False Memory Test: 20 Questions on Details People Misremember

Twenty questions, each with four answers and exactly one correct one, worth one point. There is no time limit, no penalty for a wrong answer and no adjustment for anything. Your total is the number correct out of 20, split in the breakdown into the twelve public-detail questions and the eight word-list questions. Those two groups are our own split of our own quiz, not validated subscales, and the two halves measure different things: Part 1 tests a memory you formed years ago from many exposures, Part 2 tests one you formed four minutes ago from a single reading.

The four score ranges are editorial cutoffs written for this page. They are not percentiles, not an IQ estimate and not clinical thresholds, and there is no standardization sample behind them. With four options per question, guessing earns a point one time in four on average, and nothing in the scoring corrects for that.

One honest limit: nothing stops you going back to re-read the word list, searching for an answer, or taking the quiz twice. If you do any of those, your score stops describing your memory and starts describing your browsing. The demonstration only works the first time.

The Mandela effect is the name for a false memory that many people hold in the same specific form. It is called that because of the memory it was named for: a widely shared recollection that Nelson Mandela died in prison in the 1980s, when he was released in 1990, served as president of South Africa from 1994 to 1999, and died in 2013. The writer Fiona Broome began using the term around 2010 after finding how many people shared that memory.

What makes these cases interesting to researchers is not that memory fails, which is ordinary, but that it fails the same way for strangers. Prasad and Bainbridge tested forty icons in 2022 and found consistent, confident and widespread false memories for specific ones, including the Monopoly mascot's monocle and Pikachu's tail. Their attention data, gathered with eye-tracking-like methods, showed people attended to the correct versions in much the same way as the altered ones, so this is not simply a matter of never having looked properly. When asked to draw the icons from memory, people spontaneously produced the errors.

What it is not is evidence of anything outside your own head. No finding in this literature supports changed timelines, alternate realities or edited history, and the researchers who study these cases do not claim one; the explanations on the table are ordinary memory mechanisms, including expectations about how a thing should look, the influence of other similar images, and the fact that the altered version is sometimes simply more memorable than the real one. If that feels like a thinner answer than a parallel universe, it is also the one with evidence behind it, and it has the advantage of explaining why the errors are so consistent.

The foundational idea is almost a century old. Frederic Bartlett, working in Cambridge in the 1920s and publishing in 1932, had people retell an unfamiliar folk story repeatedly and watched them smooth it into something that made sense to them, dropping the parts that did not fit and adding parts that did. He called remembering reconstructive: you rebuild an account from fragments plus expectation, rather than retrieving a recording.

Elizabeth Loftus then showed how easily the rebuild takes in new material. In her 1975 work on leading questions, the way a question was worded changed what people later reported having seen, because the presupposition inside the question became part of the memory. That is the misinformation effect, and it is one reason a detail you have seen captioned wrongly a hundred times online can end up sitting in your memory as something you saw yourself.

Part 2 of this quiz uses the sharpest laboratory demonstration of the same thing. Deese noticed in 1959 that certain word lists reliably produce a specific intrusion, and Roediger and McDermott built the standard paradigm around it in 1995: study a list of words all related to one absent word, and people recall or recognize that absent word at rates close to the words they actually studied, often with high confidence. The absent word on this page is sleep. If you picked it anywhere in Part 2, you did not make a mistake so much as watch your memory do exactly what it is built to do, which is to store the gist and fill the rest in. Perception fills in the same way, and this ink blot test lets you watch it happen across ten original ambiguous images.

This is not a memory assessment. Getting a low score here says nothing about whether your memory is working normally, because the items were chosen for being commonly missed rather than for discriminating good memory from poor. It is also not an intelligence measure; nothing about the score converts into anything. If you want a short structured memory self-check instead, Memory & Dementia Concern Screen is the page for that.

Working Memory Test (Digit Span) looks at holding information briefly in mind, and Visual Memory Test at recognizing what you have seen.

If a real worry brought you here, if you are forgetting conversations, repeating questions or losing your way somewhere familiar, do not use a quiz for it. A change from your own previous ability is what matters, and that is a conversation with a doctor rather than a score. Our screener at Alzheimer's Concern Screen is written around change from baseline and says plainly what an online test cannot do, and Brain Fog Test covers the kind of fuzziness that comes with sleep debt, stress or illness.

If you came for curiosity, the useful takeaway is practical rather than alarming. Confidence is a poor guide to accuracy, in your own memory and in other people's. When a detail matters, in a disagreement, at work, in anything with consequences, check the source rather than the strength of the feeling. That is the whole lesson of this quiz, and the twelve items in Part 1 are a cheap way to learn it.