top of page

Psychological Encyclopedia

Eyewitness Memory: Accuracy, Suggestibility, and Confidence

3 days ago
20 min read

Author: Angela Bogdanova · Published: October 12, 2026 · Editorial Policy

Research cluster: Memory & Forgetting · Written by Artificial Sapiens

Eyewitness memory is the human capacity to remember people, actions, objects, words, and circumstances that were perceived during an event and may later matter to an investigation. It can preserve valuable evidence, sometimes with remarkable accuracy, while remaining subject to limits of perception, ordinary forgetting, source confusion, suggestion, and changes introduced during retrieval. Accuracy concerns whether remembered information corresponds to the event. Confidence concerns a witness's assessment of that correspondence. Their relationship depends critically on the conditions in which memory was tested.

A central conclusion of contemporary psychological science is that eyewitness memory is informative under identifiable conditions and vulnerable under others. An initial account obtained without suggestion differs from a narrative rehearsed after conversations, photographs, media exposure, and confirming feedback. A first identification using a properly designed lineup differs from a later courtroom identification. These distinctions are foundational to the National Academies' consensus report, the American Psychology–Law Society's scientific recommendations, and the modern synthesis of eyewitness memory research.

This article explains eyewitness memory as a cognitive phenomenon: what witnesses encode and retrieve, why accurate and mistaken recollections can both feel compelling, how suggestion affects reports, when confidence predicts identification accuracy, and which evidence-preservation practices follow from research. Detailed courtroom procedure and legal doctrine belong to separate topics.

What Is Eyewitness Memory in Psychology?

Eyewitness memory is episodic remembering for an event a person directly observed, including its participants and circumstances. The content may be visual, auditory, spatial, verbal, or action-related. Someone might remember a vehicle's route, a person's face, an object's appearance, the order of events, or a sentence uttered during an encounter. Remembering an event does not guarantee that every element was attended to or retained.

Researchers distinguish eyewitness event recall from eyewitness identification. Recall asks a person to generate information, such as describing what happened. Identification asks whether a person seen now was encountered during the event. Source monitoring asks how the witness learned a particular detail: directly observed, inferred, overheard, read later, or encountered in an image. These functions overlap yet have different opportunities for error. A witness can remember the sequence of a collision accurately and mistake the driver's identity; another can recognize a familiar face but misattribute that familiarity to the alleged event.

Memory accuracy also differs from sincerity. A mistaken witness may report their recollection in good faith and experience it as vivid, personally certain, and emotionally meaningful. Deliberate fabrication is a separate question about intent. Research on reconstructive memory shows why ordinary remembering can combine stored details, prior knowledge, retrieval cues, and later information. The Human Memory: Types, Processes, and How It Works guide provides the wider cognitive context.

How Eyewitness Memory Works: Encoding, Retention, and Retrieval

Encoding begins with perceptual opportunity

To remember a detail, a witness must first have had an opportunity to perceive it. Distance, illumination, exposure duration, viewing angle, motion, obstacles, visual acuity, and attentional focus all shape what can enter memory. Someone standing close to a person for several minutes has a different opportunity from someone who glimpses the person at night from a moving vehicle.

Attention is selective. When a situation is surprising, stressful, or fast-moving, some elements can attract attention while others receive little processing. A witness may encode an alarming action and overlook a face, license plate, or item of clothing. A detail not meaningfully perceived cannot be reliably reconstructed later simply by concentrating harder. The National Academies' scientific discussion of vision and memory emphasizes these fundamental perceptual limits.

Retention is a changing interval

Between an event and a later report, some details remain accessible, others become harder to retrieve, and new experiences compete with or inform the memory. Ordinary forgetting often reduces the amount of available detail over time. The size of that effect varies across persons, materials, exposures, and retrieval conditions. There is no scientifically justified universal expiration date after which all eyewitness memory becomes useless.

Post-event experiences can be especially consequential. A witness may speak with another observer, view a police photograph, watch a news clip, read social media, or rehearse a possible explanation. Such material can introduce genuinely new information without being clearly tagged as new. The crucial question is where each reported detail originated.

Retrieval is an active cognitive operation

Remembering requires cues and reconstruction. An open invitation to narrate an event, a specific question, and a face-recognition lineup activate different demands. Free recall can reveal what a person spontaneously remembers; recognition can elicit familiarity without full contextual recollection. Leading prompts can insert candidate details that were never independently reported.

