Last lesson handed you a slogan — what comes to mind feels true — and a one-line version of the model: you judge how likely something is by how easily an example pops up. Slogans are great for remembering and useless for thinking. A slogan can’t tell you exactly what’s being measured, why the measurement usually works, or when it betrays you. This lesson turns the slogan into a tool. We’ll get the precise definition from the people who discovered it, dig out the deeper machine humming underneath it, and run the actual experiments that pinned it to the wall — including the one where almost everybody confidently believes a thing that is provably false about the letter K.
By the end you should be able to do something specific: catch yourself in the exact moment the swap happens — the instant your mind quietly trades a hard question for an easy one and hopes you won’t notice.
Guess before you peek — especially here
There’s a question about the letter K coming up that most people get wrong with total confidence. The wrong answer feels obviously right, which is precisely the lesson. When you hit it, commit to an answer out loud before reading on. The sting of “wait, really?” is the whole point.
The precise definition
Start with the analogy. Imagine a warehouse with a single, very lazy stockboy. You ask him, “Do we have a lot of red umbrellas?” He doesn’t walk the aisles and count. He just notices how fast a red umbrella springs to his memory. If three of them leap to mind instantly, he says “tons of them.” If he has to strain to recall even one, he says “barely any.” He’s not measuring the stock. He’s measuring how easy it was to think of the stock — and reporting that as if it were the same number.
That lazy stockboy is your brain, and the habit has a name. The availability heuristic is the mental rule of thumb of estimating the probability or frequency of an event by how easily instances or examples of it come to mind. It was named and dissected by psychologists Amos Tversky and Daniel Kahneman in their 1973 paper, “Availability: A heuristic for judging frequency and probability.” (A heuristic, recall, is just a fast, good-enough shortcut your mind uses in place of a full calculation.)
Now the single most important distinction in this whole lesson, because nearly everyone gets it slightly wrong: availability is about ease of retrieval, not about how much you know. The question your mind secretly answers is not “how much information do I have about X?” It’s “how fluently, how effortlessly, did an example of X arrive?” Those come apart all the time. You might know, on paper, that car crashes kill vastly more people than plane crashes — yet right after watching footage of a crash landing, plane death feels more available, more summonable, and so it feels more likely. Knowledge sat in one drawer; the heuristic reached into a different drawer entirely and grabbed the feeling of fluency.
The definition, stripped to its parts
Hard question being asked: how frequent / how probable is X? • What the mind actually measures: how easily an example of X comes to mind. • The move: report the second as if it answered the first. The whole heuristic lives in that substitution — and the next section gives the substitution its proper name.
Before you read — take a guess
Two people are asked, 'Are tornado deaths or asthma deaths more common in the US?' Person A recently saw dramatic tornado footage and says 'tornadoes, obviously.' (In reality asthma kills far more people.) Which statement best captures what the availability heuristic actually measured in Person A's head?
Attribute substitution — the engine underneath
So far availability looks like a standalone quirk. It isn’t. It’s one instance of a much bigger, deeper mechanism that Kahneman and his colleague Shane Frederick named attribute substitution — and once you see it, you’ll spot it everywhere.
Here’s the machine. Your mind is handed a target question that is genuinely hard to answer: How probable is X? How frequent? How risky? Answering it properly would mean gathering data, recalling base rates, doing arithmetic — slow, effortful work. So the mind does something sneaky: it silently swaps in a different, easier question — a heuristic attribute that happens to be quick to evaluate — answers that one instead, and then hands you the answer as though you’d answered the original. It never flags the substitution. You feel like you assessed the risk; you actually assessed how easily an example came to mind.
The analogy that makes it click: a clerk is asked a hard question — “how tall is that distant mountain?” — and instead answers an easy one he can eyeball — “how bluish and hazy does it look?” — because faraway things look hazier. Most of the time haziness tracks distance, so the swap is fine. But on a crystal-clear day the mountain looks close, and the clerk, still secretly answering the haziness question, confidently tells you it’s near. He never realized he changed the question.
| The hard target question | The easy substitute question | |
|---|---|---|
| Availability | How frequent / probable is this event? | How easily does an example of it come to mind? |
| Mountain analogy | How far away is that mountain? | How hazy and blue does it look? |
| General form | (slow, requires real data) | (fast, gut-level, instantly available) |
A fully worked substitution. A friend asks: “Is it more dangerous to fly or to drive to the airport?” The honest target question — which has the higher death rate per mile or per trip? — requires statistics you don’t have on hand. So your mind quietly substitutes: which is easier to picture going horribly wrong? And flying wins that contest in a landslide — you’ve seen the dramatic news coverage, the reconstructions, the black-box transcripts; nobody runs a special report on a routine drive to the airport. The vivid air-crash image floods in, the substitute question returns “flying, by a mile,” and your mind reports that as the answer to the original question. So you say “flying is way more dangerous” — when per mile, driving is dramatically deadlier. You never answered the question you were asked. You answered an easier one and didn’t notice the swap.
