The Four Pitfalls of Induction
What is The Four Pitfalls of Induction?
Induction looks for what several specific facts have in common and forms a judgement over a wider range; the conclusion is only probable. First principle: the conclusion cannot travel further than the evidence — a handful of cases will not carry “always”, “never”, or “everyone”. Four pitfalls. (1) Too small a sample — cheated once, so nobody can be trusted; one bad boss, so all bosses are bad; one poor exam, so the child will never amount to anything. (2) A biased sample — interviewing only the customers who left, gathering information only while upset, picking only the facts that match your current mood. (3) Mistaking a short-term swing for a long-term trend — one large order or one delayed purchase read as a direction, with no span of time to check it against. (4) Collecting support and never counterexamples — the valuable question is where the counterexamples are, because they tell you where the pattern stops holding.
The Four Pitfalls of Induction: common mistakes and how to handle them
“Turnover is high among younger staff, so hire fewer of them.” On inspection the departures clustered under one manager and within the first three months — the real pattern was that unclear goals and thin support drive people out, and age was standing in for it.
What the The Four Pitfalls of Induction questions test
identifying which of the four pitfalls a passage falls into; judging whether a sample can carry the conclusion; pointing at the part of a conclusion that outruns the evidence (the absolute wording); choosing the counterexample or extra sample that would test where the pattern stops.