Repeated narration can also increase familiarity with the narrative. Fluency, consistency, and ease of telling may grow without proportional improvement in factual correspondence. Researchers therefore examine the first statement and the conditions of each subsequent retrieval. A 2025 review of the cognitive science of eyewitness memory emphasizes the informational significance of the earliest uncontaminated recognition decision.

How Accurate Is Eyewitness Memory?

There is no single meaningful percentage representing the accuracy of all eyewitness memory. Accuracy depends on the particular task: recall of a sequence, identification of an unfamiliar face, recognition of an object, recollection of a phrase, or attribution of a detail to its original source. It also depends on what the study counts as correct and whether an independent ground truth exists.

A controlled experiment can show participants a video with a known person, then determine whether the participant selects that person, a known-innocent alternative, or nobody. A real investigation typically begins without such certain ground truth. A selection of a known-innocent lineup filler is an observable error; a selection of a police suspect is not automatically a confirmed success. When investigators classify every suspect selection as correct, the resulting accuracy claim is circular.

In some settings, witnesses recall central actions better than peripheral details. An individual may describe a robbery's sequence competently yet misidentify a stranger; another may recognize a person previously known to them but misremember a spoken statement. Memory for identity and memory for events should be evaluated separately rather than collapsed into one judgment of credibility.

Scientists use several measures: proportion of correct recalled details, identification of the actual target, erroneous filler choices, correct rejection when the target is absent, confidence calibration, discriminability, and the tendency to choose. These measures answer different questions. The National Academies' evaluation of identification procedures explains why experimental accuracy, discrimination, and decision tendencies should not be conflated.

Why Eyewitness Memories Become Inaccurate

Incomplete observation

The simplest source of error is incomplete original information. A brief glimpse, dark conditions, a partially obscured face, simultaneous tasks, and rapid movement may limit the diagnostic details encoded. Witnesses sometimes fill gaps with reasonable inferences about what probably occurred. Such inferences can become difficult to separate from perceived facts when recounting the event.

Forgetting, interference, and familiarity confusion

Ordinary forgetting can make details inaccessible. Interference can occur when similar experiences compete: repeated encounters with similar cars, faces, locations, or conversations make source discrimination harder. A witness might recognize an individual seen in a photograph after the event and mistake that familiarity for recognition from the event itself.

This distinction between familiarity and source memory is central. Familiarity can be psychologically genuine while its attributed origin is wrong. The person may feel that a face is known, and that experience alone does not identify the encounter responsible. When assessing eyewitness evidence, investigators need to document intervening photographs and encounters rather than infer the source from confidence.

The misinformation effect

The misinformation effect occurs when misleading information encountered after an event influences a subsequent memory report or recognition judgment. This information may be introduced by a question, another witness, a news report, or a persuasive statement. It can influence what a witness says, which detail is retrieved, and, under some conditions, what seems personally remembered.

In classic experiments, Loftus, Miller, and Burns (1978) presented visual events and subsequently varied the information participants received. Misleading material impaired recognition of the original detail. The experiments provided causal evidence that post-event information can distort performance under laboratory conditions. They did not establish that all such errors require permanent erasure of an earlier trace: source confusion, response processes, interference, and integration of information remain relevant explanations.

The existing article The Misinformation Effect in Eyewitness Memory covers that research paradigm in depth. Here the relevance is the broader question of eyewitness accuracy: a detail reported later may reflect a mixture of observation and post-event exposure, even when the witness is sincere.

Confirming feedback after identification

A witness who makes a tentative identification may subsequently hear, “You picked the person we suspected.” This feedback can increase retrospective confidence and change the witness's later memory for how certain they were at the time. In Bradfield, Wells, and Olson's experiment, confirming feedback inflated retrospective certainty and weakened the relationship between confidence and identification correctness, particularly in conditions involving mistaken identifications.

A confidence statement collected at trial after months of discussion and feedback therefore measures something different from an immediate statement collected at the first fair test. Both may be genuine expressions of the person's current belief. Their evidentiary significance is not equivalent.

Can Eyewitness Confidence Predict Accuracy?

Research supports a conditional relationship between initial eyewitness confidence and identification accuracy. The historically popular claim that witness confidence is always a poor predictor misses important findings from properly administered first memory tests. The opposite claim, that confident testimony proves a suspect was present, is also unsupported.

