Understanding p-Values

When your producing trend lines in Tableau or building models in Alteryx (or really any type of statistics), you'll have no doubt seen a p-value or at least seen something that looks like "p < 0.005".

These are p-values. p-values tell us the probability of getting a certain observation, given the data.

Most of the time we'll use p-values to tell us if our results are statistically significant by comparing it to our significance level, α (alpha).

If our p-value is less than (<) α, we can describe the result as statistically significant. In most cases, α is set at 0.05 (or 5%), but this will differ depending on what you are testing.

The most traditional use of a p-value is to test a null hypothesis, H₀. A null hypothesis states there is no relationship between the variables being studied.

If our p-value is found to be significant (< α), we will reject the null hypothesis in favour for our alternative hypothesis, Hₐ. Otherwise we do not reject H₀.

Importantly when using p values, a statistically significant result cannot prove our alternative hypothesis is correct (as this implies 100% certainty). Instead we say that it give 'support' or 'evidence' for out alternative hypothesis.

Author:
Jacob Kilroy
Powered by The Information Lab
1st Floor, 25 Watling Street, London, EC4M 9BR
Subscribe
to our Newsletter
Get the lastest news about The Data School and application tips
Subscribe now
© 2025 The Information Lab