Based on the research of Roger N. Shepard on continuous recognition memory.

Verbal Memory Test

How many words can you keep track of?

The Verbal Memory Test measures how well you recognize words you have already seen; the Number Memory Test and the Sequence Memory Test test memory for digits and for squares on a grid, and the Reaction Time Test and the Typing Speed Test measure how quickly you respond and type. Words appear one at a time, and for each one you decide whether it is new or whether it has already appeared during this test. Some words come back after only a few others, some after a long stretch, and the list of words you have seen keeps growing.

Every correct answer adds a point. A wrong answer costs one of your three lives, and the test ends when the last one is gone. There is no time limit, so take each word at your own pace. How many words can you keep track of? To take the test, use the panel below.

Have you seen this word before?

This is a continuous recognition test. Instead of learning a list first and being tested on it later, you are tested on every word as it arrives, so taking in new words and remembering old ones happen at the same time, and each new word competes with all the ones before it. Roger Shepard and Martha Teghtsoonian introduced the method in 1961 with a deck of cards printed with three-digit numbers. The 63 students who went through it recognized a number from the previous card almost every time, but only about three times in four once six or seven other cards had come between, and a little over half the time after 50 or 60. Calling a new number "old" by mistake also grew more common as the deck went on, rising towards about three times in ten.

This version uses words. They come from a list of several hundred common, concrete words in your language, chosen from that language's own word frequency data, of similar length, and with no two words that differ by a single letter or an accent. After the first two words, the chance that a word is one you have already seen rises quickly to an even chance. When it is, the word is chosen by how long ago it last appeared: gaps of a few words and gaps of a hundred or more both occur, short gaps more often than long ones, so the test grows harder as the words you have seen pile up. The same word never appears twice in a row, and the sequence is drawn at random as you go, so no two runs are alike.

The choice of common words is deliberate. Rare words are easier to recognize than common ones, a finding known as the word frequency effect: in a continuous recognition study of younger, middle-aged and older men, the rarer words were remembered better in every age group. Everyday words feel familiar even when they are new in the test, so recognizing them depends on remembering this particular sitting rather than on a large vocabulary. The same study found that older men did as well as younger ones when only a few words came between a word's two appearances, and fell behind over longer gaps.

Your score is the number of correct answers before your third mistake, counting both the words you rightly recognized and the new words you rightly called new. Both kinds of mistake cost a life, so answering SEEN to everything, or NEW to everything, ends the test quickly. Your score is compared with the scores of tens of thousands of players, a self-selected crowd rather than a representative sample, so read the comparison as a rough guide.

This test is a quick measure of memory for words in one sitting. It is not a clinical memory assessment, and a single score depends on attention, tiredness and how hurried you were as well as on memory. If you want a steadier figure, take it a few times and look at your typical score rather than your best one.

References

Shepard, R. N., & Teghtsoonian, M. (1961). Retention of information under conditions approaching a steady state. Journal of Experimental Psychology, 62(3), 302-309.
Poon, L. W., & Fozard, J. L. (1980). Age and word frequency effects in continuous recognition memory. Journal of Gerontology, 35(1), 77-86.
Shepard, R. N. (1967). Recognition memory for words, sentences, and pictures. Journal of Verbal Learning and Verbal Behavior, 6(1), 156-163.

Frequently Asked Questions

What does the Verbal Memory Test measure?

It measures recognition memory for words: how well you can tell the words you have already seen in the test from new ones while the list of words you have seen keeps growing. It is a test of short- and medium-term memory for this one sitting, not of vocabulary.

How is my score calculated?

Every correct answer is worth one point, whether you rightly said SEEN to a word that had appeared before or NEW to one that had not. Each wrong answer costs one of your three lives, and the test ends when the last one is gone. Your score is the number of correct answers.

What is a good score?

Of the tens of thousands of players your score is compared with, the middle player scores about 41, and one in ten scores 95 or more. Someone answering at random would score about 3 on average, and someone who pressed NEW every time about 7.

Is there a time limit?

No. Only your answers count, not how fast you give them, so take each word at your own pace. The test ends only when your three lives are used up.

Can I take the test on my phone?

Yes. On a phone the word is shown large with two big SEEN and NEW buttons at the bottom of the screen, within reach of your thumbs. On a computer you can click the buttons or use the left and right arrow keys.

Why are the words so ordinary?

On purpose. Rare words are easier to recognize than common ones, because they stand out, so a test built on unusual words partly measures vocabulary. Every language has its own list of common, concrete words, chosen from that language's own word frequency data rather than translated from English.

Verbal Memory Test

Why Use This Test?

1. Free. The Verbal Memory Test is free, runs in your browser and works with a mouse, a keyboard or a touchscreen.

2. Words in your language. Each language has its own list of common, concrete words, chosen from that language's own word frequency data.

3. Fair from the first word. Repeats come back after gaps drawn at random, never twice in a row, so neither guessing nor pattern-spotting gets you far.