Bias · what comes to mind first · 7 min read

The availability heuristic, and what comes to mind first

Tversky and Kahneman's letter-frequency study, the lethal-events survey it inspired, the ease-of-retrieval twist, and three scenarios to try the rule on.

In short

The availability heuristic is judging how common or likely something is by how easily examples of it come to mind, rather than by how often it actually happens. Amos Tversky and Daniel Kahneman named it in 1973 and showed it with a simple, testable case: people judge words starting with a given letter to be more common than words with that letter third, even when the opposite is true, because words are easier to retrieve by their first letter. Later work found the heuristic runs on the felt ease of recall itself, not only on how many examples you can list, which can make it work against you exactly when you try hardest to think of more.

An everyday example

After a week when a shark attack makes national news, more people say they are afraid to swim in the ocean, even though nothing about the actual rate of shark attacks changed that week. What changed is how easily a shark attack comes to mind: a vivid story, freshly seen, is far easier to recall than the ordinary, unreported fact that drowning from a rip current is a far more common way to die at the beach.

The mind is not lying to itself on purpose. It is using a real signal, how readily something comes to mind, as a stand-in for a different question, how often that thing actually happens, and the two only agree by coincidence.

The classic experiments

Amos Tversky and Daniel Kahneman's 1973 paper introduced the availability heuristic with a study built to separate true frequency from ease of retrieval. English has more words with the letter K as the third letter than as the first letter, but most people, asked to judge which is more common, say the opposite: more words start with K. Tversky and Kahneman argued this is because people search their memory for words by trying to recall the first letter, a search strategy that works well for K-first words and poorly for K-third words, so K-first words feel far more available even though they are, in fact, rarer. They found the same reversed judgement for several other consonants that behave the same way in English.

Lichtenstein, Slovic, Fischhoff, Layman and Combs extended the idea to judgements of risk in a 1978 study. Participants estimated how many people die each year from a range of causes, from tornadoes and botulism to diabetes and stroke. Dramatic, widely reported causes of death were consistently overestimated relative to their true toll, and common, undramatic causes were consistently underestimated. In one of the paper's most quoted comparisons, participants judged accidental deaths and deaths from disease to be roughly equally frequent, when disease is in reality by far the larger cause of death; accidents simply generate more memorable, more widely reported stories.

It is the ease, not just the amount

Norbert Schwarz and colleagues sharpened the mechanism in a 1991 study with a twist that the simple "more examples equals higher judgement" story does not predict. Participants were asked to recall either six or twelve times they had behaved assertively, then rated how assertive they considered themselves. The twelve-examples group had, by definition, generated twice as many actual instances of assertive behaviour. They rated themselves as less assertive than the six-examples group did. Recalling twelve examples was noticeably harder than recalling six, and that struggle was read as information: if assertive examples are this hard to find, I must not be very assertive, regardless of how many were actually listed by the end. The felt difficulty of the search overruled the content the search turned up.

Does it replicate?

Replication grade: Strong: reproduced across many judged categories; the ease-versus-content split is still studied

The basic letter-frequency demonstration and the pattern of over- and under-estimating causes of death have been reproduced in many forms since the 1970s, across different judged categories and different populations, and the general finding that memorable, vivid or recent cases inflate frequency estimates is one of the more consistently observed patterns in judgement research. Schwarz's ease-of-retrieval studies have also been repeated with variations in the number of examples requested, and the reversal, where asking for more examples produces lower confidence because the extra examples are harder to generate, shows up reliably within the range of set sizes those studies used.

What is debated is not whether availability affects judgement, but exactly which ingredient is doing the work in a given case: sheer content, the ease or difficulty of retrieval, or both, and how large a role each plays depends on the specific task. Schwarz's own later work found that ease of retrieval matters more when people are not otherwise motivated to think the judgement through carefully, and matters less when the content itself is what people are focused on.

Two stacks, one taller

An illustration, not measured frequencies. The left stack is topped by vivid, easily pictured cases; the right stack holds duller, more ordinary cases further down. Judging by how easily examples come to mind tracks vividness and recency as much as it tracks true frequency.

