Wednesday, October 7, 2026
They Buried
We Dug It Up
⌕ Search
InvestigationThe Signal

We Asked 900 Strangers To Draw It. They Made The Same Mistake.

Shown nothing and asked to draw well-known icons from memory, strangers produce identical errors at rates up to 78 per cent. The obvious explanations were tested and none fit. Nobody knows why, and the researchers say so.

A months-long reporting project. Documents cited below are held in The Vault and available to readers.

Nine hundred drawings, overlaid. The error is where the lines agree. Composited for They Buried

Seventy-eight per cent of nine hundred people gave him a monocle. He has never had one. He does not have one now and did not in any edition since 1936.

The people who drew it had never met each other.

Figure How often strangers make the identical error
Monopoly man drawn with a monocle78% — he has never had onePikachu drawn with a black tail tip71% — the tail is plainFruit of the Loom drawn with a cornucopia63% — there has never been oneCurious George drawn with a tail58% — he has no tailVolkswagen logo drawn without the gap41% — there is a gapControl: a clock face4% — errors random, not sharedPercentage of participants producing the same specific wrong detail, unprompted, drawing from memory.
Nine hundred participants, recruited by this newspaper, shown nothing before drawing. Pre-registered before recruitment. The full drawing set is published; you can look at all nine hundred. They Buried

What we did

We pre-registered everything on 4 March, before recruiting a single person: which five icons, the control, how a drawing would be scored, and who would do the statistics. We did not amend it.

Participants were shown nothing. No reference image, no multiple choice, no list. A blank box and an instruction to draw.

Then a control: a clock face. Errors on the clock were random, scattered and unshared — 4 per cent agreement, which is what noise looks like.

That control is the difference between this piece and every other piece about the Mandela effect, and the outlet that reported our numbers as proof of a changed timeline did not mention it.

What was ruled out, and by whom

Professor Wilhelmina Bainbridge-Osei's group tested the obvious explanations before we ever ran ours.

Not looking properly: fails, because people get it wrong having seen the image thousands of times. Copying an erroneous version seen online: fails, because the effect appears in people who cannot have encountered one. Guessing: fails, and fails interestingly — participants are confident, and confidence goes with being wrong.

Four explanations, tested, gone.

The sentence she asked for in bold

She read our draft and asked to reply at length, because her field's findings get misused constantly. Her statement runs in full at the foot of this story and it contains the most useful thing in the piece:

That a thing is unexplained is an ordinary condition in science and not an invitation.

Every time this gets written about, somebody reads "scientists cannot explain it" and hears "scientists have ruled everything else out." Those are opposite sentences. Four things have been ruled out. There are a great many things.

The two who remember the film

We interviewed two people who remember a 1990s film in which a comedian plays a genie. There is no such film. There is a different film, two years later, in which a basketball player plays a genie.

Both know it does not exist. Both looked, years ago, and found nothing. Neither has changed what they remember, and both were careful to say those are separate facts.

One of them described the video shop, the shelf, the position on the shelf, and the colour of the spine. Then he said: "And I know. I have known for eleven years. It is still on the shelf when I think about it."

Our correspondent asked whether that troubled him.

"Not any more," he said. "It taught me something about myself that I would not otherwise have believed, which is that I can be completely certain and completely wrong at the same time, and not feel any different. I think about that when I read the news."

What we got wrong

Our headline says nine hundred strangers. They are nine hundred people who take paid online surveys, which is a narrower and stranger population than the word implies, and our standards editor has flagged it in the box below rather than letting it pass.

The pre-registration says so. The headline does not. She is right.

Sources & Method

Pre-registered before a single participant was recruited: the five icons, the control, the scoring rule, and the analysis plan, all lodged publicly on 4 March 2026 and not amended. Participants were shown nothing and asked to draw from memory. A statistician who was not involved in recruitment wrote the analysis. All nine hundred drawings are published, not a selection.

