LogicalMANIPULATIVE

Conjunction Fallacy

What it is

Adding specific, plausible details to a claim so that the more detailed version feels more probable than the simpler claim it contains, when every added detail can only make it less likely.

How it works

A conjunction cannot be more probable than either of its parts: “bank teller and active feminist” is a subset of “bank teller”. Yet when Tversky and Kahneman described Linda — thirty-one, single, outspoken, a philosophy graduate concerned with social justice — about 85 percent of respondents ranked the conjunction as more likely than the single category. In a 1982 study of professional forecasters, “Russia invades Poland and the United States suspends diplomatic relations” was judged more likely than the suspension alone. The lever is representativeness: the detailed version matches the picture, and a coherent story feels like a probable one. Worked: a scenario with three independent details each 80 percent likely is, taken together, about 51 percent likely; add two more and it is under a third, while feeling more convincing at every step. Hertwig and Gigerenzer showed the effect shrinks when the question is framed as frequencies, which is useful for defense but does not save the audience of a pitch. The persuasion move is to load a forecast, a sales scenario, a prosecution narrative or a conspiracy theory with specifics, since each one raises felt plausibility while lowering actual probability. Insurance buyers pay more to cover death from a named cause than from any cause, which includes it.

Real-world examples

  • •Tversky and Kahneman (1983) reported that about 85 percent of respondents judged Linda more likely to be a bank teller active in the feminist movement than to be a bank teller, across student and expert samples.
  • •In the same paper, forecasters at a 1982 conference rated a scenario in which a Soviet invasion of Poland led to a suspension of US-Soviet relations as more probable than a suspension of relations from any cause.
  • •Johnson and colleagues (1993) found people would pay more for flight insurance covering death from terrorism than for insurance covering death from any cause, although the second policy includes the first.
  • •Investment pitches and forecasts sell a scenario — the merger closes, the new chief executive delivers, regulators approve, the market re-rates the sector — whose richness is what persuades and whose four conditions must all hold.
  • •Conspiracy narratives accumulate specific claims about names, dates and mechanisms; each addition makes the account feel more detailed and therefore more credible to an audience, while every one is another thing that must be true.

Ethical guidelines

  • ●When you present a scenario, state that its probability is at most that of its least likely component, and give that estimate.
  • ●Do not use detail as a substitute for evidence; specifics that cannot be checked add plausibility and nothing else.
  • ●In forecasting and risk work, present the simple event alongside the elaborated scenario so the audience can see that the elaboration is a subset.
  • ●Prefer frequency framing (“out of 100 people like this…”) when you want the audience to judge accurately rather than to be moved.

How to defend against it

  • ►Strip the story to its core claim and ask how likely that is on its own; every added detail can only lower the total, so the core is the ceiling.
  • ►Count the conditions that must all be true and multiply roughly; five things at 80 percent each are together about one in three.
  • ►Treat unusual specificity as a cue for skepticism rather than credibility: ask which of the details has been independently verified.
  • ►Reframe as frequencies — “of 100 people described like this, how many are both?” — which makes the subset relationship visible.
  • ►For insurance, forecasts and contracts, ask whether the broad case covers the narrow one you are being sold; if so, pay for the broad one.

From the Defense Playbook

The Base-Rate Checkminutes

Before accepting what a test result, a profile, or a striking detail seems to prove, ask how common the thing is in the first place, because evidence that sounds accurate can still be wrong most of the time when the underlying condition is rare.

"Which Step, Specifically?"minutes

When told that one action will lead through a chain of events to a dramatic outcome, lay the chain out link by link, ask what makes each link likely, and remember that the probability of the whole chain is lower than that of its weakest link.

Every playbook entry states how strong its evidence is and when not to use it. Browse the full playbook.

References

  1. Tversky, A., & Kahneman, D. (1983). Extensional versus intuitive reasoning: The conjunction fallacy in probability judgment. Psychological Review, 90(4), 293-315 · link
    The Linda problem, the forecasting scenarios and the 85 percent figure.
  2. Hertwig, R., & Gigerenzer, G. (1999). The “conjunction fallacy” revisited: How intelligent inferences look like reasoning errors. Journal of Behavioral Decision Making, 12(4), 275-305
    The critique that “probable” is read as “plausible”, and the reduction of the effect under frequency framing.
  3. Johnson, E. J., Hershey, J., Meszaros, J., & Kunreuther, H. (1993). Framing, probability distortions, and insurance decisions. Journal of Risk and Uncertainty, 7(1), 35-51
    Higher willingness to pay for insurance against a vividly named cause than against any cause including it.
Last reviewed
Suggest a correction

Detect Conjunction Fallacy in any text

Paste any message, email, or article into our free Manipulation Detector to see if Conjunction Fallacy or other techniques are being used on you.

Try the Detector FreeCreate Free Account