Strip the narrative and restate as a bare frequency claim
Translate the vivid scenario into a dry statistical question to see if it still feels probable.
Key takeaways
- What it is: Translate the vivid scenario into a dry statistical question to see if it still feels probable.
- Why it works: Rich narratives engage the brain’s pattern-matching and story-comprehension systems, which are optimized for coherence and causal plausibility — not probability. Translating "Linda is a feminist bank teller" into "what percentage of bank tellers are feminist activists?" moves the question from the associative system to a more calibrated numerical frame. Frequency format is consistently easier for people to reason about correctly than probability format, as Gigerenzen’s research on natural frequencies demonstrates.
- Evidence: Backed by randomized trials / meta-analyses.
- Avoid: Translating to frequency format but keeping the narrative framing intact ("how many people like Linda…"), which still activates representativeness — the restatement must be genuinely dry.
Why it works
Rich narratives engage the brain’s pattern-matching and story-comprehension systems, which are optimized for coherence and causal plausibility — not probability. Translating "Linda is a feminist bank teller" into "what percentage of bank tellers are feminist activists?" moves the question from the associative system to a more calibrated numerical frame. Frequency format is consistently easier for people to reason about correctly than probability format, as Gigerenzen’s research on natural frequencies demonstrates.
How to do it
- 1Restate the claim in frequency form: "Out of 100 people with this background, how many would have both of these properties?"
- 2Compare that estimate to the frequency of just one of the properties in the same group.
- 3If the conjunction rate is higher than either component rate, you’ve caught a fallacy.
- 4Use this restatement before making any decision that depends on the joint scenario being likely.
What the evidence says
RCT / meta-analysisGigerenzen and Hoffrage (1995) showed that natural frequency formats dramatically improve probabilistic reasoning accuracy compared to probability percentages — the conjunction fallacy rate drops substantially in frequency format.
Honest caveat: Frequency framing reduces but does not eliminate the fallacy; some individuals still commit it in frequency format, especially under time pressure.
- — Gigerenzen & Hoffrage (1995), How to improve Bayesian reasoning without instruction, Psychological Review
Common mistake
Translating to frequency format but keeping the narrative framing intact ("how many people like Linda…"), which still activates representativeness — the restatement must be genuinely dry.
IX Coach reformats probability claims into frequency questions during goal planning, so assessments of what "might happen" stay grounded in how often such things actually occur.
Practice this with IX Coach →