Who we spoke to

  1. Prof. Wilhelmina Bainbridge-Osei, Cognitive neuroscientist; has published on visual memory for icons. Interviewed by video call twice; advised on the protocol and was paid nothing February and June 2026 Asked repeatedly that we print that nobody has an explanation. We have printed it four times.
  2. Nine hundred participants, Recruited through a survey panel and paid the panel's standard rate. Drew from memory, shown nothing, pre-registered task order April 2026
  3. Dr. Emeka Sandoval-Trần, Statistician, retained by this newspaper. Wrote the analysis plan before data collection; report published March–May 2026
  4. Two people who remember Shazaam, Named in our notes with consent. Interviewed at length May 2026 Neither has changed their mind and both know the film does not exist. Their accounts are printed.

Documents

  • PX-1968 — Nine hundred drawings from memory, with the pre-registration and analysis plan accepted

What we could not confirm

  • Why. This is the finding and the researchers we spoke to are emphatic that it is unexplained. None of them offered a mechanism and we did not press them for one.
  • Whether our sample generalises. Nine hundred online panel participants are not a population, and the panel skews younger and more online than the country.
  • Whether the icons are representative. We chose five that circulate widely, which is a selection effect, and a control that does not.
Disclosure. This newspaper paid the survey panel $27,400 and the statistician $4,100. Prof. Bainbridge-Osei advised on the protocol and was offered a fee, which she declined.

How Others Covered This

The same events, as reported elsewhere on the same day. We list what each outlet had that we did not, as well as what we had that they did not — including where we come off worse. Why we print this.

  1. Aperture
    The Memories We Share And Never Had

    An essay on the history of collective false memory, from crowd psychology to the internet.

    Had that we did not

    A far better historical frame than ours, and an interview with a memory researcher we could not book.

    Left out

    Any new data. Theirs is a piece about the literature; ours ran the experiment.

  2. The Meridian Telegraph
    PROOF THE TIMELINE WAS CHANGED

    Reported the drawing results as evidence of timeline alteration.

    Had that we did not

    Nothing.

    Left out

    The control condition, in which errors are random and unshared, and which is the whole point of running a control.

  3. They Buriedthis newspaper
    We Asked 900 Strangers To Draw It. They Made The Same Mistake.

    Pre-registered, ran a control, published all nine hundred drawings.

    Had that we did not

    The control condition and the raw drawings.

    Left out

    We recruited online, which means our nine hundred are not nine hundred humans, they are nine hundred people who do surveys online. The paper says so; the headline says 'strangers' and implies more. — V. Ashcombe-Doyle, standards editor

Right of Reply

They Buried contacted Prof. Wilhelmina Bainbridge-Osei throughout; she read the full draft on 28 July with no deadline. She asked to reply and to be quoted at length, on the grounds that her field's findings are routinely misused. Printed unedited.

Your data look right to me and I want to say two things about what they mean.

The first is that this is a real, robust, replicated effect and it is genuinely not explained. We tested the easy answers. People are not simply failing to look at the detail — they get it wrong even when they have seen the image thousands of times. They are not copying errors they saw online — the effect shows up in people who cannot have encountered the erroneous versions. They are not guessing — they are confident, and confidence correlates with being wrong.

The second thing is the one I would like in bold. That a thing is unexplained is an ordinary condition in science and not an invitation. Most of what I work on is unexplained. Unexplained means nobody has done the work yet. It does not mean the answer is large.

Every time this is written about, somebody reads 'scientists cannot explain it' and hears 'scientists have ruled everything else out'. Those are opposite sentences. We have ruled out four things. There are a great many things.

Published unedited under our right-of-reply guarantee.

How was this story?

We publish the result, whatever it is. Reader verdicts appear on the front page and in our newsroom metrics.

3,063 verdicts · 74.0% loved it

Readers' Letters 0

Printed at once under the name you give and read by the desk afterwards; anything unfit is removed, with a note saying so, and nothing else is ever deleted — only corrected. Letters that changed something in the story carry a mark saying so, and there are 5 of those across the archive.

  1. No letters yet on this story. Yours would be the first.

Write to the desk