Normality Test Calculator

Paste a numeric sample to measure skewness and kurtosis with a Jarque–Bera test. Use this as a diagnostic, not a reason to discard large-sample experiments automatically.

0.445
Jarque–Bera statistic
0.8006
p-value
0.000
Skewness
-1.033
Excess kurtosis

The test does not reject normality at α = 0.05.

Normality tests become hypersensitive in very large samples and underpowered in small ones. For conversion rates, use a proportion test; for continuous experiment metrics, inspect the distribution and consider robust inference alongside this diagnostic.

More free A/B testing tools

All tools