The Experiment That Gave People Memories of a Childhood That Never Happened

At the beginning of the experiment, one participant was completely certain.

He had never been in a hot-air balloon.

The photograph in front of him appeared to say otherwise.

It showed him as a child, alongside members of his family, standing in the basket of a balloon. It looked like exactly the kind of slightly faded family photograph that might have been sitting in an album for years.

But it was fake.

The researchers had taken a genuine childhood photograph of him, digitally removed his image, and placed it inside another photograph. Before the experiment began, a family member had specifically confirmed that he had never taken a hot-air-balloon ride.

When the participant first saw the image, he did not suddenly produce a memory. Quite the opposite.

“No memory whatsoever,” he told the interviewer.

A little over a week later, the situation looked very different.

Now he was trying to establish exactly when the ride had happened. He thought it had been while he was in school. He estimated that the ride had cost around ten dollars. He believed it had happened on a Saturday. He described going with his parents. He estimated that the balloon had risen about 20 metres into the air. He even reasoned that his mother must have been standing on the ground taking the photograph.

None of it had happened.

And the researchers had not supplied most of those details.

He had.

That transformation—from I have absolutely no memory of this to an autobiographical account containing a time, location, people, actions and visual details—is what made Kimberley Wade and colleagues' 2002 hot-air-balloon experiment so striking.

It was not simply evidence that people can answer a question incorrectly.

It showed researchers watching a nonexistent childhood episode gradually acquire the qualities of a memory.

First, the researchers needed a childhood that looked real

Wade, Maryanne Garry, J. Don Read and D. Stephen Lindsay recruited 20 adults between 18 and 28 years old.

But the experiment actually began with their families.

A relative of each participant supplied childhood photographs taken when the participant was roughly four to eight years old. The researchers selected three genuine photographs showing moderately significant childhood experiences—things such as birthdays and family holidays.

The relative also had to verify something very specific: the participant had never been on a hot-air-balloon ride.

The researchers then used Photoshop to create a fourth photograph.

A genuine image of the participant as a child, sometimes accompanied by family members, was cut from a real family photograph and inserted into a photograph of a hot-air balloon. The finished image was printed alongside the genuine childhood photographs and presented as just another event from the participant's past.

The choice of a balloon ride was deliberate. In New Zealand, where the research was conducted, balloon rides were plausible enough that they could reasonably have occurred, but unusual enough that relatives could confidently say whether the participant had ever experienced one.

The participants were then interviewed three times over seven to sixteen days.

At each interview, they were encouraged to describe everything they could remember about the events in the photographs.

For the real photographs, this worked much as you would expect. By the first interview, participants recalled information about 56 of the 60 genuine events. By the final interview, that number had increased to 58.

Then came the balloon photograph.

At first, nobody remembered the balloon ride

This detail matters.

When participants first encountered the false photograph, not one immediately reported remembering the balloon ride.

The false memories did not appear fully formed the moment the photograph was placed in front of them.

Instead, the researchers created conditions in which participants repeatedly tried to recover a memory they had been led to believe must exist.

If someone could not remember an event, the interviewer reassured them that this was normal: people often have difficulty recalling childhood experiences they have not thought about for a long time.

They were then given time to concentrate.

If nothing emerged, the interviewer used guided imagery and mental context reinstatement.

For the balloon event, participants were asked to close their eyes and imagine themselves back inside the balloon. They were encouraged to think about who might have been there, what the weather was like, what they might have seen when they looked over the basket and what the experience could have felt like.

The fake photograph remained visible during the discussion.

After the first interview, participants took the photographs home. They were asked to spend a few minutes each night trying to remember any events they had not yet been able to recall.

Three to seven days later, the researchers asked again.

Then they asked again several days after that.

What happened across those interviews is more revealing than the final percentage alone.

You can actually watch the memory being built

One participant, identified in the paper as S.B., is perhaps the clearest example.

During the first interview, S.B. looked at the photograph and said he had never thought he had been in a balloon. He acknowledged that the child in the photograph was him, but reported no memory of the event.

Even after being encouraged to concentrate, he could not retrieve anything.

The interviewer then guided him through imagining where the ride might have taken place and who might have been present.

At that stage, he was still speculating rather than remembering.

By the third interview, however, his account had changed.

