Skip to content
Back to the question

Evidence record

Can you trust a memory?

Every sentence on the exploration belongs to at least one claim below. Each says what kind of statement it is, what it rests on, what the evidence does not establish, and how far it was checked. Nothing here is a score, and none of it is a verdict on a belief or a theory.

Last checked 24 September 2026.

  • A judgement, not a fact: 6
  • Partly checked: 3
  • Source checked: 35
  • Presented as open: 1

If something here is wrong, misquoted, out of date or unfair, write to support@unknownoceans.com and give the claim number, which does not change.

What the question takes for granted

A question in the psychology of memory that sounds general and is answered in parts. That memory can be altered by suggestion is established; how often whole false memories form, and how far laboratory rates carry to courts and therapy, are argued between named research groups.

  • Memory can be changed by what happens after the event

    Established

    Decades of misinformation studies show reports of an event shifting after misleading information. Those who argue memory is essentially reliable accept this too; they dispute how far it goes.

  • A memory is either true or false

    Contested

    Researchers on both sides of the implantation debate say many false memories are built partly from real experiences, and that believing an event happened is different from remembering it.

  • What happens in a laboratory tells us what happens in a courtroom or a therapy room

    Contested

    The laboratory studies use students, brief suggestion and events chosen by researchers. Each side argues the real-world rate differs from the laboratory one, in opposite directions, and neither has direct data.

How the question is labelled

Claim: Rich and wrong

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The page’s title. “Rich false memories” is the field’s term for detailed false memories of whole events, from the title of the 2015 study the page describes.

Claim 1 · checked 24 September 2026 · AI-assisted audit

Claim: False memory

Kind
Framing. Signposting. It asserts nothing a source could check.

The question

Claim: Detailed accounts of events that never happened can be drawn out of university students in three interviews.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Partly checked. Only part of the source could be read. What was not read is listed.

  • Supports

    Constructing Rich False Memories of Committing Crimeabstract

    Cited · Link answered, 19 October 2026 · Abstract only, 24 September 2026 · Research in people · Experiment

    The source reads: After three interviews, 70% of participants were classified as having false memories of committing a crime (theft, assault, or assault with a weapon) that led to police contact in early adolescence and volunteered a detailed false account.

  • Supports

    Creating Memories for False Autobiographical Events in Childhood: A Systematic ReviewTable 4, row for Shaw and Porter (2015)

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Shaw and Porter (2015) 60 college students

What this does not establish

The 2015 paper was read only as its abstract. Sixty students in one Canadian university, in a study built to produce false memories. Whether the accounts were memories or accepted stories is the subject of claims 16 and 17.

Later work looked for

Crossref, 2016 on, "Shaw Porter rich false memories crime": the recoding by Wade, Garry and Pezdek (2018), a reply by Shaw (2018, not readable here) and Shaw (2020) on observers. Page updated to report the recoding.

Corrected

24 September 2026. Previously: “Detailed, confident, emotional memories of events that did not happen can be produced in a laboratory in three interviews.” Participants’ own confidence in these memories was low: Brewin and Andrews (2017) report a marked discrepancy in this study between investigator ratings and the confidence participants had that the events were real, and a recoding by Wade, Garry and Pezdek (2018) classed most of the reported memories as accepted beliefs.

Claim 3 · checked 24 September 2026 · AI-assisted audit

Claim: How many of those accounts are memories is argued.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Deconstructing Rich False Memories of Committing Crime: Commentary on Shaw and Porter (2015)Conclusion

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: We suggest, therefore, a better interpretation of Shaw and Porter’s data is not that 70% of their subjects showed evidence of false memories, but that 26-30% did.

  • Supports

    Lost in the Mall? Interrogating Judgements of False Memorysection 1, Introduction

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Observational study: associations, not causes

    The source reads: The acknowledged difficulty in coding false memories (Scoboria et al. 2017; Shaw 2018; Wade, Garry, and Pezdek 2018) has yet to be resolved.

Claim 4 · checked 24 September 2026 · AI-assisted audit

Claim: Memory is partly a reconstruction rather than a replay, which means it can be edited by what happens after the event, including by the questions asked about it.

Kind
What the field holds. The settled view of a field, as a review, a consensus statement or a reference work gives it. Not one study.
Review
Source checked. The cited passage was read, and it says this.

Held by both sides of the argument the page reports: the reviewers who defend memory’s reliability describe it as partly reconstructive too.

