What this lab asks
How much of a list of words can you write down straight after reading it, and which places in the list survive best? And can a list make you remember a word that was never in it? The first test draws your serial position curve from three lists of unrelated words; the second uses the lists of a classic false-memory experiment, with published values for each list.
Serial position: three lists of unrelated words
Each list has 15 common nouns from 15 categories (an animal, a tool, a vegetable and so on), so no theme links them, and the lists never share a word. A word stays for 1.5 s, then the screen is empty for 0.5 s: one word every 2 s, the pace of the 1999 norms below. Then you have 2 minutes to write every word you remember, in any order; Done ends the list sooner.
The curve shows, for each of the 15 places, in how many of your lists you wrote its word, and the result says where your first written word came from. With three lists a point can only be 0%, 33%, 67% or 100%, so History averages the curve over your saved runs. No openly available study with comparable values for unrelated lists was found for this page, so your curve stands on its own. For the related lists of the second test, over all 36 lists of the 1999 norms, 85% of the students wrote the first word, 45% to 48% the words in places 5 to 10, and 93% the last one.
False memory: lists built around a hidden word
The 15 words of each list are all associates of one word that is never shown. After each of the 6 lists (or 4) you have 2 minutes to write the words, with the instruction of the original studies: write only words you are reasonably sure were shown. The hidden words are revealed only after the last list. An optional recognition test follows: for the words in places 1, 8 and 10 of each list, its hidden word and two new words unrelated to every list, you answer Old or New. The page remembers on this device which lists you have seen and offers set B after set A.
Where the lists come from
James Deese (1959) found that after the 12 most common associates of one word, 42% of his participants wrote that word although it was never presented. Henry Roediger and Kathleen McDermott (1995) revived the method. In their first experiment 36 students heard 6 lists of 12 words, read aloud at about one word every 1.5 s, and wrote each list for 2.5 minutes. They wrote 65% of the studied words and the hidden word after 40% of the lists, about as often as the middle words of a list, while other words not on a list appeared after only 14% of the lists. In recognition they called 86% of the studied words old, 84% of the hidden words and 2% of unrelated words. In their second experiment (30 students, 15-word lists from tape) the hidden word came after 55% of the lists, more often than the words in places 4 to 11 (47%).
Michael Stadler, Henry Roediger and Kathleen McDermott (1999) published norms for 36 such lists: 205 students heard them in a male voice, one word every 2 s, in two sets of 18 lists (95 and 110 students), and wrote each list for 2 minutes, again only words they were reasonably sure of. The hidden word of the strongest list was written by 65% of the students, that of the weakest by 10% (mean 40%); in the recognition test 84% and 27% called it old (mean 66%). Endel Tulving named the method the DRM paradigm, after Deese, Roediger and McDermott.
The 14 lists and their 1999 values (names the hidden words)
In the order a run uses them. Wrote it and Called it old are the shares of students in the 1999 norms (Tables 1 and 2); the English words of each list are those of the paper's appendix, in its order.
| Set | Hidden word | Wrote it | Called it old | Slovak list |
|---|---|---|---|---|
| A | sleep | 61% | 80% | spánok |
| A | chair | 54% | 74% | stolička |
| A | needle | 52% | 68% | ihla |
| A | doctor | 60% | 71% | lekár |
| A | smoke | 54% | 73% | dym |
| A | music | 34% | 69% | hudba |
| A | slow | 42% | 69% | pomalý |
| A | bread | 31% | 64% | chlieb |
| B | window | 65% | 84% | okno |
| B | sweet | 54% | 78% | sladký |
| B | river | 42% | 67% | rieka |
| B | girl | 32% | 58% | dievča |
| B | thief | 23% | 70% | zlodej |
| B | black | 34% | 49% | čierna |
The Slovak lists are translations made for this page, without norms. A word with no close Slovak equivalent, or whose translation would contain the hidden word, was replaced, for example vankúš (pillow, from the 1995 version of the list) for slumber and Dunaj (the Danube) for Mississippi. Lists with brand names, or whose Slovak translation would merge words, were left out.
Things to try
- Run Serial position two or three times and look at the averaged curve in History: are both ends higher than the middle?
- Write the last words first in one run and start from the beginning in another; Roediger and McDermott asked for the last few words first, calling it a standard instruction for the task. Does the end of your curve change?
- After a False memory run, check whether you wrote the hidden words of the lists with the highest 1999 values.
- In the recognition test, compare how often you called the hidden words old with the studied words from places 8 and 10.
- If you read English and Slovak, run False memory in the other language with the other set.