Attribute substitution always pairs a HARD target question with an EASY substitute the mind answers instead. Match each hard question to the easy one availability quietly swaps in.
Pick a term, then click its definition.
Kahneman and Frederick describe attribute substitution as the deeper mechanism behind the availability heuristic. What is the relationship between the two?
Why ease usually works — and how it gets corrupted
If availability were simply wrong, evolution would have scrapped it. The reason it survives is that, on average, it’s a pretty good bet. In the ordinary course of life, the things you encounter more often genuinely are the things easiest to recall. You can summon a thousand examples of “rainy day” and far fewer of “hailstorm” because rainy days really are more common in your experience. Ease of recall correlates with true frequency — so most of the time, reading frequency off ease gets you a roughly right answer for free. That correlation is the heuristic’s whole license to exist.
Here’s the trouble. Ease of recall is influenced by frequency, yes — but it’s also influenced by a pile of things that have nothing to do with how often something happens. A memory can come easily because the event was vivid (a gory, concrete image sticks; a dry statistic slides off), because it was recent (last week feels more typical than last decade), because it was emotional (fear and outrage burn memories deep), because it was salient (it stood out, grabbed attention), or because it was repeated (the same story replayed twenty times feels like twenty events). Each of these inflates ease without touching frequency — and the heuristic, blind to why the example came easily, dutifully reports the inflated ease as inflated frequency.
So the right mental picture is not “availability is broken.” It’s: availability is a usually-good bet that fails in a small, predictable set of ways. Whenever something boosts the summonability of an example by a route other than “it actually happens a lot,” the heuristic gets fooled — and those routes (vividness, recency, emotion, repetition) are exactly the next lesson’s subject.
Fill in why the availability heuristic both survives AND systematically misleads.
Pick the right option for each blank, then check.
The heuristic survives because ease of recall true frequency — common things really are easier to remember, so reading frequency off ease is usually . But it gets corrupted because ease is ALSO inflated by vividness, recency, emotion, and , none of which track how often the event happens. So availability is a usually-good bet that fails in ways.
The classic experiments that pinned it down
A model isn’t science until someone designs an experiment whose result it predicts and rival explanations don’t. Tversky and Kahneman’s 1973 paper did exactly that, several times over. Here are the three demonstrations that nailed availability to the wall.
The letter-K experiment
This is the famous one. Subjects were asked: in typical English text, are there more words that begin with the letter K (like kangaroo, kitchen*) or more words with K in the third position (like* ask, cake, like*)? The overwhelming majority answered first position — most people guessed there are roughly twice as many words starting with K.
They were wrong, and not by a little. In ordinary English text there are about three times more words with K in the third position than words starting with K. So why is everybody so confident of the opposite? Because of how you search your memory. Your mental dictionary is indexed by first letter — that’s how recall works — so words starting with K (king, kite, keep, kind…) come pouring out effortlessly, while words with K in slot three (ask, lake, acknowledge…) require a slow, awkward, almost impossible search. First-position words are far easier to retrieve, so they feel far more numerous. Ease of search, not true frequency, drove the answer. Tversky and Kahneman ran this with several consonants (K, L, N, R, V) that are all actually more common in third position — and for every one, people wrongly judged first position more frequent. The pattern was rock solid.
You might think people just don’t know much about letter frequencies, and that’s the problem. But notice: the subjects weren’t asked to recall a fact they’d never learned. They were asked to generate examples, and they could — they just generated first-position words far more fluently than third-position ones. The knowledge gap isn’t the story; the retrieval asymmetry is. Both kinds of word exist in everyone’s vocabulary in roughly the true 1-to-3 ratio. What differs is the cost of pulling each kind up, and that cost — not the underlying count — is what the mind read as frequency. This is why the K experiment is such clean evidence for availability specifically: it isolates retrieval ease from everything else and shows that ease alone bends the judgment.
In the letter-K experiment, most people judged that more English words START with K than have K in the THIRD position — but the truth is the reverse (about 3× more in third position). What does this demonstrate?