He now placed the experience during a particular period of school. He constructed an approximate price for the ride. He believed it had happened on a Saturday. He identified his parents as likely companions and described how high the balloon had gone. He also decided that his mother was probably on the ground taking the photograph.

The experimenters had not inserted those details into the picture.

The participant had supplied them while trying to make sense of an event that he had been told belonged to his childhood.

Another participant, F.J., showed a more ambiguous progression.

At the first interview, F.J.'s reaction was essentially disbelief: he had not known he had ever been in a balloon and had never seen the photograph before.

He then tried to work out where such an event could have occurred.

Perhaps, he reasoned, it had happened at a fair in a particular city where hot-air balloons were sometimes present.

By the final interview, that inference had begun to acquire a different quality. He still expressed uncertainty, but said he thought he remembered being in the balloon and described seeing a road, people and a large field below.

The uncertainty here is important. False-memory formation is not necessarily a switch from absolute disbelief to absolute certainty.

For some participants, it looked more like a gradual movement from reasoning, to imagining, to familiarity, to something that increasingly resembled remembering.

That is precisely why researchers distinguish between a clear false memory, a partial false memory, imagery without a claimed memory and complete rejection of the suggested event.

By the final interview, half of the group reported at least some false recollection

At the end of the first interview, seven of the 20 participants had already begun reporting either partial or clear false memories of the balloon event.

By the third interview, ten participants—50% of the sample—were classified as having either a clear or partial false memory.

Four met the researchers' stricter standard for a clear false memory.

Six were classified as having partial false memories.

The participants with clear false memories were, on average, 66.7% confident that the balloon ride had actually occurred.

That confidence was still lower than participants' confidence in their genuine recalled events, which averaged more than 90%, but it was far from simple guessing.

One of the most fascinating analyses came when Wade and colleagues broke the participants' reports into individual pieces of information and asked a simple question:

How much of what people were saying could actually have come from the photograph?

The answer was surprisingly little.

Across the interviews, fewer than 30% of the details in the reports could generally be traced directly to information visible in the photographs.

Participants were generating the rest themselves.

They supplied locations.

Times.

People.

Emotions.

Sensory information.

Descriptions of what the event had supposedly been like.

And as the interviews progressed, the proportion of emotional material in the false accounts increased.

The photograph did not contain the complete false memory.

It appears to have acted more like a starting structure around which a memory could be constructed.

That distinction is crucial.

The study was not simply showing participants a fake picture and watching them repeat what they saw.

People were adding autobiographical material that was never in the picture at all.

The stranger part: researchers had already done this without photographs

The balloon experiment did not appear from nowhere.

Years earlier, researchers had discovered that a fictional childhood event could sometimes be implanted using little more than a believable story attributed to the participant's family.

The best-known example is Elizabeth Loftus and Jacqueline Pickrell's 1995 “lost in the mall” study.

Twenty-four participants were given descriptions of several childhood events. Their relatives had supplied information about the genuine events.

The researchers added one that was false.

Participants were told that when they were about five years old they had become lost for an extended period in a shopping mall or large department store, cried, were helped by an elderly woman and were eventually reunited with their family.

The story was personalized with genuine information about the participant and their family—but relatives had confirmed that the event itself had not happened.

Participants first wrote what they remembered about each event and later discussed them in two interviews.

By the end of the study, approximately 25% reported some memory of the fabricated event (Loftus & Pickrell, 1995).

Nearly three decades later, Murphy and colleagues conducted a much larger preregistered replication with 123 participants.

Using a procedure modelled on the original experiment, they again combined genuine childhood events supplied by relatives with a fabricated getting-lost event.

Under their coding system, 35% of participants produced reports classified as full or partial false memories. The breakdown was approximately 8% full and 27% partial false memories. The study also separately measured what participants themselves said they remembered, highlighting an important distinction between researchers' classifications of a narrative and participants' own subjective claims of recollection.

The phenomenon therefore does not depend on a Photoshop trick.

Family-linked suggestion alone can sometimes be enough.

Researchers then tried inventing different childhoods

Other researchers began testing what kinds of false events could become incorporated into autobiographical memory.

Ira Hyman, Troy Husband and F. James Billings contacted participants' parents to obtain genuine childhood stories, then inserted fabricated events among them.

Depending on the experiment, these included things like an accident at a wedding involving a punch bowl, setting off a sprinkler in a store or being left alone in a car.

