A/B Test: Quiz Version Comparison

This project explores whether a new version of a quiz (Version B) leads to significantly higher student scores compared to the original version (Version A). The A/B test was conducted on 100 students, randomly assigned to one of the two versions.

---

Objective

To determine if Quiz Version B improves student performance compared to Quiz Version A using a randomized A/B testing framework and statistical analysis.

---

Methodology

1. Group Assignment

   • 100 students were randomly assigned to Group A or B using set.seed(42) for reproducibility.
   • Assignment was saved in group_assignment.csv.

2. Quiz Administration

   • Students received different versions of the quiz based on their group.
   • Scores were recorded and added to the CSV (group_assignment_with_scores.csv).

3. Analysis

   • An independent two-sample t-test was conducted to compare average scores.
   • A boxplot was generated to visualize score distribution.

---

Project Structure

AB_Test/
├── data/
│   ├── group_assignment.csv
│   └── group_assignment_with_scores.csv
├── output/
│   ├── score_boxplot.png
│   └── ttest_results.txt
├── scripts/
│   ├── 01_assign_groups.R
│   └── 02_analyze_results.R

---

Results

Mean Scores

• Group A: 74.06
• Group B: 78.89

Welch Two-Sample t-test

t = -2.6797, df = 97.338, p-value = 0.008653
95% CI: [-8.41, -1.25]

• Statistically significant difference (p < 0.05)
• Conclusion: Version B resulted in significantly higher quiz scores.

Visualization

---

Tools Used

• R (base, dplyr, ggplot2)
• RStudio
• Git/GitHub for version control

---

How to Reproduce

1. Clone the repo
2. Open in RStudio
3. Run 01_assign_groups.R to generate student group assignments
4. Add scores to group_assignment_with_scores.csv
5. Run 02_analyze_results.R for analysis and plots

---

Author

Asta Musa
Monitoring, Evaluation & Data Analytics Specialist
LinkedIn | GitHub