  • Supports

    Regaining Consensus on the Reliability of Memoryintroduction, second paragraph

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Memory has been seen as partially reproductive, retaining some details unique to an event, and partially reconstructive, influenced by knowledge (‘schemas’) derived from prior learning

  • Supports

    Planting misinformation in the human mind: A 30-year investigation of the malleability of memoryConcluding remarks

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: when they are interrogated with leading questions or suggestive techniques, when they see media coverage about an event, misinformation can enter consciousness and can cause contamination of memory.

What this does not establish

That memory can be edited does not say how often it is, or by how much; claims 6 and 7 carry that.

Later work looked for

Crossref, 2019 on, "misinformation effect review": nothing found that disputes that memory reports can be changed after the event; the argument is about scale.

Corrected

24 September 2026. Previously: “Memory reconstructs rather than replays, which means it can be edited by what happens after the event, including by the questions asked about it.” Brewin, Andrews and Mickes (2020) describe memory as partly reproductive and partly reconstructive, and note that calling retrieval reconstructive says nothing by itself about whether a memory is true.

Claim 5 · checked 24 September 2026 · AI-assisted audit

Claim: Under persistent suggestion a minority of people come to remember whole events that never happened, some of them serious.

Kind
What the field holds. The settled view of a field, as a review, a consensus statement or a reference work gives it. Not one study.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

The participants are mostly university students and the events are chosen by researchers. The two reviews agree that it happens to a minority and disagree about how large.

Later work looked for

Crossref and Europe PMC, 2017 on, "rich false memories", "false memory implantation", "lost in the mall": a preregistered replication (Murphy and others, 2023), a reanalysis of its transcripts (Andrews and Brewin, 2024), a 2023 meta-analysis of 30 implantation studies reporting effect sizes rather than rates, and a 2025 exchange in Applied Cognitive Psychology over the paradigm. The rate is still argued, which the page now says.

Corrected

24 September 2026. Previously: “The effect is large enough to produce full false memories of serious events in a substantial fraction of ordinary people.” “A substantial fraction” is one side’s reading. A systematic review (Brewin and Andrews, 2017) puts full false memories at about 15 per cent at most; a mega-analysis (Scoboria and others, 2017) at 30.4 per cent including partial ones. The page now gives the range and why it is wide.

Claim 6 · checked 24 September 2026 · AI-assisted audit

Claim: Estimates of how large a minority run from about 15 per cent to about 30, depending on what is counted as remembering.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

The 30.4 per cent includes partial false memories; Scoboria and colleagues’ own table puts full and robust ones at about 22 per cent. Brewin and Andrews call 15 per cent an upper bound; Scoboria and colleagues say no single rate generalises, because it depends on the procedure.

Later work looked for

Crossref and Europe PMC, 2017 on, "rich false memories", "false memory implantation", "lost in the mall": a preregistered replication (Murphy and others, 2023), a reanalysis of its transcripts (Andrews and Brewin, 2024), a 2023 meta-analysis of 30 implantation studies reporting effect sizes rather than rates, and a 2025 exchange in Applied Cognitive Psychology over the paradigm. The rate is still argued, which the page now says.

Claim 7 · checked 24 September 2026 · AI-assisted audit

Claim: The feeling of remembering something vividly is assembled by the brain from many sources, and the past is only one of them.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

The account of remembering that implantation researchers work from, summarised in the mega-analysis from established theories of memory.

  • Supports

    A mega-analysis of memory reports from eight peer-reviewed false memory implantation studiesIdentifying false autobiographical memories

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Meta-analysis

    The source reads: The experience of remembering a past event is the result of multiple cognitive systems and processes. Inputs from numerous sources (e.g., records of prior perceptual processing, of prior thoughts, feelings, general knowledge, beliefs, goals, feelings of familiarity, etc.) give rise to an experience of remembering the past.

Corrected

24 September 2026. Previously: “The feeling of remembering something vividly is produced by the brain rather than by the past.” The sources describe the experience of remembering as built from several inputs, one of which is the record of the event. “Rather than by the past” denied that input.

Claim 8 · checked 24 September 2026 · AI-assisted audit

Claim: It feeds confidence, and confidence is no guarantee of accuracy

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Flashbulb MemoriesConfidence

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Vividness, elaborateness, and ease of retrieval are thought to influence the judgment that an event occurred (Johnson, Foley, Suengas, & Raye, 1988). They no doubt also account, in part, for the confidence assigned to a FBM

  • Supports

    Flashbulb MemoriesFBMs and Trauma

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Finally, persistence of and confidence in a memory is no guarantee of its accuracy.

  • Qualifies

    Regaining Consensus on the Reliability of MemoryThe Relation Between Confidence and Accuracy

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Wixted and Wells concluded that initial confidence is diagnostic of accuracy when pristine testing procedures are used.