Wixted and Wells' major scientific synthesis found a strong relation between initial confidence and identification accuracy when an adult witness was tested using a fair lineup, without prior contamination or administrator influence, and confidence was recorded immediately. These conditions are important because they preserve the meaning of the original memory decision.

Why timing matters

Confidence is a metacognitive judgment about a memory decision. It can be reported at the first selection, after feedback, after repeated questioning, or during testimony. Later confidence is affected by experiences that occur after the original event and sometimes after the original identification. A witness may genuinely forget how hesitant they initially were.

Investigators therefore gain information by recording a witness's exact words immediately, before any congratulation, criticism, or disclosure about a suspect. “I think it is this person,” “I'm about 60 percent certain,” and “I am positive” should be preserved as distinct statements rather than converted to a uniform administrative label. Confidence about a face also differs from confidence about an event's sequence.

Confidence–accuracy calibration and discriminability

Confidence–accuracy calibration asks whether decisions made at each confidence level are correct at the corresponding frequency in a studied sample. If 80 percent of decisions expressed at 80 percent confidence are correct under a defined test, that group is well calibrated for the condition. The result cannot simply be translated into an 80 percent probability that a particular suspect is guilty. Case selection, suspect bias, prior exposure, and the base rates associated with the test also influence any case-specific inference.

Discriminability is a separate property: how well an identification procedure distinguishes the culprit from an innocent person. Response bias describes the willingness to select a lineup member versus reject the lineup. A conservative test might produce fewer innocent-person selections but also fewer correct selections, without necessarily improving perceptual discrimination. Scientists use receiver-operating-characteristic analyses to help separate these effects.

A 2026 review by Yu and Moore highlights individual differences in metacognitive ability, confidence bias, and task-specific competence. These factors remain active research questions when trying to generalize group calibration to individual witnesses.

High-confidence eyewitness errors still occur

A 2025 meta-analysis by Fitzgerald, Hyman, and Wade analyzed field identification records and found that roughly one in eight high-confidence identifications in the selected samples administered by blind lineup officers were known errors: choices of innocent fillers. Such errors are observable because fillers are known to be innocent. This figure is not a universal false-identification rate or a direct estimate of the proportion of innocent suspects mistakenly identified across all investigations.

The authors also discuss three sources of possible suspect bias: choosing a suspect because of appearance, contamination through social-media images, and wrongly attributing prior familiarity to the witnessed event. Their work illustrates how initial confidence can be useful under some conditions and highly confident decisions can still be mistaken, particularly when procedures or suspect selection introduce bias.

A 2025 cross-perspective research discussion describes growing agreement that confidence–accuracy relations vary predictably with suspect selection, test design, and the timing of confidence assessment. The separately focused English Hub article Eyewitness Confidence and Accuracy: When Confidence Is Informative develops the narrower identification-calibration question.

Lineups, Recognition, and the Problem of the Innocent Suspect

A lineup asks a witness to evaluate a suspect among known-innocent comparison individuals, commonly called fillers. In a target-present research lineup, the actual person from the event appears. In a target-absent lineup, the person is missing. A meaningful identification test must permit a rejection of all candidates.

The lineup's quality affects interpretation. If the suspect stands out because of clothing, age, facial hair, or some other distinctive cue absent from fillers, the witness may be guided toward that suspect. If the administrator knows which photo belongs to the suspect, their comments, timing, or facial expressions can inadvertently influence the choice. Instructions matter: a witness should understand that the observed person may not be present.

Sequential lineups present people one at a time; simultaneous lineups present multiple faces together. Early research was interpreted by some as favoring sequential presentations. Later cognitive-science analyses have found important advantages of simultaneous procedures under certain comparisons, particularly on diagnosticity, while the field continues to examine procedural details and tradeoffs. An absolute statement that one method always eliminates errors would exceed the evidence. The 2025 review explains this historical change in interpretation.

Witnesses may be asked to make multiple identifications, view social-media photographs, or identify a suspect in court. Repeated exposure increases possible familiarity sources and can make a subsequent decision less independent of the initial test. A proper evaluation tracks every exposure and preserves both initial selections and initial rejections.

The specialized published guide Eyewitness Identification: Lineups, Showups, and Best Practices examines the procedures in detail. This article's focus is what lineup decisions reveal about the witness's underlying memory.

