Physics
Models tested against measurement, with error accounted for.
What counts as evidence: Quantitative measurement or simulation, compared against a stated theoretical prediction.
- State the prediction you're testing before collecting data, not after.
- Report and propagate measurement uncertainty — a result without error bars isn't complete.
- Simple setups (pendulums, circuits, optics) can still produce genuine, checkable results.
- Simulation and modeling are legitimate substitutes when equipment isn't available.