Confidence.js

Confidence.js is a light-weight JavaScript library to help you make sense of your A/B Test results. Given an A/B test (or "Split Test") result set, Confidence.js will tell you if a statistical "winner" can be determined.

Documentation

* API Reference and Examples
* How It Works

Getting Started

Download confidence.js here

Include confidence.js in your HTML.
`` HTML

`

You're all ready! Start testing... with confidence.

$3

We have many examples in the API Reference. Here's a simple one to get you started:

` js
// Create a new test result
var myConfidence = new Confidence();

// Add variant A results
myConfidence.addVariant({
id: 'A',
name: 'Variant A',
conversionCount: 2500,
eventCount: 3000
});

// Add variant B results
myConfidence.addVariant({
id: 'B',
name: 'Variant B',
conversionCount: 1500,
eventCount: 3000
});

// Get results, using Z-Test method
zTestResult = myConfidence.getResult();
/*
{
hasWinner: true,
hasEnoughData: true,
winnerID: 'A',
winnerName: 'Variant A',
confidencePercent: 95.00,
confidenceInterval: { min: 82, max: 84.67 },
readable: 'With 95% confidence, the true population
parameter of the "Variant A" variant will
fall between 82% and 84.67%.'
}
*/
`

$3

Confidence.js provides two methods for evaluating an A/B test:
- The Z-Test Method
- Chi Square Test and Marascuillo's Procedure

Selecting an appropriate method depends on your data set and use case - both have advantages.

Learn how each method works here.

$3

`
npm install
npm test
`

Feature Requests and TODO

- implement "Confidence to beat baseline" comparison (in progress)
- add removeVariant` function

Issues and Questions

Found a bug? Create an issue here on GitHub!

For general questions, tweet me @jessicaraygun.

Authors

Developed and maintained by Jessica Thomas, Data Scientist @ sendwithus.com, with guidance and tutelage from Statistics Mastermind Emily Malcolm.