The famous-names study
The second demonstration is even slicker. Tversky and Kahneman read subjects a list of about 39 names, then asked: were there more men or more women on the list? The trick was in the construction. In one version, the list contained a few very famous women (think household names of the day) mixed with many less famous men — and in another version, the genders were flipped: a few famous men among many less-famous women.
The result: people judged whichever gender had the famous names to be the more numerous — even when that gender was actually in the minority on the list. A handful of celebrities tipped the count. Why? Because famous names are easier to recall. When asked “were there more men or women?”, subjects didn’t tally; they did a quick mental retrieval — and the famous names came back fast and bright, so that gender felt like it had more entries. Once again, retrieval fluency impersonated frequency: a few highly available examples outweighed a larger pile of forgettable ones.
The “-ing” demonstration: when availability beats logic outright
The third one is my favorite because it shows availability overpowering pure logic. Tversky and Kahneman asked people to estimate, in a few pages of a novel, how many words would fit the form ”_ _ _ _ ing” (a seven-letter word ending in -ing) versus how many would fit ”_ _ _ _ _ n _” (a seven-letter word with n in the sixth, second-to-last, position). People estimated far more “-ing” words.
But look closely: every single “-ing” word is also a ”_ _ _ _ _ n _” word — the n in -ing sits in exactly that second-to-last slot. The set of “-ing” words is a strict subset of the “n second-to-last” words. It is logically impossible for there to be more of the subset than the whole. Yet people confidently estimated the opposite — because the category “-ing” is a vivid, easy-to-search pattern (you can rattle off running, jumping, walking…) while “n in sixth position” is an awkward, alien search that surfaces almost nothing. Availability didn’t just bend the estimate; it broke a law of logic and the subjects never felt it break. Ease of generating examples beat a containment relationship that a moment’s reflection would have made obvious.
Fill in the logic of the -ing demonstration.
Pick the right option for each blank, then check.
People estimated more words of the form '_ _ _ _ ing' than words with n in the second-to-last position, even though every -ing word is an n-second-to-last word — making -ing a strict of the larger category. So it is logically for -ing words to be more numerous. The estimate flipped because the -ing pattern is than 'n in the sixth slot' — availability overrode logic.
Here’s the whole experimental haul in one view — three different setups, one identical culprit:
| Experiment | What subjects judged | What they answered | The truth | What it isolates |
|---|---|---|---|---|
| Letter K | More words with K first, or K third? | ”First” (K is easier to recall first) | ~3× more have K third | Ease of memory search |
| Famous names | More men or women on the list? | The gender with famous names | That gender was the minority | A few highly available examples outweigh many forgettable ones |
| ”-ing” words | More “____ing” or “_____n_” words? | More “-ing” | Impossible — “-ing” is a subset | Availability overriding logic itself |
Match each classic experiment to the specific thing it demonstrates about the availability heuristic.
Pick a term, then click its definition.
Availability of an instance vs. the felt ease of the search
Now a subtlety that turns the whole thing one notch deeper — and it’s so counterintuitive it’s worth slowing down for. So far we’ve talked as if availability works through content: an example comes to mind, and that example tilts your judgment. But sometimes it’s not the content you retrieve that drives the verdict — it’s the felt ease or difficulty of the retrieving itself.
The landmark demonstration is Norbert Schwarz and colleagues (1991). They asked people to recall examples of their own assertive behavior, then rate how assertive they thought they were. The twist: one group was asked to list 6 examples; another group was asked to list 12. Now, common sense says the group that generated twelve examples of their own assertiveness should come away feeling more assertive — they just dug up a mountain of evidence for it.
The opposite happened. The group asked for 12 rated themselves as less assertive than the group asked for only 6. Why? Because listing twelve examples is hard — you run dry around seven or eight and have to grind for the rest. And the mind reads that struggle as a signal: “If being assertive came naturally to me, surely more examples would have come easily — the fact that I’m straining must mean I’m not very assertive.” The group asked for only six breezed through, the ease itself whispered “see, lots of evidence,” and they rated themselves high. Same direction of recall (both groups dredged up assertive memories), opposite conclusion — driven entirely by how easy or hard the dredging felt.
This is the deepest version of the model. Availability isn’t only “what examples did I retrieve?” It’s also “how did the retrieving feel?” The metacognitive experience of fluency or difficulty is itself an input — which is why being asked to think of more reasons for a belief can paradoxically weaken it, because the strain of finding the later reasons reads as “there must not be that many.”