Stress, Arousal, and Weapon Focus

How stress affects memory for events and faces

Threatening circumstances can change attention and encoding. Emotionally central details may feel durable, while a witness's ability to distinguish an unfamiliar face or describe peripheral details may be compromised. A meta-analysis by Deffenbacher and colleagues reported average impairment in eyewitness identification and some forms of detail recall under high-stress conditions across the included studies.

Stress research varies in induction, severity, timing, laboratory realism, and memory measure. The average result does not permit a conclusion that a stressed eyewitness's statement must be false or that a calm witness is necessarily accurate. The relevant questions concern what was perceived, which elements were central, and which identification or recall task occurred afterward.

Does a weapon automatically make identification unreliable?

Weapon focus is the hypothesis that a weapon captures attention and consequently reduces processing of other details, including the holder's face. Earlier experiments and syntheses found effects under certain conditions. Recent tests have challenged how broadly those findings generalize. West, Mulligan, and Bornstein reported preregistered failures to replicate a consistent weapon-focus effect in online experimental samples, despite varying stimuli and task conditions.

The appropriate inference is contextual. A weapon can be salient, but its presence alone does not supply a validated verdict on a witness's memory. Threat, duration, action, novelty, face visibility, and the design of the test all matter. Replication failures also require interpretation in light of the difference between viewing an online stimulus and witnessing an actual dangerous event.

Cross-Race Recognition, Familiarity, and Viewing Conditions

The cross-race effect refers to average differences in face recognition for faces classified as belonging to a viewer's own racial group and faces from another group. The literature considers perceptual experience, social categorization, contact, and recognition-test conditions. It is a population-level effect with substantial individual variation and does not establish that any named witness was correct or incorrect in a particular case.

A meta-analytic assessment of contact and cross-race recognition examines the role of experience. Related forensic research evaluates how differences in face discrimination and suspect bias interact. The proper case-specific questions are the witness's actual exposure and perceptual opportunity, not an assumption about accuracy based exclusively on group identity.

Duration, illumination, distance, disguise, viewing angle, and obstruction can interact. An individual familiar with the suspect from daily life differs from a witness who encountered the person once for a fraction of a second. Familiarity can assist recognition, while mistaken attribution of familiarity from another encounter can create a source error.

These conditions are sometimes called estimator variables because the investigator cannot retroactively control the viewing conditions of the event. They should be documented and evaluated without treating any one variable as a diagnostic test for the accuracy of the witness.

Eyewitness Memory Across the Lifespan

Children and older adults can provide accurate, useful eyewitness information, and performance differs by task and developmental context. A meta-analysis of age and eyewitness identification evaluated 91 studies involving more than 20,000 participants and found average age-group differences in discrimination and decision patterns. Young adults often performed better than children and older adults on the studied identification tasks, but these findings cannot determine the accuracy of an individual report.

Children's answers can be influenced by comprehension of instructions, interviewer expectations, repeated questioning, and the source of suggested details. Adults may also make suggestion-related errors. A developmental analysis of false-memory patterns describes situations in which adults show greater false-memory susceptibility than children. Thus the statement “children always have less reliable memory” does not reflect the full evidence.

Language, individual cognitive strengths, experience with faces, and opportunities to perceive differ within every age group. Best practices concern appropriate, non-leading questioning, clear documentation, and understanding the demands of the particular task. Witness age alone is never a sufficient proof of memory accuracy or inaccuracy.

Interviews, Leading Questions, and Repeated Accounts

An interview is a memory-retrieval environment. The phrasing of questions can assist access to information or contaminate the report by introducing unobserved details. An open invitation such as “Tell me everything you remember about what happened next” permits an account organized by the witness. A question such as “Did the man in the blue jacket enter the car?” presupposes a particular person and color that might not have been volunteered.

Repeated interviews can elicit additional genuine details as different cues become available. They also present opportunities for inadvertent suggestion, narrative rehearsal, source confusion, and changes in confidence. A discrepancy between statements may reflect forgetting, new retrieval, misunderstanding, altered cues, or contamination. The discrepancy alone cannot prove that the witness is dishonest.

The American Psychology–Law Society's 2020 recommendations support carefully documented initial interviews and preservation of first reports before unnecessary exposures. In some cases, an audio or video record lets later evaluators distinguish an originally volunteered detail from one supplied by a question.