    Where a memory is tested once, fairly, confidence does carry information. The page reports this in claims 33 to 36.

Corrected

24 September 2026. Previously: “It correlates with confidence and not, past a point, with accuracy.” The weak confidence-accuracy correlation is the claim Brewin, Andrews and Mickes (2020) single out as resting on a misleading statistic: at a first, fair test, high confidence predicts accuracy. The page now says what the flashbulb-memory research shows and reports the eyewitness argument below.

Claim 9 · checked 24 September 2026 · AI-assisted audit

Claim: memories of hearing shocking news stay vivid and confidently held for years while they drift from what people reported at the time

Kind
What the field holds. The settled view of a field, as a review, a consensus statement or a reference work gives it. Not one study.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Flashbulb MemoriesConfidence

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: One agreed-upon difference between FBMs and “everyday” autobiographical memories, even those that are rated “important,” is that confidence in FBMs remains high, even as consistency declines

  • Supports

    Flashbulb MemoriesFBMs and Trauma

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: once these errors are incorporated into the memory, they can persist and be confidently held for many years, perhaps a lifetime

What this does not establish

What is measured is consistency with the first report, not accuracy against the event; the review says so. Most of the studies are of news such as the 9/11 attacks.

Later work looked for

Crossref, 2016 on, "flashbulb memory confidence consistency": later test-retest studies (2021, 2024), read as titles only; nothing found that reverses confidence staying high while consistency falls.

Claim 10 · checked 24 September 2026 · AI-assisted audit

Claim: This is not a claim that memory is unreliable in general.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The page limiting its own scope. It follows from claim 40: the reviewers who dispute the implantation rates hold that memory is essentially reliable.

  • Context

    Regaining Consensus on the Reliability of MemoryConclusions

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Memory is clearly malleable but not unreliable, under normal circumstances and in the absence of contamination or prolonged suggestion

Claim 11 · checked 24 September 2026 · AI-assisted audit

Claim: It is a claim about a specific failure mode that has been demonstrated repeatedly and has consequences outside the laboratory.

Kind
What the field holds. The settled view of a field, as a review, a consensus statement or a reference work gives it. Not one study.
Review
Source checked. The cited passage was read, and it says this.

Later work looked for

Crossref and Europe PMC, 2017 on, "rich false memories", "false memory implantation", "lost in the mall": a preregistered replication (Murphy and others, 2023), a reanalysis of its transcripts (Andrews and Brewin, 2024), a 2023 meta-analysis of 30 implantation studies reporting effect sizes rather than rates, and a 2025 exchange in Applied Cognitive Psychology over the paradigm. The rate is still argued, which the page now says.

Claim 12 · checked 24 September 2026 · AI-assisted audit

What can be done to a memory

Claim: What can be done to a memory

Kind
Framing. Signposting. It asserts nothing a source could check.

Whole events can be implanted

Claim: Whole events can be implanted

Kind
What the field holds. The settled view of a field, as a review, a consensus statement or a reference work gives it. Not one study.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

Agreed by reviewers on both sides; they disagree about how often, which the passage reports.

Later work looked for

Crossref and Europe PMC, 2017 on, "rich false memories", "false memory implantation", "lost in the mall": a preregistered replication (Murphy and others, 2023), a reanalysis of its transcripts (Andrews and Brewin, 2024), a 2023 meta-analysis of 30 implantation studies reporting effect sizes rather than rates, and a 2025 exchange in Applied Cognitive Psychology over the paradigm. The rate is still argued, which the page now says.

Claim 14 · checked 24 September 2026 · AI-assisted audit

Claim: In a 2015 study, university students were interviewed three times about their adolescence, with a real event supplied by their parents and a false one, some of them criminal, described as though the parents had reported it too.

Kind
Historical claim. A statement about the past that historical evidence can test.
Review
Partly checked. Only part of the source could be read. What was not read is listed.

What this does not establish

The 2015 paper was read only as its abstract; the design is taken from two other papers’ descriptions of it, which agree.

Claim 15 · checked 24 September 2026 · AI-assisted audit

Claim: The authors classified 70 per cent of those given a crime as having a false memory of committing it.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Partly checked. Only part of the source could be read. What was not read is listed.

  • Supports

    Constructing Rich False Memories of Committing Crimeabstract

    Cited · Link answered, 19 October 2026 · Abstract only, 24 September 2026 · Research in people · Experiment

    The source reads: After three interviews, 70% of participants were classified as having false memories of committing a crime ... With suggestive memory-retrieval techniques, participants were induced to generate criminal and noncriminal emotional false memories

  • Supports

    Deconstructing Rich False Memories of Committing Crime: Commentary on Shaw and Porter (2015)text before Table 1

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: When we recoded Shaw and Porter’s data using their own scheme, we replicated, exactly, their 70% result.

