Evidence (Evaluating What's True)

A hierarchy for evaluating the reliability of evidence: randomized controlled experiments provide the strongest evidence, followed by large observational studies, small studies, case reports, expert opinion, and finally anecdote. Most business decisions rely on the weakest evidence types—anecdotes and expert opinion—while ignoring or not seeking the stronger types. Understanding this hierarchy doesn't make all evidence equal; it helps you calibrate confidence appropriately.

When to use it

When evaluating any factual claim—business case, market research, health advice, investment thesis; when deciding how much confidence to place in a recommendation; when designing research or experiments to test your own hypotheses; when encountering persuasive stories that 'prove' a point.

How it can help

Before accepting any claim, ask: what's the evidence, and where does it fall on the hierarchy? A single customer testimonial (anecdote) should carry far less weight than a properly controlled A/B test (experiment). Expert opinion without supporting data is just one person's pattern-matching. This model prevents the common error of treating vivid stories as proof while ignoring boring but reliable statistics.