"The normal-distribution trick is not something you carry into a lab. It is the lens you carry into a Tuesday afternoon."
The Quest Was Practice; Daily Life Is the Game
Nine tracks of derivation, examples, and dismantling. The point was never to convert the citizen into a statistician — it was to install a lens that operates in daily life, automatically, on whatever statistical claim the world presents. This epilogue track is the operating manual.
The Five Reflexes
After completing this quest, the citizen-statistician should reach for five reflexes:
- 'What shape is the distribution?' Before responding to any 'average X,' inspect the distribution, population, and purpose. Mean, median, and percentiles answer different questions; real data need not be either normal or a pure power law.
- 'What is the prior?' Before turning a likelihood, test result, or DNA statistic into a posterior claim, ask for the relevant base rate and the full evidence model. A p-value is not a likelihood and need not be converted into a posterior to be used correctly.
- 'What is missing from this sample?' Before drawing inference from any data, ask who was excluded by the sampling process. Self-selection, survivorship, non-response, geographic filtering — name the filter.
- 'Is this regression to the mean?' Before attributing 'decline' or 'slump' to character, ask whether the previous extreme was partly noise that simply did not repeat. Include regression as a candidate when selection was based on an extreme, then compare causal explanations with an appropriate design.
- 'Am I weighting confirming and disconfirming evidence symmetrically?' Confirmation bias is reliable; the corrective is deliberate symmetry. What evidence would, if observed, reduce your confidence? Go look for it.
The Compound Effect
Each reflex on its own catches a fraction of citizen-statistical errors. Together they catch many common errors, though no checklist replaces domain knowledge or validation. The journalism, advertising, politics, and casual conversation you encounter is shot through with statistical claims that fail at least one of these reflexes. Picking them off routinely is the lens-skill in action.
The lens does not make you cynical or paralyzed. It makes you calibrated. You can still believe statistics; you just believe them with the right weight, after the right questions, and with the right humility about what is hidden. The citizen-statistician is not the person who rejects all numbers; it is the person who reads numbers honestly.