Availability Heuristic
The availability heuristic is judging how common or how large a category is by how easily instances of it come to mind. Nothing is counted. The speed with which examples arrive becomes the answer, and because the swap happens out of sight the estimate feels like a report about the world rather than a reading of your own memory. The originating paper is Amos Tversky and Daniel Kahneman's, in 1973, with Tversky as first author. The version of the heuristic most often quoted is the one the research has since moved past.
What it is
Tversky and Kahneman's 1973 paper introduced availability as one of a small set of heuristics people use in place of probability. The intuitive reading of it is about quantity: you recall more plane crashes than car crashes, so you overestimate how dangerous flying is.
The demonstration that changed what the heuristic is about is not theirs. It is Norbert Schwarz's. He asked people to list instances of their own assertive behaviour, some of them six instances and some of them twelve, and then to rate how assertive they were. If availability meant quantity retrieved, twelve should have produced higher self-ratings. It produced lower ones. Twelve is hard to generate. Six is easy. What people read off was not the size of the pile but the effort of building it.
Kahneman takes the result seriously enough to propose renaming the thing an unexplained unavailability heuristic, because the inference rests on retrieval feeling harder than it should have. The clinching evidence is that the effect can be switched off by giving the difficulty a false cause: tell people the background music is impairing recall and the struggle no longer counts as information about them.