Once again, participants did not simply receive an obviously fictional story.

The event was presented alongside genuine material obtained from someone who knew their childhood.

Across repeated interviews, some participants began producing information about the fabricated experiences.

One of the striking findings from Hyman's research was that participants who generated relevant background knowledge early in the interviewing process were more likely to later produce false recollections.

That makes intuitive sense.

To create a convincing autobiographical event, the mind needs material.

A wedding provides knowledge about relatives, clothing, rooms, tables, food and social rituals. A shopping trip brings stores, parking lots, parents and familiar places. A school prank can draw on real teachers, classmates and classrooms.

The invented episode can be constructed from genuine pieces.

This also helps explain one of the most important discoveries in the literature:

imagining the event can make a difference.

What happened when researchers told people to imagine harder?

In 1996, Hyman and Joel Pentland tested whether guided mental imagery would increase false remembering.

Sixty-five college students were asked about true childhood events supplied by their parents and a fabricated event involving spilling a punch bowl at a wedding.

When participants initially failed to remember an event, some were instructed simply to think about it.

Others were guided through constructing an image.

The difference was substantial.

About 12% of participants in the “think” condition eventually produced some false recollective experience, compared with approximately 36% in the guided-imagery condition.

Using stricter criteria, full false memories were also more frequent in the imagery condition: roughly 25% compared with 9% in the think condition.

The implication is not that visualization automatically turns fantasy into memory.

It does not.

Rather, imagination can manufacture some of the features that later make a mental representation feel memory-like: visual detail, spatial context, emotion, familiarity and narrative structure.

That possibility was reinforced by another influential experiment.

Garry, Manning, Loftus and Sherman (1996) asked participants to rate how likely various childhood experiences were to have happened to them. Later, participants imagined some of those events. When they rated the childhood experiences again, imagining certain events increased their confidence that those events had actually occurred.

The effect became known as imagination inflation.

Thinking vividly about a possible past can sometimes make that past feel more plausible.

Then researchers removed the fake photograph entirely

The obvious criticism of Wade's balloon study is that most people do not regularly encounter carefully Photoshopped evidence of fictional childhood adventures.

So Lindsay and colleagues asked a cleverer question:

Could a completely genuine photograph still help create a false memory?

They recruited 45 undergraduates.

Parents supplied information about two real school experiences and confirmed that a third event had never happened.

The invented event was a childhood prank.

Participants were told that when they were in Grade 1 or 2, they had brought a brightly coloured toy called Slime to school and secretly placed it in their teacher's desk before the teacher arrived. The story included the participant's name and the real name of their teacher.

Some participants received only the narratives.

Others also received an authentic class photograph from the relevant school year.

The photograph had not been altered.

It contained the participant's actual classmates.

Their actual school context.

Their real childhood appearance.

And absolutely no evidence of anyone putting Slime in a teacher's desk.

Participants were encouraged to work on recovering the memory for a week. They used guided imagery, repeatedly revisited the story and were contacted midway through the week to encourage continued recall attempts.

The photograph made an enormous difference.

By the second session, 65.2% of participants in the photo condition were judged to be experiencing memories of the invented event, compared with 27.3% in the no-photo condition. If looser criteria incorporating imagery were used, the percentages were higher still.

More strikingly, participants classified as having false memories rated those experiences as approximately as compelling as their memories for the genuine childhood events on measures such as reliving, remembering and confidence.

During debriefing, most participants could identify the school prank as the likely false event once explicitly told that one event had been fabricated.

But several were visibly surprised when researchers confirmed it.

One participant's reaction was essentially: if nobody had told me it was false, I would have left believing I had done it.

Others expressed disbelief because they felt that they could remember parts of it.

The photograph had not shown the invented event.

It had supplied the world in which the invented event could take place.

And strangely, the photograph was not even the strongest suggestion

After the balloon study, Garry and Wade directly compared fake photographs with verbal descriptions.

You might expect photographs to dominate.

They did not.

Participants were presented with a false childhood balloon ride either through a doctored photograph or through a personalized written narrative.

Reports of imagery or memory occurred in about 50% of the photograph group but 82% of the narrative group.

Under stricter memory criteria, the difference was smaller but remained in the same direction.

One possible reason is that photographs constrain imagination.

A photograph determines what the balloon looked like, where people stood and what the surroundings contained.