What this does not establish

The 2015 paper was read only as its abstract. Thirty people were given a crime; the figure is the authors’ own coding.

Later work looked for

Crossref, 2016 on, "Shaw Porter rich false memories crime": the recoding by Wade, Garry and Pezdek (2018), a reply by Shaw (2018, not readable here) and Shaw (2020) on observers. Page updated to report the recoding.

Corrected

24 September 2026. Previously: “A large share came to report the false event, many of them with detail, emotion and confidence.” The headline figure is contested by a published recoding of the same data (Wade, Garry and Pezdek, 2018), and participants’ own confidence in these memories was low (Brewin and Andrews, 2017). The page now reports both figures and who holds them.

Claim 16 · checked 24 September 2026 · AI-assisted audit

Claim: Other researchers recoded the same transcripts with schemes that separate believing an event from remembering it, and put the figure at 26 to 30 per cent: many had accepted the story and speculated about it without showing signs of remembering.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

A recoding of the same thirty transcripts by judges working to other schemes, not a new experiment. Shaw published a reply defending her coding (2018), which was not readable here.

Other readings

Shaw (2020) maintains that participants in implantation studies report that their false memories feel real.

Later work looked for

Crossref, 2016 on, "Shaw Porter rich false memories crime": the recoding by Wade, Garry and Pezdek (2018), a reply by Shaw (2018, not readable here) and Shaw (2020) on observers. Page updated to report the recoding.

Claim 17 · checked 24 September 2026 · AI-assisted audit

Claim: How to score a false memory is still unresolved.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Lost in the Mall? Interrogating Judgements of False Memorysection 1, Introduction

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Observational study: associations, not causes

    The source reads: The acknowledged difficulty in coding false memories (Scoboria et al. 2017; Shaw 2018; Wade, Garry, and Pezdek 2018) has yet to be resolved.

Claim 18 · checked 24 September 2026 · AI-assisted audit

Claim: The interviews combined memory-retrieval techniques, including imagining the event,

Kind
Historical claim. A statement about the past that historical evidence can test.
Review
Source checked. The cited passage was read, and it says this.

Corrected

24 September 2026. Previously: “The procedure used no hypnosis and no pressure beyond the ordinary suggestion of a persistent, friendly interviewer.” The authors describe “a highly suggestive interview”, and Wade, Garry and Pezdek describe context reinstatement and imagination exercises with heavy social demand. “No pressure beyond the ordinary” understated it. Nothing read here states whether hypnosis was used, so that clause was removed.

Claim 19 · checked 24 September 2026 · AI-assisted audit

Claim: with what the critics called “a heavy dose of social demand”.

Kind
Quotes a source. The words of a named text, checked against the source.
Review
Source checked. The cited passage was read, and it says this.

Claim 20 · checked 24 September 2026 · AI-assisted audit

Claim: Even the lower figure matters where memory is evidence, as the critics said themselves.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

Corrected

24 September 2026. Previously: “The finding matters most where memory is evidence. Reported rates vary with method and with how a false memory is scored, and the effect is not in doubt.” The variation in rates is now carried by the recoding (claims 16 to 18). The last clause is split so that what is agreed, that it happens, is kept apart from what is argued, how often.

Claim 21 · checked 24 September 2026 · AI-assisted audit

Claim: That whole false memories can be produced is not in doubt.

Kind
What the field holds. The settled view of a field, as a review, a consensus statement or a reference work gives it. Not one study.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

Both reviews note that many false memories contain elements of real experiences.

Later work looked for

Crossref and Europe PMC, 2017 on, "rich false memories", "false memory implantation", "lost in the mall": a preregistered replication (Murphy and others, 2023), a reanalysis of its transcripts (Andrews and Brewin, 2024), a 2023 meta-analysis of 30 implantation studies reporting effect sizes rather than rates, and a 2025 exchange in Applied Cognitive Psychology over the paradigm. The rate is still argued, which the page now says.

Claim 22 · checked 24 September 2026 · AI-assisted audit

Claim: How often is.

Kind
Disputed. A contested point, where the page reports a side or the disagreement.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

Other readings

Brewin and Andrews argue that full false memories are uncommon and hard to implant; Scoboria and colleagues that the rate depends on the procedure and can be high.