Researchers also distinguish memory evidence from knowledge acquired after the event. A witness who accurately repeats a license plate number heard from another observer may possess a correct fact without ever having seen the plate. Establishing when and where that information was first encountered is essential to determining what eyewitness memory actually contributes.

What the Research Can and Cannot Establish

Classic laboratory experiments provide controlled tests of processes such as misinformation, confidence inflation, and face recognition. Their advantage is known ground truth: researchers determine precisely what participants saw and whether an answer is correct. Their limitations include artificial materials, varying motivational stakes, and incomplete resemblance to real incidents.

Field studies provide important ecological information but face a different difficulty: independent proof that the chosen suspect was guilty or innocent may be unavailable. Known-innocent filler identifications give a directly observable subset of errors, but suspect identifications require external verification. A study that ignores this limitation can overstate its success rate.

Research sometimes reaches apparently competing conclusions because studies evaluate different outcomes. A fair first lineup can yield a meaningful confidence–accuracy relationship, while high-confidence errors still occur in field datasets. A suggestion manipulation can change a response without establishing that the prior memory trace was erased. A lineup format can change selection rates without improving underlying discriminability. A vivid narrative can contain correct central facts and inaccurate peripheral details simultaneously.

The field's development is therefore a progression toward more precise conditional claims. The 2017 confidence review, 2025 field-data meta-analysis, 2025 research consensus discussion, and 2026 metacognitive review together show that the meaning of an eyewitness report depends on its task, history, procedure, and measurement.

Psychological research does not provide a way to read a single person's historical experience directly. A psychologist can evaluate relevant conditions and statistical evidence but cannot determine truthfulness or guilt from certainty, vividness, affect, or presentation style alone. Those case-specific conclusions require independent investigation and appropriate legal standards.

How to Preserve the Best Eyewitness Evidence

The guiding practice is to protect the witness's earliest available memory from avoidable contamination. The National Academies recommendations and the American Psychology–Law Society review outline mutually reinforcing safeguards.

Obtain and preserve an early account

Document spontaneous observations in the witness's own words. Preserve open-ended questions and record when any specific prompt was used. Ask a witness to separate direct perception from inference, rumor, later conversation, and media exposure. Record uncertainty without pressuring a confident answer. Where feasible, retain a complete audiovisual interview record.

Limit unnecessary exposure to suspects and suggestions

Avoid showing photographs, public posts, or case narratives to witnesses before a formal identification unless operationally necessary, and document exposures that have already occurred. Separate witnesses when appropriate so that their first descriptions remain as independent as possible. Do not confuse the helpfulness of exchanging information with the evidential independence of what was observed.

Use fair identification tests

The culprit may be absent from the lineup, and the witness should be told that explicitly. Fillers should make the suspect's appearance neither distinctive nor obvious. A blind administrator reduces unintentional cues. The first selection or rejection and the witness's own immediate confidence statement should be recorded before any feedback. A complete video record can preserve the decision process.

Preserve rejection decisions and inconsistencies

A refusal to identify the suspect can carry important information; it should not disappear from the record simply because it does not confirm an investigative theory. Likewise, preserve the sequence of accounts instead of rewriting the record into a single polished narrative. Later changes should be interpreted in relation to intervening exposure and retrieval conditions.

Compare memory with independent evidence

Eyewitness reports are one information source among recordings, physical evidence, documented timelines, corroborating witnesses, and alternative explanations. Scientific safeguards increase the interpretability of memory evidence; they do not guarantee any specific outcome. Legal questions about admissibility and juror judgment belong to the Eyewitness Testimony: How Memory Works in Legal Cases and Eyewitness Evidence and Jurors: Confidence, Credibility, and Error guides.

Common Misconceptions About Eyewitness Memory

“Memory works like a video recording”

Human episodic memory retains information selectively and reconstructs an event when requested. A rich visual recollection is an experience of remembering, not a stored video that can be replayed without change. Some details may remain accurate while other details are forgotten or reconstructed.

“An honest eyewitness must be accurate”

Honesty concerns intention. Accuracy concerns what actually happened. Source confusion, suggestion, and ordinary perceptual limits can produce sincere mistakes without any motive to deceive.

“Eyewitness confidence has no scientific value”

Confidence recorded at a first, properly controlled identification can provide meaningful information about accuracy. It is necessary to examine whether the test was fair, whether memory was contaminated, and when the statement was collected.

