The full loop: from question to revised theory

Observation, question, hypothesis, prediction, design, controls, measurement, analysis, peer criticism, replication, meta-analysis, revision.

It is a loop, not a ladder

Textbook diagrams end at 'conclusion'. Real science ends at 'revision' and starts again. A theory is a current best account with a known error bar, not a final verdict.

Controls are the whole trick

Any intervention produces change. The control group answers the only question that matters: change compared to what?

Where it breaks

Flexible analysis, undisclosed dropped conditions, small samples, and publishing only the results that worked. These produce literatures full of findings that dissolve on replication.