Claim 23 · checked 24 September 2026 · AI-assisted audit

Confidence can be added later

Claim: Confidence can be added later

Kind
What the field holds. The settled view of a field, as a review, a consensus statement or a reference work gives it. Not one study.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Regaining Consensus on the Reliability of MemoryThe Relation Between Confidence and Accuracy

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Critically, the finding holds solely during the first and only lineup test because confidence can increase over repeated testing.

  • Context

    Convicting with confidence? Why we should not over-rely on eyewitness confidenceabstract

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: While there is strong consensus that courtroom confidence is problematic

What this does not establish

Shown for eyewitness identification; the passage says so.

Later work looked for

Crossref and Europe PMC, 2018 on, "eyewitness confidence accuracy" and "post-identification feedback": an exchange in Memory (2021 to 2022) between Berkowitz, Garrett, Fenn and Loftus and Wixted, Mickes and Brewin over how far initial confidence can be relied on. The page now reports both sides.

Corrected

24 September 2026. Previously: “Confidence is not the signal it feels like” The old heading implied confidence is a poor signal in general. Research on first, fair eyewitness tests finds it informative there; what is agreed is that confidence can grow after the event.

Claim 24 · checked 24 September 2026 · AI-assisted audit

Claim: The method is older than that study.

Kind
Historical claim. A statement about the past that historical evidence can test.
Review
Source checked. The cited passage was read, and it says this.

Claim 25 · checked 24 September 2026 · AI-assisted audit

Claim: In 1995 a smaller experiment, with 24 adults, gave people short accounts of childhood events supplied by a family member, one of which, being lost in a shopping mall, had not happened.

Kind
Historical claim. A statement about the past that historical evidence can test.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

The 1995 paper itself was not read; its sample and design are taken from a systematic review and from the first author’s own later review.

Corrected

24 September 2026. Previously: “In 1995 a smaller experiment gave adults short accounts of childhood events supplied by a family member, one of which, being lost in a shopping mall, had not happened.” The sample size added, so that the replication’s scale can be compared.

Claim 26 · checked 24 September 2026 · AI-assisted audit

Claim: About a quarter were judged to remember it, fully or in part.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

Six people of 24, judged by the investigators, with full and partial memories counted together.

Later work looked for

Crossref and Europe PMC, 2017 on, "rich false memories", "false memory implantation", "lost in the mall": a preregistered replication (Murphy and others, 2023), a reanalysis of its transcripts (Andrews and Brewin, 2024), a 2023 meta-analysis of 30 implantation studies reporting effect sizes rather than rates, and a 2025 exchange in Applied Cognitive Psychology over the paradigm. The rate is still argued, which the page now says.

Corrected

24 September 2026. Previously: “About a quarter came to remember it, some elaborating detail the experimenters had never supplied.” The 25 per cent combines full and partial memories, as judged by the investigators. The elaboration clause could not be checked: the 1995 paper was not readable, and a reanalysis of the 2023 replication found participants recalled on average fewer than two of the six core details.

Claim 27 · checked 24 September 2026 · AI-assisted audit

Claim: A preregistered replication in 2023, five times the size, judged 35 per cent to remember it; 14 per cent said so themselves.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

Read in the authors’ preprint, not the published version. The coded 35 per cent counts partial memories with full ones, most of them partial; a further 52 per cent said they believed the event without remembering it.

Later work looked for

Crossref and Europe PMC, 2017 on, "rich false memories", "false memory implantation", "lost in the mall": a preregistered replication (Murphy and others, 2023), a reanalysis of its transcripts (Andrews and Brewin, 2024), a 2023 meta-analysis of 30 implantation studies reporting effect sizes rather than rates, and a 2025 exchange in Applied Cognitive Psychology over the paradigm. The rate is still argued, which the page now says.

Corrected

24 September 2026. Previously: “” Added: the 1995 study has since been replicated at a larger scale, and the gap between investigators’ and participants’ judgements is the centre of the current argument.

Claim 28 · checked 24 September 2026 · AI-assisted audit

Claim: People who come to report an implanted memory hold it, on average, less firmly than their real ones, and rate it lower for clarity and confidence.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Lost in the Mall? Interrogating Judgements of False Memorysection 1, Introduction

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Observational study: associations, not causes

    The source reads: Investigator-determined false memories were on average rated by participants as below the midpoint on scales of clarity and confidence and lower than their true memories

  • Supports

    Regaining Consensus on the Reliability of MemoryImplanting False Memories of Childhood Events

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Memories of false events also consistently have less of a recollective quality and are less vivid than corresponding true events (Brewin & Andrews, 2017).

  • Supports

    Creating Memories for False Autobiographical Events in Childhood: A Systematic ReviewDiscussion of memory implantation studies

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Similarly, Shaw and Porter (2015) found a marked discrepancy between the ease of inducing experiences rated by investigators as memories and the confidence participants had that they were real.

