Section 023 · Chapter 4, Statistical Tests for AI Quality

P-Values: Evidence, Not Permission

P-values can support a comparison, but they do not decide whether a product is safe, useful, or worth shipping.

p-valuep values evidence permission

What to do

  1. Version B reduced automatic refunds from 18% of conversations to 14%, and the p-value is 0.003.
  2. Define runnable checks that exercise p-value and p values evidence permission.
  3. Set acceptable outcomes and blocker failures for p-value and p values evidence permission before running the evaluation.

Evidence to preserve

  • Preserve the inputs, versions, configurations, raw outcomes, and results for p-value, p values evidence permission needed to reproduce work on P-Values: Evidence, Not Permission.
  • Report results for p-value, p values evidence permission by relevant slice, separate blocker failures from averages, state uncertainty and blind spots, and connect the result to a release decision.

Expert note

Expert reports treat p-values as continuous evidence, not as a yes/no permission slip. A useful report pairs the p-value with the other facts needed to make a responsible decision:

Continue the conversation

Apply this to your context.

Save your product context once, then open a focused conversation that combines it with this concept.

Cite this page

Jason Arbon. "P-Values: Evidence, Not Permission." Testing AI Knowledge Edition, section 23.

https://jarbon.ai/testing-ai/knowledge/ch023-p-values-evidence-permission.html

Shared across the Knowledge Edition

Adapt every concept to your world.

Describe your product, role, users, risks, constraints, or current quality problem. This stays in this browser until you choose to send it to ChatGPT.

Saved only in this browser.0 / 2400