The counterintuitive kicker
Asking someone to list more arguments for their position can make them less sure of it — because dredging up the eighth and ninth reason is hard, and that difficulty itself signals “I guess there aren’t many good reasons.” It’s not the content that flipped them; it’s the felt effort of recall. Availability runs on the experience of remembering, not just the things remembered.
In Schwarz et al. (1991), people asked to list 12 examples of their own assertiveness rated themselves as LESS assertive than people asked to list only 6. What does this reveal about how availability works?
The pitfall, and when to be on guard
We can now state the single mistake all of this collapses into. The pitfall is confusing “I can think of an example fast” with “this is common.” Speed of recall feels like data. It arrives with a little glow of confidence — of course it’s frequent, look how instantly the example showed up — and that glow is exactly the trap, because the glow is blind to why the example showed up. It can’t tell “this happens constantly” from “this was on the news last night,” from “this is the kind of thing that’s easy to picture,” from “I just heard a horror story about it.” All of those feel identical from the inside: fast, fluent, available.
So here’s the practical alarm. Be most suspicious precisely when a number, a probability, or a frequency feels obvious — especially right after a vivid example. The very feeling of obviousness is the tell. If you just watched footage, heard a scary anecdote, or read a dramatic headline, and your estimate of “how common is this?” arrives instant and confident, that’s not your knowledge talking — that’s the availability heuristic handing you the ease of recall dressed up as a fact. The fix you’ll build over the rest of the course starts with one question, the same one from lesson 1: “Am I judging the real frequency here, or just how fast an example came to mind?”
When to (deliberately) reach for it
To be fair to the lazy stockboy: availability is not a defect to be exterminated. It’s a fast, free estimator that’s usually right, and you couldn’t function without trusting your sense of “how common is this?” thousands of times a day. The skill isn’t to override it always — that’s exhausting and unnecessary. The skill is to know when its license expires. Lean on availability for low-stakes, everyday frequency calls in a stable environment you’ve sampled fairly (“is the bakery usually busy on Saturdays?”). Distrust it — and do the slow count instead — whenever the stakes are high, or whenever the examples reaching you have been filtered by vividness, drama, recency, emotion, or repetition (which, in a media-saturated world, is most of the time the subject is anything alarming).
Sort each statement by what it's really reporting: the EASE OF RECALL (how fast/fluently an example came up) or the TRUE FREQUENCY (the actual base rate).
Place each item in the right group.
- According to the CDC, heart disease causes about 1 in 4 deaths in the US
- K must start more words than it sits third in — K-initial words pour out so easily
- Per mile traveled, driving has a far higher fatality rate than flying
- It feels like shark attacks are everywhere — I can picture three news stories instantly
- Plane crashes seem common because the footage is so vivid and replayed
- In ordinary English text, about 3× more words have K in the third position
Recap
You’ve turned the slogan into a tool. The shape to carry forward:
Big picture
- What Availability Is
- The precise definition
- Judging frequency/probability by ease of recall
- Kahneman & Tversky, 1973
- Ease of RETRIEVAL, not how much you know
- Attribute substitution (the engine)
- Hard target question swapped for an easy one
- Answer the easy one, report it as the hard one
- Availability = one instance of this swap
- Why ease usually works — and breaks
- Ease correlates with frequency (so it survives)
- Corrupted by vividness, recency, emotion, repetition
- A usually-good bet that fails predictably
- The classic experiments
- Letter K: ease of search, not frequency (~3× more K-third)
- Famous names: a few salient instances outweigh many
- -ing words: availability beats logic (subset judged bigger)
- The felt ease itself (Schwarz 1991)
- Listing 12 assertive acts → felt LESS assertive than listing 6
- Difficulty of recall is itself a signal
- The alarm
- Suspect any number that FEELS obvious
- Especially right after a vivid example
- The precise definition
Check yourself: what availability is
A colleague says, "I know flying is statistically safe, but after seeing that crash report I just feel like it's risky." Which part of the availability heuristic does this most precisely illustrate?
Check your answer to continue.
“Availability heuristic” is now a precise tool, not a slogan: a specific instance of attribute substitution, where the mind answers “how easily does an example come to mind?” in place of “how frequent is this?” — and the lab experiments show it bending judgment, and even logic, every time. But we’ve left one promise unpaid. We keep saying ease gets corrupted by vividness, recency, emotion, and repetition. Next, in What Cranks Availability Up, we put each of those four engines under the microscope — exactly how every one of them jacks up the summonability of a memory without touching how often the thing actually happens.