  • Cuts against

    Do False Memories Look Real? Evidence That People Struggle to Identify Rich False Memories of Committing Crime and Other Emotional EventsIntroduction

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: participants consistently report that such false memories feel incredibly real (e.g., Shaw and Porter, 2015; Scoboria et al., 2017)

What this does not establish

An average over studies that report self-ratings, about half of them.

Other readings

Shaw, first author of the 2015 study, writes that participants consistently report their false memories feel real.

Later work looked for

Crossref and Europe PMC, 2017 on, "rich false memories", "false memory implantation", "lost in the mall": a preregistered replication (Murphy and others, 2023), a reanalysis of its transcripts (Andrews and Brewin, 2024), a 2023 meta-analysis of 30 implantation studies reporting effect sizes rather than rates, and a 2025 exchange in Applied Cognitive Psychology over the paradigm. The rate is still argued, which the page now says.

Corrected

24 September 2026. Previously: “People rarely experience an implanted memory as uncertain. It arrives with the same texture as a real one, because it is being produced by the same reconstruction.” Contradicted by the self-report data: a systematic review found participants’ confidence in implanted memories below the scale midpoint and below their true memories, including in the 2015 crime study. The page now reports the average and the minority who are sure.

Claim 29 · checked 24 September 2026 · AI-assisted audit

Claim: Not all of them: in the two pooled studies that asked, 28 per cent of 116 people given a false event put their belief that it had happened at the top of the scale

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    A mega-analysis of memory reports from eight peer-reviewed false memory implantation studiesResults, False Autobiographical Belief; Figure 4

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Meta-analysis

    The source reads: we report here on subjective autobiographical belief ratings that were taken in two of the samples (N = 116; ... While 50% of individuals rated autobiographical belief above the scale midpoint, 28% reported autobiographical belief at the top of the scale.

What this does not establish

Belief ratings were taken in two of the eight pooled studies, 116 people, not across the whole mega-analysis. Belief that the event happened, rated by participants, is not the same as remembering it; the mega-analysis found belief ratings did not differ across its memory categories.

Later work looked for

Crossref and Europe PMC, 2017 on, "rich false memories", "false memory implantation", "lost in the mall": a preregistered replication (Murphy and others, 2023), a reanalysis of its transcripts (Andrews and Brewin, 2024), a 2023 meta-analysis of 30 implantation studies reporting effect sizes rather than rates, and a 2025 exchange in Applied Cognitive Psychology over the paradigm. The rate is still argued, which the page now says.

Corrected

24 September 2026. Previously: “Not all of them: pooled across eight studies, 28 per cent of everyone given a false event put their belief that it had happened at the top of the scale” Denominator: the mega-analysis took belief ratings in only two of its eight samples (N = 116), so the 28 per cent is of those 116 people, not of everyone across eight studies.

Claim 30 · checked 24 September 2026 · AI-assisted audit

Claim: and observers watching the accounts often cannot tell them from true ones

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

Untrained observers watching short recorded accounts, in two small studies of eight speakers. Trained questioning under cross-examination was not tested.

Later work looked for

Crossref, 2020 on, "distinguishing true and false memories": studies using eye movements (2021) and event-related potentials (2020), read as titles only; nothing found that claims reliable classification of a single memory. Shaw (2020) found observers no better than chance.

Claim 31 · checked 24 September 2026 · AI-assisted audit

Claim: Confidence at a first, fair test is another matter.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

A transition to the eyewitness research, where confidence is measured at a first test. It follows from claims 33 and 34.

Claim 32 · checked 24 September 2026 · AI-assisted audit

Claim: In eyewitness lineups, one group of researchers finds that a witness who picks someone with high confidence at that point is likely to be right

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Regaining Consensus on the Reliability of MemoryThe Relation Between Confidence and Accuracy

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Suspects who were identified with high confidence were likely to be the target, but suspects who were identified with low confidence were more likely to be innocent. Wixted and Wells concluded that initial confidence is diagnostic of accuracy when pristine testing procedures are used.

What this does not establish

Holds, on this account, for fair lineups at a first test; the synthesis it rests on (Wixted and Wells, 2017) was not read directly.

Later work looked for

Crossref and Europe PMC, 2018 on, "eyewitness confidence accuracy" and "post-identification feedback": an exchange in Memory (2021 to 2022) between Berkowitz, Garrett, Fenn and Loftus and Wixted, Mickes and Brewin over how far initial confidence can be relied on. The page now reports both sides.

