Lesson 2 — Why a Pattern Alone Does Not Prove a Cause

Summary

Learners go deeper than “correlation is not causation.” A pattern is only a starting point. To call something a cause, they check for a third variable, then ask for a mechanism (a “how”) and a fair test — and learn to stay open to new evidence. They run real claims through a four-step checklist.

Objectives

• Distinguish correlation from causation and explain why a pattern alone does not prove a cause. (D06.S1.07.01)

Connection

An umbrella shop notices that on the days it sells the most umbrellas, it rains the most. Do the umbrellas cause the rain? No — rain causes umbrella sales, and the pattern alone cannot tell you which way the cause runs, or whether a third thing (a coming storm) drives both. Every day, people make decisions — what to buy, what to believe, how to vote — on patterns just like this. Knowing what it would take to show a real cause is what protects you from being fooled.

Materials

• Causation-checklist page
• Claim cards
• Highlighters

Preparation

• Copy the causation-checklist page and claim cards.
• Have the worked example ready: “umbrella sales rise on rainy days” — the pattern does not prove umbrellas cause rain; the mechanism (rain → people buy umbrellas) and the order in time point the other way.
• Retrieval: recall from Lesson 1 what a third variable is and the ice-cream example.

Facilitator note

This lesson is written to the learner (“you”). The idea to land: a pattern is necessary but never sufficient — a cause also needs a plausible mechanism and, where possible, a fair test, and all causal claims stay revisable. Use explicit instruction with the umbrella worked example first, then guided practice on one claim card before independent work (Kirschner, Sweller & Clark, 2006). The checklist is a simplified, learner-friendly version of the “Hill criteria” for causation used in epidemiology (e.g., strength, consistency, plausible mechanism, temporality — which came first). Keep it simple: pattern, third variable, mechanism, test. The egalitarian lens (docs/philosophy.md §4): the demand for a mechanism and a test is what stopped false blame — for example, in old medicine, illnesses were blamed on a person’s character or group until careful tests found real microbial causes; the pattern “these people are sick” was real but the cause was mis-assigned. The global lens: different peoples developed careful causal reasoning independently — trial-and-error agriculture (which crop, which soil), traditional medicine, and navigation all demanded “what actually causes this?” — so the skill is a human universal, not one culture’s.

Procedure

1. Gather (5 min). Last time you met correlation and causation. Quick recall: what is a third variable? Say one example from memory.
2. The checklist (10 min). To decide if a pattern proves a cause, ask four questions in order: (1) Is there a real pattern? (2) Could a third thing cause both? (3) Is there a mechanism — a “how” that explains it? (4) Does a fair test (or many tests) show it, with a control and repeated trials?
3. Worked example (10 min). Claim: “Selling umbrellas causes rain.” (1) Yes, a pattern: umbrella sales and rain rise together. (2) Third thing? A coming storm causes both. (3) Mechanism? None — umbrellas have no way to change the weather. (4) Test? A fair test would fail. Verdict: correlation, not causation — and the cause actually runs the other way (rain → umbrella sales).
4. Run your own (20 min). Take your claim cards. For each, walk the four questions and write one line per step. Mark the verdict: pattern only or supported cause. Check with your partner — did you both check for a third thing and a mechanism?
5. Close (5 min). Remember: a cause needs more than a pattern. It needs a “how” and a test — and even then, science stays open to better evidence.

Differentiation

• Support: Provide the checklist as a sentence starter (“The pattern is ___. A third thing could be ___. The mechanism would be ___. A test would ___.”); work one card together.
• Extension: For one claim, weigh two competing explanations and name one piece of evidence that would decide between them.

Assessment

• Formative (observation/self): Can the learner run a claim through the four questions and explain why a pattern alone is insufficient?
• Self-check: The learner asks, “Did I check for a third variable AND a mechanism AND a test before calling anything a cause?”

Home connection

Ask someone at home about a rule or remedy they follow (“this makes my plants grow,” “this makes me sleepy”). Run the four questions together, kindly — and see whether it is a pattern or a supported cause.

Resources

• The idea that causation requires more than correlation (plausible mechanism, temporality, and, where possible, experimental evidence) underlies modern epidemiology and statistics; its classic statement is Sir Austin Bradford Hill’s “The Environment and Disease: Association or Causation?” (1965, Proceedings of the Royal Society of Medicine). The umbrella/rain example is a standard illustration of reverse causation. The value that demanding evidence protects people from false blame is a value commitment (docs/philosophy.md §5).
• On explicit instruction for novice reasoning skills: Kirschner, Sweller & Clark (2006), https://doi.org/10.1207/s15326985ep4102_1