Question
What is disconfirming evidence?
Quick Answer
Actively looking for evidence against your current belief is the fastest path to calibration.
Example: You believe your startup's churn is caused by poor onboarding. Instead of gathering more data about onboarding friction (which would confirm what you already suspect), you deliberately investigate the three strongest alternative explanations: pricing misalignment, feature gaps, and support response time. You discover that 68% of churned users never opened a support ticket — they left because a competitor launched a feature you deprioritized. The disconfirming evidence didn't just correct your belief. It changed your roadmap.
This concept is part of Perceptual Calibration in the How to Think curriculum, which builds the epistemic infrastructure for perceptual calibration.
Learn more in these lessons
- Seek disconfirming evidence
- Distinguish validation from confirmation
Looking for evidence that supports your schema is not the same as rigorously testing it.
Go beyond the answer
This topic is part of The Staff Engineer’s Thinking Infrastructure — a guided sequence of 20 lessons.
After completing this path you will be able to consciously choose your level of abstraction, build shared mental models across teams through ADRs and RFCs, make architecture decisions that persist beyond your involvement, structure your own time in an undefined role, and manage the invisible emotional load of technical leadership.
Start the path →