A narrative leaves gaps.

And people can fill those gaps with material that already feels familiar to them.

A fictional childhood event may become more convincing when someone can unknowingly build it from pieces of their actual childhood.

That turns the simple idea that “photos cause false memories” into something more interesting.

The critical ingredient may not be photographs themselves.

It may be the opportunity to construct a plausible autobiographical scene and then lose track of where its components came from.

Not every invented memory works

The research also shows that memory is not infinitely suggestible.

Kathy Pezdek, Kimberly Finger and Dandle Hodge tested whether plausibility constrains false-memory creation.

In one experiment, Jewish and Catholic participants were given genuine childhood events alongside invented descriptions of religious experiences associated with each tradition.

False recollection was more likely when the suggested event fit the participant's own religious background.

In another experiment, participants were given two false childhood events.

One involved becoming lost while shopping.

The other involved receiving an enema.

False recollections occurred for the plausible shopping event.

None occurred for the deliberately less plausible enema event.

The mind, in other words, does not construct autobiographical memories out of nothing.

It uses what it already knows.

A suggested event fits more easily when it is compatible with existing knowledge about one's childhood, family, culture and environment.

What does it mean to “have” a false memory?

This question has become one of the major debates in the field.

Suppose a participant says:

“I can picture it.”

Is that a false memory?

What if they say:

“I think it probably happened.”

What if they describe ten details but repeatedly insist they are only guessing?

What if an independent researcher believes the participant sounds like they are remembering, but the participant later says they never truly experienced it as a memory?

Those distinctions can dramatically alter reported implantation rates.

Brewin and Andrews reviewed the false-autobiographical-memory literature and argued that studies have not always used identical definitions. When they applied more stringent criteria across studies, the proportion of participants showing some type of recollective experience was much larger than the proportion showing what they considered a convincing full false memory.

Across the implantation studies they reviewed, approximately 47% of participants showed some false recollective experience, whereas their stricter estimate for full false memories was around 15%.

Scoboria and colleagues approached the same problem by pooling transcript data from eight implantation studies involving 423 cases.

They coded the accounts for seven characteristics, including whether participants accepted the suggestion, added information beyond it, produced imagery, created a coherent story, expressed emotion and explicitly described remembering.

Using that framework, 30.4% of reports were classified as false memories, with an additional group accepting aspects of the false event without meeting the full memory classification.

So there is no scientifically responsible single sentence such as “researchers can implant memories in half of people.”

Sometimes studies find rates near that level.

Sometimes much lower.

The event matters.

The interview method matters.

Imagery matters.

Personal details matter.

Photographs sometimes matter.

The number of sessions matters.

And, critically, the definition of memory matters.

Why can imagining something eventually feel like remembering it?

One major explanation comes from the source-monitoring framework, developed by Marcia Johnson and colleagues.

When we remember, the mind has to infer where mental information originated.

A visual scene in your head might have come from something you experienced.

Or something you imagined yesterday.

Or a story your mother told repeatedly.

Or a photograph.

Or a dream.

Usually we discriminate among these sources remarkably well.

But source monitoring is not perfect.

Imagined experiences can contain visual and emotional detail. Real memories can be vague. Stories heard repeatedly can become extremely familiar. A childhood photograph can contain accurate information even while the story attached to it is false.

Over repeated acts of imagination, the content can become familiar while the origin of that content becomes less obvious.

That is why researchers are particularly interested in the combination of suggestion + repeated retrieval attempts + imagery + autobiographical knowledge.

The person is not necessarily retrieving a hidden memory.

They may be constructing one.

This account fits modern descriptions of episodic memory as reconstructive rather than reproductive. Contemporary memory textbooks emphasize that remembering involves integrating stored information with knowledge, expectations, inference and current retrieval cues rather than replaying a literal copy of the past (Baddeley et al., 2020; Eysenck & Keane, 2020; Goldstein, 2019).

The unsettling part is not that memory fails

The most interesting lesson from these experiments is not that memory is “bad.”

Nor is it that every uncertain childhood memory should suddenly be distrusted.

The majority of participants in many implantation experiments do not develop clear false memories. The studies generally involve carefully constructed suggestive conditions, repeated interviews and personalized information. And ethical limits mean that researchers cannot experimentally implant every kind of real-world event, particularly severe trauma.

The findings are therefore evidence of a capacity, not proof that any specific memory is false.