The diagram on this page is an illustration, not a measurement of real frequencies. It draws two stacks of examples: a short stack of vivid, easily pictured cases sitting on top, and a taller stack of ordinary, easily forgotten cases underneath. The point of drawing the vivid stack as shorter is that availability tracks how easy something is to reach, not how many of them there really are.

Try it: three scenarios

Three short situations that turn on how easily something comes to mind, rather than on how often it actually happens.

1. After a week of vivid news coverage of a plane crash, a person judges flying to be riskier than driving the same distance. What does this most directly illustrate?

That vivid, recent coverage makes an event easier to recall, which can be mistaken for the event being more frequentThat flying has just become statistically more dangerousThat the person has calculated the actual accident rates for both

Why

Nothing about the true frequency of plane or car accidents changed that week; only how easily a plane crash could be recalled changed, and that ease was read as information about likelihood.

2. Asked to name five friends who exercise regularly, most people answer quickly and confidently. Asked to name fifteen, they struggle and often report feeling less confident that exercising regularly is common among their friends. What does the second part show?

That people have fewer friends than they thoughtThat the difficulty of the second recall task was itself read as information, on top of whatever was actually recalledThat fifteen is an unreasonable number to ask for

Why

The struggle to reach fifteen names is itself a signal, and people weigh that felt difficulty alongside, or even instead of, the actual count of names they produced.

3. A company's safety report lists two rare but dramatic incidents from the news in detail, and a common, undramatic cause of workplace injury in one line with no story attached. Why might readers misjudge which risk matters most?

Because the dramatic incidents are easier to picture and recall, regardless of how often each type of incident actually happensBecause the report's authors made a factual error in the numbersBecause rare incidents are always more dangerous than common ones

Why

A story with vivid detail is more available in memory than a one-line statistic, which can make the rarer, more memorable risk feel larger than the risk that actually causes more harm.

Nothing you type on this page leaves your browser tab: no answer, guess or score is sent anywhere or stored.

How to catch it

Availability is hardest to notice from the inside, because the feeling of "I can think of several" or "nothing comes to mind" feels like a fact about the world rather than a fact about your own memory search.

When a risk feels large because of a recent story, ask for a base rate: how often does this actually happen, out of how many opportunities, over what period.
Notice when a judgement is driven by one or two vivid cases rather than by any actual count.
If a task asks you to generate a long list and it starts feeling hard, remember that the difficulty itself is not evidence about the answer; it may only be evidence that the list is long.
Ask what kind of event would be reported or remembered even if it were rare, and discount accordingly: dramatic, unusual and recent events all have an availability advantage that has nothing to do with frequency.
Where real numbers exist, such as public health or safety statistics, look them up rather than estimating from memory.

Check yourself: three questions

1. In Tversky and Kahneman's 1973 letter study, why did most people wrongly judge that more English words start with K than have K as the third letter?

Because people search memory for words mainly by first letter, which makes K-first words easier to retrieve even though K-third words are more numerousBecause K-first words are genuinely more numerousBecause participants were only shown examples of K-first words

Why

The paper's point was exactly this mismatch: retrieval by first letter is an easy search strategy, so K-first words felt more available, even though English actually has more words with K in the third position.

2. In Lichtenstein and colleagues' 1978 study, how did participants compare accidental deaths and deaths from disease?

They judged the two roughly equally frequent, though disease deaths far outnumber accidental deathsThey correctly judged disease deaths as far more frequentThey judged accidents as rare and disease as common by an even larger margin than reality

Why

Accidents generate more vivid, widely reported stories than most disease deaths, and the study found participants judged the two categories as roughly equally frequent despite the large real gap between them.

3. In Schwarz and colleagues' 1991 study, why did the group asked to recall twelve assertive examples rate themselves as LESS assertive than the group asked to recall only six?

Because the difficulty of recalling twelve examples was itself read as information, overriding the fact that twelve is more content than sixBecause the twelve-example group was less assertive to begin withBecause they were given less time to answer

Why

The twelve-example group generated more actual instances but found the task harder, and that felt difficulty was used as a signal of low assertiveness, outweighing the larger raw count of examples.

Sources

Text on this page is original to MyTestAtlas, written from the studies listed. The diagram is drawn by this site and is not a copy of any published figure.