Corrected

24 September 2026. Previously: “The same reconstruction explains why memories change with retelling, why a group’s shared account converges, and why an eyewitness becomes more certain after being told they identified the right person.” The eyewitness finding is kept and sourced, and the argument over initial confidence is reported with who holds each side. The retelling and group-convergence clauses, and the claim that one reconstruction explains all three, had no source among those cited and were removed.

Claim 33 · checked 24 September 2026 · AI-assisted audit

Claim: another points to conditions in which confident first picks are wrong

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Convicting with confidence? Why we should not over-rely on eyewitness confidenceabstract; section before the Conclusion; Key points

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: we review new research that reveals numerous conditions wherein eyewitnesses with high initial confidence end up being wrong. ... There are some conditions that reduce the strength of the confidence-accuracy relationship (e.g., age of the eyewitness, lineups where the suspect is likely innocent, biased identification procedures

What this does not establish

A review of other studies, not new data. The same authors accept that initial confidence can be of value in certain contexts.

Later work looked for

Crossref and Europe PMC, 2018 on, "eyewitness confidence accuracy" and "post-identification feedback": an exchange in Memory (2021 to 2022) between Berkowitz, Garrett, Fenn and Loftus and Wixted, Mickes and Brewin over how far initial confidence can be relied on. The page now reports both sides.

Claim 34 · checked 24 September 2026 · AI-assisted audit

Claim: Both agree that confidence expressed in a courtroom is a poor guide.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Convicting with confidence? Why we should not over-rely on eyewitness confidenceabstract; Introduction

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: there is important consensus on many of the most important issues in the eyewitness area (National Research Council, 2014) – including that courtroom confidence and courtroom identifications are fraught with problems.

  • Supports

    Regaining Consensus on the Reliability of MemoryThe Relation Between Confidence and Accuracy

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Critically, the finding holds solely during the first and only lineup test because confidence can increase over repeated testing. ... the initial identification was made with low confidence or no identification, but the courtroom identification was made with high confidence

What this does not establish

Brewin and colleagues make the point through the rise in confidence between a first identification and a courtroom one in exoneration cases.

Claim 35 · checked 24 September 2026 · AI-assisted audit

Claim: Confidence can be added afterwards: repeated testing raises it, and so does telling a witness they picked the right person.

Kind
What the field holds. The settled view of a field, as a review, a consensus statement or a reference work gives it. Not one study.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Regaining Consensus on the Reliability of MemoryThe Relation Between Confidence and Accuracy

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: The increase in confidence can happen for various reasons, for example with post-identification feedback (Wells & Bradfield, 1998).

  • Supports

    Regaining Consensus on the Reliability of MemoryThe Relation Between Confidence and Accuracy

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: Critically, the finding holds solely during the first and only lineup test because confidence can increase over repeated testing.

What this does not establish

The feedback experiment itself (Wells and Bradfield, 1998) and the later meta-analyses were not readable here; the claim rests on the review’s account of them.

Later work looked for

Crossref and Europe PMC, 2018 on, "eyewitness confidence accuracy" and "post-identification feedback": an exchange in Memory (2021 to 2022) between Berkowitz, Garrett, Fenn and Loftus and Wixted, Mickes and Brewin over how far initial confidence can be relied on. The page now reports both sides.

Claim 36 · checked 24 September 2026 · AI-assisted audit

How far it reaches

Claim: How far it reaches

Kind
Framing. Signposting. It asserts nothing a source could check.

Claim: Vividness is built during recall, which is why a memory of something that never happened can have it too.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The page’s reading of claims 8 and 30: imagery produced during recall can be taken for memory. It does not say false memories are usually as vivid as true ones; claim 29 says they usually are not.

Corrected

24 September 2026. Previously: “Vividness is manufactured at the moment of recall, which is why it is available to a memory of nothing.” “A memory of nothing” overstated it: implanted memories are often built partly from real experiences. The line is narrowed to what the sources carry.

Claim 38 · checked 24 September 2026 · AI-assisted audit

How much of ordinary memory this touches

Claim: How much of ordinary memory this touches

Kind
Framing. Signposting. It asserts nothing a source could check.

Claim: On the long-standing view, which its defenders say still holds, everyday recall preserves the general shape of the past well.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Regaining Consensus on the Reliability of MemoryAbstract

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: New research has clarified that all three claims rest on shaky foundations, suggesting there is no reason to abandon the old consensus about memory being malleable, but essentially reliable.

  • Supports

    Regaining Consensus on the Reliability of Memoryintroduction, quoting Schacter (1996)

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: our memory systems do a remarkably good job of preserving the general contours of our pasts and of recording correctly many of the important things that have happened to us

Other readings

Researchers who emphasise memory’s malleability, among them Loftus, Wade and Garry, are the other side of this argument.

