🧪 A/B Test Statistical Significance Calculator

A robust, web-based utility designed to calculate the statistical significance of A/B tests. Product managers, marketers, and data analysts can seamlessly input visitor and conversion data for Control and Variant groups to quickly determine conversion rates, uplift, p-values, and confidence intervals.

💡 About

This is a lightweight, static frontend application (HTML, CSS, JavaScript) that performs standard A/B test calculations. It serves as a reliable tool for quick, offline checks and as an educational resource for interpreting split-test results without relying on heavy third-party platforms.

✨ Features

• Instant Calculations: Rapidly compute conversion rates, absolute uplift, and relative uplift.
• Statistical Rigor: Outputs accurate Z-scores, p-values, and confidence intervals for conversion rates.
• Actionable Output: Clear, plain-language guidance on statistical significance thresholds.
• Robust Testing: Backed by a rigorous testing suite (unit tests and property‑based tests via Jest + fast-check) to validate calculation correctness across extreme edge cases.
• Zero Dependencies: A clean, responsive UI built with vanilla HTML/CSS/JS. Its small production footprint means it is completely portable and easy to host.

🧮 The Math

This tool calculates significance using a two‑proportion Z‑test (normal approximation). The Z-score is calculated as:

Z = (p₁ - p₂) / √[ p(1 - p) * (1/n₁ + 1/n₂) ]

Where:

• p₁, p₂ = Sample proportions (conversion rates) for the Control and Variant groups.
• n₁, n₂ = Sample sizes (total visitors).
• p = The pooled proportion, calculated as (x₁ + x₂) / (n₁ + n₂), where x₁ and x₂ are the total conversions.

Note on Methodology: > The p‑value is derived from the Z-score and is two‑tailed by default. Confidence intervals utilize the normal approximation method. For smaller sample sizes or extremely low conversion counts, consider using Wilson score intervals or exact methods.

🚀 Usage

1. Launch: Open the live demo or open index.html locally in any web browser.
2. Input Data: Enter your experiment values for both the Control (A) and Variant (B) groups:
   • Visitors (Total sample size)
   • Conversions (Successful actions)
3. Analyze Results: The calculator will immediately display:
   • Conversion rates for each group
   • Absolute and relative uplift
   • Z-score and p‑value
   • Confidence intervals
4. Interpret: Use the calculated p‑value alongside your practical business context to make a decision.

Example Interpretation:

• p < 0.05: Commonly considered statistically significant at the 5% alpha level.
• Tip: Always check the magnitude of the uplift. A difference can be statistically significant but practically insignificant if the relative uplift is negligible.

💻 Local Setup & Installation

Because this is a static web app, no complex build pipelines are required. You can run it directly from your file system or serve it using a lightweight static server.

1. Clone the repository:

   git clone [https://github.com/Deekshithaa-Y-M/ab-test-calculator.git](https://github.com/Deekshithaa-Y-M/ab-test-calculator.git)