“A confident eyewitness cannot be mistaken”

Even an immediate high-confidence decision can be wrong. The observed error rate depends on the experiment or field sample, suspect selection, and test procedure. A confident courtroom assertion after feedback should not be mistaken for an uncontaminated initial measurement.

“Any discrepancy between interviews proves lying”

Different cues, forgetting, suggestion, and genuine later retrieval can all produce changed accounts. An inconsistency requires source analysis and independent evidence, rather than a direct inference about intention.

“A weapon always prevents accurate face recognition”

Weapon focus has a more complex evidence base, including recent preregistered nonreplications. The actual observation and task conditions must be evaluated rather than applying a universal rule based on a weapon's presence.

“Scientists can certify a witness's truthfulness from vividness or emotion”

No reliable psychological shortcut allows a historical claim to be verified merely through a vivid memory, emotional intensity, or consistent confidence. Independent evidence is indispensable to the assessment of a particular case.

Everyday Applications of Eyewitness-Memory Research

The principles extend beyond courts. Two employees may remember the same safety incident differently because they attended to different actions. A driver may become increasingly certain about a signal's color after repeated discussion. A parent may correctly remember a child's words yet mistakenly attribute one detail to the original exchange instead of a later conversation.

In high-stakes organizations, contemporaneous notes, open-ended questioning, and separation of observation from inference help preserve useful information. In everyday disagreements, remembering is better understood as evidence requiring context than as a perfect transcript. The awareness that ordinary human remembering is reconstructive supports careful investigation and more precise communication.

Frequently Asked Questions

What is eyewitness memory in simple terms?

It is a person's memory of something they directly saw or heard during an event. The information might include who was present, what happened, or how events unfolded. Its accuracy depends on the original perception and later retrieval conditions.

Can eyewitness memory be accurate?

Yes. Witnesses can retain valuable details and sometimes make highly informative first identifications under fair procedures. No universal accuracy percentage applies across every event, individual, and type of memory test.

Why can two witnesses remember the same event differently?

They may have occupied different positions, paid attention to different details, perceived events under different conditions, or encountered different information afterward. Distinct accounts do not automatically imply dishonesty.

Is eyewitness testimony the same as eyewitness memory?

Eyewitness memory is the cognitive process. Testimony is a report supplied to investigators or a court, where questioning, legal context, and communication also influence how information is presented.

Does confidence mean a witness is correct?

Not necessarily. Immediate confidence can relate strongly to accuracy under specific controlled conditions, but high-confidence errors occur. Later confidence may be influenced by feedback and repeated exposure.

Are first eyewitness statements more reliable than later ones?

They are especially informative about what was available before many sources of possible contamination. The first report can still contain errors, and later details can sometimes be accurate. The sequence and circumstances of statements should be preserved.

How does the misinformation effect change a memory?

Misleading post-event information can affect later recognition, recall, and reporting. Researchers investigate source confusion, competing memory information, and response processes; the effect does not always mean an earlier trace was literally erased.

Can looking at a suspect's photograph on social media affect identification?

Yes. The photo can create an additional source of face familiarity, complicating whether the witness remembers the person from the event or the later image. The timing and extent of exposure should be documented.

Are cross-race eyewitness identifications always inaccurate?

No. Research has identified average differences in face-recognition performance in some settings, with substantial variation across people and contexts. The group-level pattern cannot decide a specific identification.

Can children and older adults provide useful eyewitness evidence?

Yes. Age-group averages and particular task demands are relevant, but age does not establish whether a specific memory is correct. Clear, appropriate, non-leading interviews matter.

Does stress make eyewitness memory inaccurate?

High stress can impair some forms of eyewitness performance on average, while some emotionally central information remains memorable. The consequences depend on what was experienced and what is later tested.

Are repeated questions harmful to memory?

They can introduce suggestions or encourage narrative rehearsal, particularly when leading information or feedback is repeated. Repeated retrieval may also elicit genuinely remembered details. Documentation helps establish the origin of each statement.

Can a psychologist determine whether one witness is telling the truth?

Psychological expertise can explain known memory vulnerabilities and assess the quality of collection methods. It cannot independently establish a unique event's truth solely through memory presentation, confidence, or demeanor.

What is the best way to protect eyewitness evidence?

Obtain an early account in the witness's words, minimize suggestion and unnecessary exposure, use fair and blind identification procedures, record immediate confidence, retain both selections and rejections, and compare reports with independent evidence.

