated from t. apa_note() says which tests were done, how d was standardized, and what the abbreviations stand for.apa_htest() runs paired t tests and Wilcoxon signed-rank tests of paired variables with Pair() in the formula (e.g., apa_htest(d, Pair(pre, post) ~ 1), or several pairs joined with +). Paired t tests show d~av~ by default, which is comparable to d for independent groups; d~z~, d~rm~, and Hedges' g versions come from effectsize::repeated_measures_d(). New pair_labels argument sets the headings above each pair's descriptive statistics.apa_anova() makes F tables of the fixed effects of linear mixed models from lme4::lmer() or lmerTest::lmer(), and formats ANOVA tables from lmerTest's anova(). Tests come from lmerTest, with Kenward-Roger (default, as in apa_parameters()) or Satterthwaite (ddf = "Satterthwaite") denominator degrees of freedom, and are Type III by default. The table shows df~effect~, df~error~, F, and p. Partial effect sizes, computed from F and its degrees of freedom, are shown only when requested (e.g., effect_size = "eta"). apa_note() gives the type of tests, the degrees-of-freedom method, and how the effect size was computed.apa_parameters() and apa_performance() support mixed models fit with glmmTMB::glmmTMB(). Fixed effects are tested with z tests, coefficients of log and logit links are exponentiated by default (e.g., IRR for negative binomial models), and random effects are shown as for lme4 models. Zero-inflation and dispersion parts appear under their own headings ("Zero-inflation model", "Dispersion model"), and apa_note() gives their scale. glmmTMB models also work in lists of models and in apa_performance_comparison().apa_parameters() and apa_performance() support linear mixed models fit with nlme::lme(). Fixed effects are tested with t tests using nlme's degrees of freedom, random effects are shown as for lme4 models, and parameters of residual correlation structures (e.g., corAR1()) and variance functions (e.g., varIdent()) are shown below the residual SD. apa_note() names the structures and says how the degrees of freedom were computed. lme models also work in lists of models and in apa_performance_comparison().apa_performance() makes fit-index tables for lavaan models (e.g., from lavaan::cfa() and lavaan::sem()): the model χ^2^ test, CFI, TLI, RMSEA with its 90% confidence interval, and SRMR (or any of these plus AIC and BIC). Models fit with robust estimators (e.g., MLR or WLSMV) show the scaled χ^2^ test and robust (or scaled) indices. apa_note() gives the sample size, the estimator, and how missing data were handled, and defines the abbreviations. Lists of lavaan models give one row per model.apa_parameters() makes parameter tables for lavaan models. Fully standardized estimates (β) with their standard errors and p-values are shown by default (standardize = FALSE for unstandardized estimates), under headings for factor loadings (e.g., Visual → x1), regressions (e.g., Speed ← Visual), correlations (e.g., Visual ↔ Textual), and defined parameters such as indirect effects. Variances, intercepts, and thresholds can be added. Parameters fixed to set the scale are not tested, and apa_note() names them, along with the estimator, the standardization, and robust or bootstrap standard errors.apa_flextable() no longer re-formats row labels that apa_parameters() (for lme4 and glmmTMB models) and apa_htest() have already formatted. Before, random-effect labels such as "Subject: r (Constant, Days)" lost their spacing and commas in flextables.apa_flextable(note = TRUE), apa_cor(), apa_chisq(), and footers styled by apa_style(). line_spacing now sets only the spacing of the header and body.