Why it happens
Roediger and McDermott (1995) start from an idea of Underwood (1965): while reading a word such as hot, people may think of an associate such as cold, and later take that thought for something they were shown. The associate may also become active unnoticed, as activation spreads through related words. In their second experiment people said they remembered the moment the hidden word was presented about as often as for real words. The authors suggest the word may come to mind while the list is read, and that people may then claim to remember it through a failure to tell what was perceived from what was only thought of (reality monitoring).
How the words are scored
- Letters only. Capitals, spaces and punctuation are ignored, and Slovak letters count without accents (postel is posteľ).
- Exact, then form, then typo. English forms are the endings -s, -es, -ed, -ing, -ly, and -ies for -y (candies counts for candy; sitting does not count for sit). Slovak forms share a beginning of at least 4 letters that covers all but the last 2 letters of the longer word, lengths at most 3 apart (unavená counts for unavený, chrápe for chrápať). A typo is one wrong, missing, extra or swapped letter, two from 9 letters; words of up to 3 letters count only when exact (pit does not count for pot).
- Known words are themselves. A correctly spelt word from any list of the page is that word, never a typo of another: lake does not count for wake.
- Each studied word once. An answer goes to the best matching studied word not yet written; when two match equally, to the earlier one in the list. A second answer for the same word is a repeat: crossed out, not counted.
- Hidden words have a fixed list of forms (plurals and other endings, in Slovak also a common synonym). From 5 letters a slip of the basic form also counts: two neighbouring letters swapped, a letter typed twice, or one of a double letter left out. Other typos do not, because they turn real words such as sheep or chief into a word that was never shown. A studied word never counts as a hidden word.
- Other answers are counted apart: a word from an earlier list of the run, or any other word.
How the result is calculated
Studied words written is your share of all studied words of the run, with a 95% Wilson interval. When whole lists go better or worse than chance would make them, the interval widens by the design effect 1 + (m − 1)ρ, with m = 15 words per list and ρ, the share of the variation that lies between lists, estimated from your per-list results and never below 0. The page's calculation was run on simulated people (November 2026, 4,000 runs per case) whose chance of writing a word follows a U-shaped curve, with lists that vary at random: with 3 lists the interval held the true value in 95% to 97% of runs and was 25 to 31 percentage points wide, and when whole lists differed even more, in 91% to 93%. Hidden words written after k of N lists has its own Wilson interval: 3 of 6 gives 19% to 81%.
Compared with the 1999 norms
For English lists, the results show the published values of the same lists beside yours: the hidden words, the curve by place and the recognition test; for new words, the 2% of the 1995 study. The students heard each word and wrote with a pen; you read and type, which is slower. The Slovak lists have no norms, and the English values are shown for orientation only.
Why there is no percentile
The norms are averages for lists, not ranks of people. Lists keep their order well: with the students split into two halves, the lists' false-recall values correlated at .80 (.85 for recognition), and students who noticed the themes produced the hidden word after the same lists as those who did not (.72). The paper gives no value for how well a person keeps their place from one run to the next, so History compares you with your own runs.
Limits
- Slovak lists. Translations without norms; some associations are weaker or stronger than in English.
- Tolerance. Forms and typos can now and then credit a different word (tire would count for tired). Your answers in This run show how each one was scored.
- Knowing the trick. Once you know how the lists work, the effect may weaken; the 1999 paper does not say by how much, so the page only offers lists you have not seen.
- Few lists. One run shows whether the illusion happened, not how prone to it you are.
- Timing. Words last whole frames of your screen, so 1.5 s can be a frame longer or shorter; the result says when a word was further off. A hidden tab pauses the run, and Resume shows the rest of the list from the word that was cut short.
- Not a medical test. This is not a clinical memory test and cannot show memory problems or dementia.
Controls
- Keyboard: Enter, Space or a comma adds the word; Backspace in the empty field takes the last word back; Tab reaches Done and your words. Recognition: F or ← Old, J or → New. Esc stops a run, except while you write.
- Touch and mouse: tap or click a word to remove it; Done ends the list early; Old and New are under the display.
- Pauses: for 0.7 s after a list, a result or Stop the buttons are dimmed and ignore keys still pressed. A second click within 0.7 s of a start does not stop the run, so a double click is safe.
- Language: the link at the end of History switches the app and the word lists to Slovak from the next run.
Sources
- Roediger, H. L., III, & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21(4), 803–814. doi:10.1037/0278-7393.21.4.803
- Stadler, M. A., Roediger, H. L., III, & McDermott, K. B. (1999). Norms for word lists that create false memories. Memory & Cognition, 27(3), 494–500. doi:10.3758/BF03211543