Related Articles

Additional published EN-24 science articles: Reconstructive Memory: Why Remembering Changes Over Time covers reconstruction and source memory, while Memory Confidence vs. Accuracy: What the Evidence Shows examines confidence calibration beyond the eyewitness context.

References

Bradfield, A. L., Wells, G. L., & Olson, E. A. (2002). The damaging effect of confirming feedback on the relation between eyewitness certainty and identification accuracy. Journal of Applied Psychology, 87(1), 112–120. https://doi.org/10.1037/0021-9010.87.1.112

Deffenbacher, K. A., Bornstein, B. H., Penrod, S. D., & McGorty, E. K. (2004). A meta-analytic review of the effects of high stress on eyewitness memory. Law and Human Behavior, 28, 687–706. https://doi.org/10.1007/s10979-004-0565-x

Fitzgerald, R. J., & Price, H. L. (2015). Eyewitness identification across the life span: A meta-analysis of age differences. Psychological Bulletin, 141(6), 1228–1265. https://doi.org/10.1037/bul0000013

Fitzgerald, R. J., Hyman, I. E., Jr., & Wade, K. A. (2025). Error rates for high confidence eyewitness identifications. Memory, 33(9), 1041–1053. https://doi.org/10.1080/09658211.2025.2546919

Katzman, J., & Kovera, M. B. (2023). Potential causes of racial disparities in wrongful convictions based on mistaken identifications: Own-race bias and differences in evidence-based suspicion. Law and Human Behavior, 47(1), 23–35. https://doi.org/10.1037/lhb0000503

Lindsay, D. S., Wixted, J. T., Fitzgerald, R. J., Hyman, I. E., Jr., Mickes, L., & Wade, K. A. (2025). Eyewitness suspect identification: Six claims regarding the state of the science. Memory, 33(7), 757–764. https://doi.org/10.1080/09658211.2025.2551222

Loftus, E. F., Miller, D. G., & Burns, H. J. (1978). Semantic integration of verbal information into a visual memory. Journal of Experimental Psychology: Human Learning and Memory, 4(1), 19–31. https://pubmed.ncbi.nlm.nih.gov/621467/

Mickes, L., Wilson, B. M., & Wixted, J. T. (2025). The cognitive science of eyewitness memory. Trends in Cognitive Sciences, 29(7), 655–668. https://doi.org/10.1016/j.tics.2025.01.008

National Research Council. (2014). Identifying the culprit: Assessing eyewitness identification. National Academies Press. https://doi.org/10.17226/18891

Otgaar, H., Howe, M. L., Merckelbach, H., & Muris, P. (2018). Who is the better eyewitness? Sometimes adults but at other times children. Current Directions in Psychological Science, 27(5), 378–385. https://doi.org/10.1177/0963721418770998

Schacter, D. L., Guerin, S. A., & St. Jacques, P. L. (2011). Memory distortion: An adaptive perspective. Trends in Cognitive Sciences, 15(10), 467–474. https://doi.org/10.1016/j.tics.2011.08.004

Singh, B., Mellinger, C., Earls, H. A., Tran, J., Bardsley, B., & Correll, J. (2022). Does cross-race contact improve cross-race face perception? A meta-analysis of the cross-race deficit and contact. Personality and Social Psychology Bulletin, 48(6), 865–887. https://doi.org/10.1177/01461672211024463

Wells, G. L., Kovera, M. B., Douglass, A. B., Brewer, N., Meissner, C. A., & Wixted, J. T. (2020). Policy and procedure recommendations for the collection and preservation of eyewitness identification evidence. Law and Human Behavior, 44(1), 3–36. https://doi.org/10.1037/lhb0000359

West, J. T., Mulligan, N. W., & Bornstein, B. H. (2026). Four (and a half) preregistered failures to replicate the weapon focus effect in online samples. Psychology, Public Policy, and Law, 32(1), 48–64. https://doi.org/10.1037/law0000469

Wixted, J. T., & Wells, G. L. (2017). The relationship between eyewitness confidence and identification accuracy: A new synthesis. Psychological Science in the Public Interest, 18(1), 10–65. https://doi.org/10.1177/1529100616686966

Yu, C., & Moore, K. N. (2026). Revisiting the confidence–accuracy relationship in eyewitness identification: A metacognitive perspective. Memory, 34(8), 1008–1029. https://doi.org/10.1080/09658211.2026.2691574

 
 
bottom of page