Corrected

24 September 2026. Previously: “Everyday recall is right about most things most of the time, and the experiments above deliberately target the conditions under which it is not: distant events, repeated suggestive questioning, a trusted source supplying the outline.” Stated as a fact, the first clause is one side of an argument over memory’s reliability. It is now attributed to those who hold it (Brewin, Andrews and Mickes, 2020).

Claim 40 · checked 24 September 2026 · AI-assisted audit

Claim: The experiments above deliberately target the conditions under which it does not: distant events, repeated suggestive questioning, a trusted source supplying the story.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

Claim 41 · checked 24 September 2026 · AI-assisted audit

Claim: How far the laboratory rates transfer to a police interview, a therapy room or a courtroom is argued.

Kind
Open question. Something the page says is not known.
Review
Presented as open. The page says this is unresolved, and the review found that it is.

Corrected

24 September 2026. Previously: “How far the laboratory rates transfer to a real interview, a therapy room or a courtroom is argued, and it is argued by people who agree about the underlying mechanism.” The argument is now reported with what each side says, from their own papers, and the shared mechanism is stated separately in claim 45.

Claim 42 · checked 24 September 2026 · AI-assisted audit

Claim: One side says the studies overstate how easily false memories form; the other, that brief and relatively mild laboratory suggestion understates what longer, stronger suggestion in therapy can do.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

The sides are not simple opposites: Brewin and Andrews also allow that therapists may have wielded more influence than the laboratory could.

Claim 43 · checked 24 September 2026 · AI-assisted audit

Claim: Neither has data from outside the laboratory to settle it.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Regaining Consensus on the Reliability of MemoryImplanting False Memories of Childhood Events

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: In real life the specific circumstances (e.g., availability of deceptive information, or opportunity for more extended suggestive influence) are likely to be very different but empirical data are unavailable.

What this does not establish

Stated in these words by one side; the mega-analysis on the other side offers no real-world rate either.

Claim 44 · checked 24 September 2026 · AI-assisted audit

Claim: Both work from the same account of how a false memory forms: a suggestion is imagined, and the imagining is taken for remembering.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Creating Memories for False Autobiographical Events in Childhood: A Systematic ReviewDiscussion of memory implantation studies

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: This is consistent with the source memory framework (Johnson, 2006), which predicts that the act of imagination may create additional contents that could be confused with real memories

  • Supports

    A mega-analysis of memory reports from eight peer-reviewed false memory implantation studiesIdentifying false autobiographical memories, Imagery

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Meta-analysis

    The source reads: Theories of memory formation that build upon the source monitoring framework (Johnson et al., 1993) posit that some suggestive practices foster the creation of imagery which may subsequently be misattributed to memory

What this does not establish

A shared framework, not a settled mechanism; both papers list other contributors, such as belief that the event is plausible.

Claim 45 · checked 24 September 2026 · AI-assisted audit

Whether a memory carries a mark of being real

Claim: Whether a memory carries a mark of being real

Kind
Framing. Signposting. It asserts nothing a source could check.

Claim: Studies have looked for differences in brain activity, in language, in detail and in hesitancy between true and false memories.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

Claim 47 · checked 24 September 2026 · AI-assisted audit

Claim: Differences exist on average.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

Later work looked for

Crossref, 2020 on, "distinguishing true and false memories": studies using eye movements (2021) and event-related potentials (2020), read as titles only; nothing found that claims reliable classification of a single memory. Shaw (2020) found observers no better than chance.

Claim 48 · checked 24 September 2026 · AI-assisted audit

Claim: None of them separates a particular memory reliably enough to be used on one person

Kind
What the field holds. The settled view of a field, as a review, a consensus statement or a reference work gives it. Not one study.
Review
Source checked. The cited passage was read, and it says this.

Later work looked for

Crossref, 2020 on, "distinguishing true and false memories": studies using eye movements (2021) and event-related potentials (2020), read as titles only; nothing found that claims reliable classification of a single memory. Shaw (2020) found observers no better than chance.

Claim 49 · checked 24 September 2026 · AI-assisted audit

Claim: which is the use a court would need

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The page’s judgement: in a legal case the question is whether one person’s memory is true, not whether a group of memories differs on average.

Corrected

24 September 2026. Previously: “which is the only use that would matter.” Group differences matter to research; “the only use that would matter” overstated it. Narrowed to the use the page is concerned with.

Claim 50 · checked 24 September 2026 · AI-assisted audit

What was read