Scientific Testing

Scientific testing is the practice of trying to refute a hypothesis by experiment or observation. The aim isn't to confirm; it's to break.

CF / CR's view of testing

• Test to break, not to confirm. A test that can't fail tells you nothing.
• One refutation is decisive. A failed test refutes the hypothesis (assuming the test itself is sound).
• Passing tests don't confirm. A million passed tests still leave open the next refutation.

What makes a test good

1. It could in principle fail. Otherwise it's not a test.
2. The result is observable. You can tell whether it failed.
3. The experiment is reproducible. Others can run the same test.
4. It targets the hypothesis specifically. It isolates the relevant variable.

What CF adds

CF emphasises that:

• A test is itself an IGC — it can be wrong, biased, or poorly designed.
• "The experiment failed" is a non-refuted IGC only if the experimental setup is sound.
• Test results should be public so others can criticise them.

"An important form of criticism is scientific testing. Hypotheses which disagree with experiment are wrong." — criticalfallibilism.com