What makes the experiments remarkable is something subtler.

People were not simply agreeing with researchers.

Some began with resistance.

They said they had never been in the balloon.

They said they had never seen the photograph.

They could not remember the event at all.

Then they tried to remember.

They imagined.

They searched genuine childhood knowledge.

They constructed possibilities.

Those possibilities became familiar.

More details appeared.

And for some participants, the final product eventually carried enough of the subjective qualities of remembering that it became difficult to distinguish from something they had lived.

The hot-air-balloon photograph did not contain a childhood memory.

It gave the mind somewhere to begin building one.

And the rest came from pieces that were real:

a real face.

real parents.

real childhood knowledge.

real feelings.

real imagination.

assembled into an event that never happened.

References

Baddeley, A., Eysenck, M. W., & Anderson, M. C. (2020). Memory (3rd ed.). Routledge.

Brewin, C. R., & Andrews, B. (2017). Creating memories for false autobiographical events in childhood: A systematic review. Applied Cognitive Psychology, 31(1), 2–23. https://doi.org/10.1002/acp.3220

Eysenck, M. W., & Keane, M. T. (2020). Cognitive psychology: A student's handbook (8th ed.). Psychology Press.

Garry, M., Manning, C. G., Loftus, E. F., & Sherman, S. J. (1996). Imagination inflation: Imagining a childhood event inflates confidence that it occurred. Psychonomic Bulletin & Review, 3(2), 208–214. https://doi.org/10.3758/BF03212420

Garry, M., & Wade, K. A. (2005). Actually, a picture is worth less than 45 words: Narratives produce more false memories than photographs do. Psychonomic Bulletin & Review, 12(2), 359–366. https://doi.org/10.3758/BF03196385

Goldstein, E. B. (2019). Cognitive psychology: Connecting mind, research, and everyday experience (5th ed.). Cengage.

Hyman, I. E., Jr., Husband, T. H., & Billings, F. J. (1995). False memories of childhood experiences. Applied Cognitive Psychology, 9(3), 181–197. https://doi.org/10.1002/acp.2350090302

Hyman, I. E., Jr., & Pentland, J. (1996). The role of mental imagery in the creation of false childhood memories. Journal of Memory and Language, 35(2), 101–117. https://doi.org/10.1006/jmla.1996.0006

Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring. Psychological Bulletin, 114(1), 3–28. https://doi.org/10.1037/0033-2909.114.1.3

Lindsay, D. S., Hagen, L., Read, J. D., Wade, K. A., & Garry, M. (2004). True photographs and false memories. Psychological Science, 15(3), 149–154. https://doi.org/10.1111/j.0956-7976.2004.01503002.x

Loftus, E. F., & Pickrell, J. E. (1995). The formation of false memories. Psychiatric Annals, 25(12), 720–725. https://doi.org/10.3928/0048-5713-19951201-07

Murphy, G., Dawson, C. A., Huston, C., Ballantyne, L., Barrett, E., Cowman, C. S., Fitzsimons, C., Maher, J., Ryan, K. M., & Greene, C. M. (2023). Lost in the mall again: A preregistered replication and extension of Loftus & Pickrell (1995). Memory, 31(6), 818–830. https://doi.org/10.1080/09658211.2023.2198327

Pezdek, K., Finger, K., & Hodge, D. (1997). Planting false childhood memories: The role of event plausibility. Psychological Science, 8(6), 437–441. https://doi.org/10.1111/j.1467-9280.1997.tb00457.x

Scoboria, A., Wade, K. A., Lindsay, D. S., Azad, T., Strange, D., Ost, J., & Hyman, I. E. (2017). A mega-analysis of memory reports from eight peer-reviewed false memory implantation studies. Memory, 25(2), 146–163. https://doi.org/10.1080/09658211.2016.1260747

Wade, K. A., Garry, M., Read, J. D., & Lindsay, D. S. (2002). A picture is worth a thousand lies: Using false photographs to create false childhood memories. Psychonomic Bulletin & Review, 9(3), 597–603. https://doi.org/10.3758/BF03196318

Wade, K. A., Garry, M., Nash, R. A., & Harper, D. N. (2010). Anchoring effects in the development of false childhood memories. Psychonomic Bulletin & Review, 17(1), 66–72. https://doi.org/10.3758/PBR.17.1.66