Experimental Design with R — Statistical Testing Toolkit

A comprehensive R toolkit for analyzing experimental data using statistical tests — from visual exploration to assumptions checking, hypothesis testing, and post-hoc analysis with both parametric and non-parametric methods.

Datasets Used

Dataset	Description	Source
ToothGrowth	Tooth length by supplement and dose	datasets
PlantGrowth	Plant weights under different conditions	datasets
mice2	Mouse weights before/after treatment	datarium
selfesteem	Self-esteem scores at 3 time points	datarium
anxiety	Anxiety scores across activity levels and time	datarium
poison.data	Survival time under poisons and treatments	BHH2
dental, rat, respiration	Advanced non-parametric analysis data	nparLD

Key Topics Covered

Parametric Tests

• One-sample, two-sample (independent & paired) t-tests
• One-way and two-way ANOVA
• Repeated measures ANOVA
• Mixed ANOVA (between and within factors)
• Post-hoc tests (Tukey HSD, Bonferroni-adjusted t-tests)

Non-Parametric Alternatives

• Wilcoxon signed-rank test
• Mann-Whitney U test
• Kruskal-Wallis test
• Friedman test
• Non-parametric mixed ANOVA (nparLD)

Assumptions Testing

• Normality: Shapiro-Wilk, Jarque-Bera, Anderson-Darling, Lilliefors
• Homogeneity of variances: Bartlett, Fligner-Killeen, Levene
• Outliers: Z-scores and IQR
• Sphericity: Mauchly's test (repeated measures)

Getting Started

Prerequisites

• R 4.0+
• RStudio (recommended)

Installation

git clone https://github.com/Wobbly1212/statistical-tests-experimental-design-R.git
cd statistical-tests-experimental-design-R

Install all required packages:

source("install_packages.R")

Usage

1. Open experimental_design.R in RStudio
2. Run section by section based on your use case
3. Ensure sonno-long.csv (if used) is in your working directory

Project Structure

statistical-tests-experimental-design-R/
├── experimental_design.R   # Main analysis script
├── install_packages.R      # Dependency installer
├── LICENSE
└── README.md

Author

Diako Darabi

License

This project is licensed under the MIT License.