Ask a room full of people which is more dangerous, flying or driving, and most now know the correct answer is driving. Ask them which they feel more anxious about, and the answers reverse.
Knowing a statistic and feeling a probability are different processes, and the second one runs on a shortcut that has nothing to do with statistics.
The shortcut
In the early 1970s Amos Tversky and Daniel Kahneman proposed that when people judge how likely or how frequent something is, they do not retrieve base rates. They ask themselves a much easier question: how easily can I think of an example?
If examples come quickly, the event feels common. If they come slowly, it feels rare. They called this the availability heuristic.
Most of the time it works. Genuinely common events do produce more memories than genuinely rare ones. The shortcut is fast, cheap, and usually approximately right.
It fails whenever something makes an event memorable for reasons unrelated to how often it happens.
The classic demonstration
Tversky and Kahneman asked participants whether more English words begin with the letter K or have K as their third letter. Most said the first.
The opposite is true — roughly twice as many words have K in third position. But retrieving words by their first letter is easy, and retrieving them by their third letter is hard. Ease of recall was mistaken for frequency.
A related study asked participants to estimate the relative frequency of various causes of death. Dramatic causes — tornadoes, homicide, accidents — were systematically overestimated, while quiet ones — stroke, diabetes, asthma — were underestimated. The pattern tracked media coverage rather closely.
Why this matters more now than it did in 1973
The availability heuristic evolved in an environment where the events you heard about were, roughly, events that happened near you. Frequency in your memory correlated with frequency in your world.
Modern media breaks that correlation completely. A news feed is, by construction, a selection of the most unusual events on the planet. Nothing is reported because it is typical.
Combine that with algorithmic ranking optimised for engagement — and unusual, frightening or outrageous content engages more reliably than ordinary content — and you get a machine that manufactures availability without any corresponding change in frequency.
This is why people's sense of how dangerous the world is, how common a particular crime is, or how prevalent an extreme opinion is, tends to be badly miscalibrated in a specific direction: toward whatever is most reportable.
Where it distorts everyday judgement
Risk perception. After a widely reported plane crash, air travel bookings drop and driving increases — a shift that, in the months following the September 11 attacks, was estimated to have caused additional road deaths well in excess of any plausible aviation risk. The behavioural response to a vivid risk created a larger statistical risk.
Hiring and management. A single memorable failure by an employee can outweigh a year of consistent competence, because the failure is more available at review time than the ordinary weeks.
Investing. Investors overweight recent market events. A crash three years ago feels more probable than a crash twelve years ago, though the market does not remember either.
Medicine. Doctors are trained to be aware of this. A physician who recently diagnosed a rare condition becomes measurably more likely to consider that condition in subsequent patients — sometimes usefully, sometimes not.
Personal relationships. Arguments are more available than pleasant evenings, because they carry emotional charge. Asked how a relationship is going, people recall the conflicts first, which is one reason keeping any kind of record of good days changes the answer.
Fear of flying, public speaking, spiders, and sharks. All well-documented, all overrepresented in vivid media, all statistically minor compared to the things people do not fear at all.
How to counteract it
Ask for the base rate, explicitly. When something feels likely, ask: out of every thousand people in this situation, how many does this actually happen to? The question forces a shift from recall to arithmetic. Even a rough guess is usually better calibrated than the feeling.
Notice the source of the example. If the reason an event feels common is that you saw three stories about it this week, that is information about the news cycle, not about the world. The correct follow-up question is whether the underlying rate changed or the coverage did.
Prefer denominators to anecdotes. "Three people were harmed by this" means nothing without knowing how many were exposed. Media reporting almost never supplies the denominator, because the denominator is what makes a story boring.
Keep written records for decisions you make repeatedly. Managers who keep brief notes on performance throughout the year give demonstrably different reviews from those relying on memory, because the notes provide a sample rather than a highlight reel.
Deliberately seek disconfirming examples. If you believe a thing happens often, spend two minutes trying to recall instances where it did not. The effort of retrieval is itself informative — if counterexamples are also easy to find, your original estimate was built on availability rather than frequency.
A caveat worth keeping
Availability is not always misleading. Sometimes the reason an example comes easily is that the event genuinely is common, and dismissing your intuition on principle is its own error.
The heuristic is a problem specifically when memorability and frequency have been decoupled — by media, by emotional intensity, by recency, or by the vividness of a single case. Those are the conditions to watch for, not the shortcut itself.
Summary
Your brain estimates probability by measuring how easily examples come to mind, and then reports the result as a feeling of likelihood. The method is efficient and mostly reliable in a world where what you hear about reflects what happens around you.
You do not live in that world any more. What reaches you is filtered for how unusual and how emotionally striking it is, which means your sense of what is common is being shaped by exactly the wrong variable.
The correction is simple to state and hard to remember: when something feels likely, ask how often it actually happens — and notice whether your evidence is a statistic or a story you saw.
Be the first to share a thought.