Loading...
Loading...
Check if your A/B test data follows a normal distribution using bootstrap resampling and the Jarque-Bera test.
The Normality Test helps you verify whether your data follows a normal distribution, which is a key assumption for many statistical tests used in experimentation. Paste your raw data and the tool will run a Jarque-Bera test and display a histogram so you can visually and statistically assess the distribution of your sample.
Standard statistical tests used in A/B testing — Z-tests, T-tests, and ANOVA, assume that the sampling distribution of the test statistic is approximately normal. This assumption is often satisfied by the Central Limit Theorem (CLT), which states that the mean of a sufficiently large sample will be approximately normally distributed regardless of the underlying data distribution.
However, the CLT has limits. Revenue data with extreme outliers, heavily skewed metrics, or bimodal distributions may require very large sample sizes before the CLT kicks in. This tool helps you verify whether your specific data meets the normality assumption at your current sample size.
Rather than testing each group individually, this tool checks the normality of the bootstrap distribution of mean differences. This is the distribution that actually matters for hypothesis testing. It's the distribution from which your test statistic is drawn.
The bootstrap generates thousands of resampled datasets, computes the mean difference for each, and then applies the Jarque-Bera test to this distribution. This approach is more relevant than testing raw data normality because even non-normal raw data can produce a normal sampling distribution of means (thanks to the CLT), and vice versa.
Explore more A/B testing and statistics tools
Analyze A/B test results with frequentist Z-tests and T-tests.
Detect sample ratio mismatch in your experiment.
Calculate standard deviation, variance, and coefficient of variation.