Verbal Memory Test

Recognition is a different job from recall

The number and visual tests on this site ask you to produce something from memory. This one asks only whether a word has already been shown, which is a weaker demand and a different capacity — people who stall at eight digits can still climb a long way here.

The score is a count of correct judgements, not a level. There is no rising difficulty except the growing set of words you have to keep track of, so a high score is mostly a matter of not mixing up two similar ones after a few dozen have gone by.

Three lives exist so that one slip does not end a run that was otherwise intact. A mis-tap on a word you had actually remembered would otherwise decide the score, which would make the test a measure of button accuracy.

Why some words are easier to hold than others

Concrete nouns stick. A word you can picture — a river, a shoe, a window — is cheaper to keep than an abstract one of the same length, which is why lists built of everyday objects produce higher scores than lists built of adjectives.

The list here is everyday vocabulary in the language the page is set to, not a translation of an English list. A Spanish run is a Spanish recognition task; serving English words under a Spanish UI would measure the wrong thing.

Similar pairs are the usual way a long run dies. After forty words the difference between two short ones that share a shape is easy to lose, and the mistake is almost always calling a new word seen rather than the other way around.

What the lives are for, and what they are not

They absorb a single error. They do not let a run recover once the set in your head has actually collapsed — two more mistakes usually follow quickly, which is why a score that used all three lives and then ended is typically within a handful of the score that would have ended on the first one.

There is no drop-back and no retry of the same word. A missed judgement is spent and the next word is drawn, so the score cannot be farmed by sitting on a word you are unsure about.

Mouse, tap and key all share one distribution. The score is a count, not a latency, so the input path cannot move it, and splitting the sample would only delay the chart.

This is a measurement toy, not a medical or psychological test. No score here says anything about your health, attention or cognition, and it is not a screening tool of any kind.

Questions

What is a good score on verbal memory?

The chart on this page is the answer from the attempts recorded here. A fixed number quoted without a sample behind it is a guess, and this test in particular varies with the language the words are in.

Does the word list change with the language?

Yes. Each language has its own everyday vocabulary. Switching language mid-run keeps the list you started with, so the test does not change language underneath you.

Why are there three lives?

So a single mis-tap does not end an otherwise good run. Once the set of words has actually been lost, the remaining lives rarely last long.