rk.rstatix: Modern & Robust Statistics for RKWard

A Modern RKWard GUI Plugin for Robust Statistical Tests using the rstatix Package

rk.rstatix brings the power of "Tidy" statistics to the RKWard graphical interface. By leveraging the popular rstatix package, this plugin bypasses the messy text outputs of base R, providing researchers with clean, pipe-friendly, and publication-ready HTML tables. It focuses on filling critical methodological gaps in standard statistical software, such as handling unequal variances, adjusting for covariates, and diagnosing extreme outliers.

---

🌟 Key Features

• "Tidy" Results: All statistical tests automatically return tibble data frames, which RKWard renders into beautiful, Excel-ready HTML tables.
• Robust Methodologies: Easily perform tests that do not assume equal variances (heteroscedasticity), which is the reality of most experimental data.
• "Auto-Focus" Magic UI: The user interface features smart variable selectors. When you select a dataset, the available variables list instantly filters to show only the columns within that specific dataframe.
• Safe R Code Generation: Uses isolated Javascript parsing to protect column names with special characters and avoid namespace collisions.
• Multilingual: Fully translated into English, Spanish, French, German, and Portuguese (Brazil).

---

What's New in Version 0.0.1

🚀 Initial Release: The Psychology & Social Sciences Toolkit

This first release introduces three highly-requested analytical modules that are traditionally tedious to code by hand:

• Welch ANOVA & Games-Howell: The gold standard for one-way comparisons when Levene's test fails. Includes automatic Eta-Squared ($\eta^2$) effect size calculation and dynamic list-saving capabilities.
• ANCOVA (Analysis of Covariance): Control for one or multiple continuous confounding variables. Automatically calculates and compares Estimated Marginal Means (EMMeans) with Bonferroni adjustments.
• Outlier Diagnostics: Quickly identify mild ($1.5 \times IQR$) or extreme ($3 \times IQR$) univariate outliers grouped by experimental conditions, helping you clean your data before parametric testing.

---

⚙️ Prerequisites

You must have RKWard installed along with the following R packages. The plugin will automatically load them when required:

install.packages(c("rstatix", "dplyr"))

---

🚀 Installation

You can install this plugin directly from GitHub using remotes inside your RKWard console:

# Install the plugin
library("remotes")
# install_github("AlfCano/rk.rstatix")
install.packages("rk.rstatix", repos = c('https://alfcano.r-universe.dev', 'https://cloud.r-project.org'))

Once installed, restart RKWard, navigate to Settings -> Configure RKWard -> Plugins, and ensure rk.rstatix is activated.

---

🛠️ Usage Workflow

This plugin adds three new powerful tools to your RKWard menus:

1. Robust ANOVA (Welch & Games-Howell)

Located under: Analysis ➔ Means

• The Problem: Standard Fisher ANOVA assumes equal variances across groups. If violated, Type I error rates inflate.
• The Solution: Select your Dependent Variable and Grouping Factor. The plugin calculates a Welch ANOVA.
• Options: Automatically append the $\eta^2$ effect size and compute the Games-Howell post-hoc test. Save all generated tables directly into a single organized List object in your R workspace.

2. ANCOVA (Analysis of Covariance)

Located under: Analysis ➔ Psychology / Social Science ➔ Frequentist

• The Problem: You need to evaluate the effect of a categorical treatment while controlling for pre-existing continuous covariates (e.g., age, baseline anxiety).
• The Solution: Input your DV, Grouping Factor, and select one or more Covariates.
• Options: Extract Estimated Marginal Means (EMMeans) adjusted for the covariates, and perform Bonferroni-corrected pairwise comparisons.

3. Identify Outliers

Located under: Analysis ➔ Psychology / Social Science ➔ Frequentist

• The Problem: Reviewers often demand proof that your data doesn't contain extreme outliers before running parametric tests.
• The Solution: Select a numeric variable and an optional grouping factor. The plugin will output a table containing only the anomalous participants.
• Options: Toggle between identifying all outliers (beyond $1.5 \times IQR$) or strictly EXTREME outliers (beyond $3 \times IQR$).

---

🌍 Internationalization (i18n)

The graphical interface automatically adapts to your RKWard language settings. Currently supported languages:

• 🇺🇸 English (Default)
• 🇪🇸 Spanish (Español)
• 🇫🇷 French (Français)
• 🇩🇪 German (Deutsch)
• 🇧🇷 Portuguese (Português do Brasil)

---

📝 License and Author

Author: Alfonso Cano (@AlfCano)
Email: alfonso.cano@correo.buap.mx

• Assisted by: Gemini, a large language model from Google.
• License: GPL (>= 3)

This project is licensed under the GPL (>= 3) License.