Trap/Cognitive Bias/No. 0775
Probability Neglect
Probability neglect is giving too little weight to an outcome’s likelihood when strong emotions draw attention to its consequences. Named by Cass Sunstein in 2002, it describes reduced sensitivity to odds, rather than necessarily ignoring them entirely, and differs from risk aversion.
- Evidence
- Well established
- Read
- 6 min
- Links
- 9 connections
01You've seen this when…
- in life
You hear about a shark attack and drop the beach trip. Learning that the risk is far lower than you assumed barely changes how dangerous swimming feels.
- at work
A manager watches a demonstration of a catastrophic server failure. The next budget goes toward preventing that scenario; its likelihood never makes the slide.
- out in the world
A council debates a new chemical plant. Speakers describe a toxic leak in detail, while estimates of how often such leaks occur barely enter the discussion.
02The idea
Once you can picture the disaster, the picture can become the argument. A plane goes down. A child gets poisoned. A company loses every customer record. The consequences feel so important that their likelihood slips out of the decision.
Probability neglect is giving too little weight to how likely an outcome is, especially when the outcome carries strong emotion. It can occur even when you understand the numbers. You might understand the difference between a 1% chance and a 99% chance, yet show surprisingly little difference in worry or willingness to pay for protection.
Severity and probability both belong in the decision. The mistake is allowing severity to crowd out probability.
This differs from risk aversion: someone can consider the odds carefully and still pay extra to avoid a loss. It also differs from probability weighting, where choices reflect probabilities weighted out of proportion to their numerical size. The ideas can overlap; neglect emphasizes unusually weak sensitivity to changes in likelihood.
03Why it happens
- The outcome is easier to imagine than its frequency. A burning building is a scene. A one-in-several-thousand annual probability is not. Salience gives the scene more attention than the number.
- Feeling supplies a fast answer. Fear can answer how dangerous something seems before you have worked through how often it happens. That is closely related to the affect heuristic: judging through the feeling an option produces.
- Possibility can become the stopping point. Once a frightening event moves from impossible to possible, people may stop distinguishing among very different probabilities. The fact that it could happen becomes enough.
- Precaution feels easier to defend than a trade-off. Paying to prevent a vivid disaster can seem responsible without further explanation. Spending that money elsewhere requires comparing risks, including less memorable harms.
Strong emotion can sometimes increase attention to probability: fear can motivate careful investigation. The warning sign is a judgment that barely changes when credible likelihood estimates change substantially.
04A worked example
In a 2001 experiment, psychologists Yuval Rottenstreich and Christopher Hsee asked participants how much they would pay to avoid either a monetary loss or a brief, painful electric shock. The stated probability was either 1% or 99%.
What it looks like People are pricing protection against two unpleasant outcomes. For either outcome, protection against an almost certain event should generally be worth more than protection against a remote possibility.
What’s actually going on Stated willingness to pay changed much less between the low and high probabilities for the electric shock than for the monetary loss. The emotionally charged outcome made the difference in likelihood matter less. Participants still gave probability some weight in their valuations.
What would have helped A useful decision check would separate the two judgments: first describe the harm, then compare protection against that same harm at different probabilities. Holding the outcome fixed makes the missing distinction easier to notice. This is a proposed safeguard, not a remedy tested in that experiment.
The study measured stated valuations. It did not show what people would actually spend during an emergency, or establish that every frightening risk produces the same response.
05How to spot it
A useful diagnostic is to hold the outcome constant and vary only its probability. Little movement in judgment is only a clue: a safety requirement or an unacceptable loss could also explain it.
06What to do instead
- Write severity and likelihood separately. Describe what would happen, then estimate how often it happens. An elaborate description of harm still needs a likelihood estimate.
- Use a shared denominator and time period. Compare events per 10,000 exposures, or per year. Put lifetime risk and risk from one trip on the same scale before comparing them. A probability without an exposure period can mislead.
- Look outside the story. Use a reference class: similar journeys, facilities, treatments or projects. Ask what happened across those cases to put the memorable one in context.
- Compare the protection with alternatives. Estimate how much risk each option removes for its cost. For repeatable financial losses, expected value offers a starting point. For safety decisions, also consider severity, distribution of harm and binding constraints.
- Test whether the odds affect your choice. Reassess the same outcome at the low and high ends of a credible probability range. If your choice stays identical, identify why: a firm constraint, an affordable precaution, or a reaction to the image itself.
- Keep uncertainty visible. If the likelihood is poorly known, record a range and the evidence behind it. Distinguish uncertainty from measurable risk. Treat missing information as a reason to withhold both reassurance and certainty of disaster.
07When it isn’t probability neglect
A rare danger can deserve substantial attention. Insuring your home may be sensible because losing it would overwhelm your finances. Avoiding even a small risk of ruin can be rational when you cannot recover from the loss.
Likewise, a tiny risk repeated across millions of exposures can produce many casualties. The total burden depends on both the probability per event and the number of exposures.
Costs matter too. A cheap precaution may be worthwhile across a wide range of probabilities, so an unchanged decision can coexist with changed reasoning. Ethical duties and safety standards can also rule out choices that a simple average would favor.
Probability neglect is therefore identified by missing or weak use of likelihood. A different tolerance for danger can explain why someone is more cautious than you. Even when you recognize probability neglect, examine frightened people’s evidence before dismissing their concerns.
08Roots
In 2002, legal scholar Cass Sunstein brought research about electric shocks into the Yale Law Journal. His concern was bigger than a laboratory choice: how should law respond when a frightening possibility dominates public attention? Terrorism and toxic exposures made the problem concrete. The harm could be appalling, and deciding how much to spend preventing it required considering its likelihood.
Sunstein named the pattern probability neglect. He drew on earlier psychological work, including Rottenstreich and Hsee’s experiments and the account of risk as feelings developed by George Loewenstein and colleagues. The name connected a laboratory finding about sensitivity to odds with a practical question for policy: whether public concern reflects likelihood, consequences, or the emotional force of an imagined event.
An adaptive explanation is possible, but speculative. Rapidly avoiding a frightening threat might sometimes be safer than pausing to calculate, especially when information is poor and the cost of escape is low. Evidence would be needed to establish probability neglect as an evolved adaptation. A response that helps in one setting can become costly when it directs a household budget or a national safety program.
09How solid is this?
Experiments support a limited finding: reduced sensitivity to probability for some emotionally charged outcomes. Claims of universal or complete disregard of odds go beyond that evidence. The strength depends on the task and framing; conclusions about real-world spending require evidence beyond stated willingness to pay.
10Connections
- Often confused with Risk Aversion
- Countered by Expected Value, Reference-Class Forecasting
- Part of Affect Heuristic, Salience Bias, Probability Weighting
- See also Risk vs. Uncertainty, Risk of Ruin, Zero-Risk Bias
11Origin and sources
Named by Cass Sunstein in 2002, drawing on earlier experimental research into emotional responses to risky outcomes.
- [1]Sunstein, C. R. (2002). Probability Neglect: Emotions, Worst Cases, and Law. Yale Law Journal, 112(1), 61–107.
- [2]Rottenstreich, Y., & Hsee, C. K. (2001). Money, Kisses, and Electric Shocks: On the Affective Psychology of Risk. Psychological Science, 12(3), 185–190.
- [3]Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch, N. (2001). Risk as feelings. Psychological Bulletin, 127(2), 267–286.
Suggest an edit· Updated 2026-10-02