MEPX
Chapter 2 of 10All chapters
  1. 1The method
  2. 2Designing an experiment
  3. 3When you cannot experiment
  4. 4Measurement
  5. 5Reading a result
  6. 6Hypotheses, theories and laws
  7. 7How science is published
  8. 8Replication
  9. 9Telling it apart
  10. 10Using it yourself

Chapter 2 of 10

Designing an experiment

Controls, randomisation and blinding.

The three pillars

A control group shows what happens anyway. Randomisation balances unknown differences. Blinding stops expectations from shifting the outcome on either side.

  • Without a control, recovery, seasons and attention all masquerade as effects.
  • Placebo effects are real and are exactly what a control measures.

Variables

Change one thing at a time, hold the rest fixed, and record what you held fixed. A study that varies two things cannot attribute the result to either.