Two-Way ANOVA in R — Practical & Production Guide

📌 Overview

Two-Way ANOVA (Analysis of Variance) is a statistical technique used to evaluate:

• The effect of two categorical independent variables (factors)
• On a single continuous dependent variable
• Including whether there is a statistical interaction between the two factors

This method is widely applied in:

• Experimental design
• A/B testing
• Behavioral and social sciences
• Business and product analytics

---

⚙️ Prerequisites

Before running the analysis, ensure:

• R ≥ 3.6 installed
• Basic familiarity with R syntax
• Foundational understanding of ANOVA concepts

---

📦 Installation

Install required packages:

install.packages(c("ggplot2", "dplyr", "car"))

Load them:

library(ggplot2)
library(dplyr)
library(car)

---

🔄 End-to-End Workflow

1. Load Data

data <- read.csv("path_to_your_dataset.csv")
head(data)

---

2. Inspect & Prepare Data

Check structure:

str(data)

Ensure:

• Independent variables → Factors
• Dependent variable → Numeric

Convert if needed:

data$Factor1 <- as.factor(data$Factor1)
data$Factor2 <- as.factor(data$Factor2)

---

3. Fit the Two-Way ANOVA Model

anova_model <- aov(DependentVariable ~ Factor1 * Factor2, data = data)
summary(anova_model)

💡 The * operator includes:

• Main effect of Factor1
• Main effect of Factor2
• Interaction effect Factor1:Factor2

---

✅ Assumption Checks (Don’t Skip This)

4. Homogeneity of Variance

leveneTest(DependentVariable ~ Factor1 * Factor2, data = data)

✔️ Interpretation:

• p > 0.05 → assumption satisfied
• p ≤ 0.05 → variance is unequal (consider alternatives)

---

5. Normality of Residuals

Q-Q Plot

qqnorm(residuals(anova_model))
qqline(residuals(anova_model))

Shapiro-Wilk Test

shapiro.test(residuals(anova_model))

✔️ Interpretation:

• p > 0.05 → residuals approximately normal

---

📊 Interpreting Results

From:

summary(anova_model)

Component	Interpretation
Factor1	Effect of first independent variable
Factor2	Effect of second independent variable
Factor1:Factor2	Interaction between the two variables

🔑 Key Insight

• If interaction is significant, it takes priority
• This means: 👉 The effect of one factor depends on the other

⚠️ Do not interpret main effects in isolation when interaction is significant.

---

🔍 Post-Hoc Analysis

If significant effects are found:

TukeyHSD(anova_model)

This identifies:

• Which specific group pairs differ
• The magnitude and direction of differences

---

📈 Visualization (Interaction Plot)

ggplot(data, aes(x = Factor1, y = DependentVariable, color = Factor2)) +
  geom_point(position = position_jitterdodge()) +
  stat_summary(fun = mean, geom = "line", aes(group = Factor2)) +
  labs(
    title = "Interaction Plot",
    x = "Factor 1",
    y = "Dependent Variable"
  ) +
  theme_minimal()

💡 Why this matters:

• Helps visually confirm interaction effects
• Makes interpretation faster and clearer

---

🧪 Reproducible Example

data <- data.frame(
  Factor1 = factor(rep(c("A", "B"), each = 10)),
  Factor2 = factor(rep(c("X", "Y"), 10)),
  DependentVariable = c(
    23, 21, 19, 30, 29, 31, 23, 22, 24, 26,
    30, 28, 25, 27, 29, 22, 24, 21, 23, 20
  )
)

anova_model <- aov(DependentVariable ~ Factor1 * Factor2, data = data)
summary(anova_model)

---

⚠️ Common Pitfalls

• Treating numeric variables as categorical incorrectly
• Ignoring interaction effects
• Skipping assumption checks
• Using small sample sizes (low statistical power)
• Over-interpreting insignificant results

---

✅ Best Practices

• Always visualize before and after modeling
• Validate assumptions before drawing conclusions
• Use post-hoc tests for deeper insights
• Keep your data clean and structured
• Document your workflow for reproducibility

---