MEPX
Chapter 10 of 10All chapters
  1. 1What correlation measures
  2. 2The five explanations
  3. 3Confounders
  4. 4Experiments
  5. 5Natural experiments
  6. 6Weighing causal evidence
  7. 7Selection effects
  8. 8Regression to the mean
  9. 9Language and reporting
  10. 10Acting under uncertainty

Chapter 10 of 10

Acting under uncertainty

Decisions cannot always wait.

Proportionate action

Sometimes you must decide before causation is established. The question becomes the cost of acting wrongly against the cost of waiting, which is a judgement not a statistic.

  • Reversible low-cost actions justify weaker evidence.
  • Irreversible expensive ones deserve to wait for stronger.

In your own work

Product changes, process changes and personal experiments all deserve a comparison group where possible. Running a genuine test is usually cheaper than arguing about the result for a year.