Text & documents
Numbers & maths
Data & formats
Security
Development & DevOps
Artificial Intelligence
Finance
Health & Wellness
Productivity
Games & Entertainment
Multimedia & design
Business
How to use
What a p-value is

Under the null hypothesis (H₀) the statistic follows a known distribution. The p-value (blue area) is the probability of observing a result at least as extreme if H₀ were true. If it falls in the rejection region (red, defined by α) → you reject H₀. Note: not rejecting ≠ accepting H₀, and a small p does not measure the effect size.

Hypothesis testingp-values, t-test, χ², ANOVA and power
Hypothesis testingp-values, t-test, χ², ANOVA and power — visual and interactive

Move the observed statistic and the α level. The blue area is the p-value; the red one is the rejection region. You reject H₀ when the blue enters the red.

p-valuerejection region (α)observed statistic
Statistic z
2.1
Level α
0.05
Tails
p0.0357
critical value±1.96
p < α → reject H₀ (significant result)

The p-value is NOT the probability that H₀ is true, nor does it measure the effect size. α=0.05 is just a convention. In ML it's used to validate whether a model improvement is real or noise.