Snap Shot
- The question: How reliable is what an eyewitness remembers and reports about a crime, including descriptions and identifications of a person, and what determines that reliability?
- What the research shows: Accuracy depends on conditions at the event (estimator variables) and on how the evidence is collected (system variables). Mistaken identification contributed to a large share of DNA exonerations, but memory tested once, early, and with sound procedures is often informative: high-confidence initial identifications from fair, double-blind lineups are usually accurate, and lineup rejections carry information too.
- What it does not show: Group research cannot show that a particular witness’s description or identification is accurate or mistaken. Exoneration data show that misidentification can lead to wrongful conviction, not how often identifications in general are wrong, and confidence expressed at trial, after repeated testing, is a poor guide to accuracy.
- The law: Due process screens police-arranged identification procedures that are unnecessarily suggestive and unreliable under the totality of the circumstances (Stovall, Simmons, Biggers, Manson, Perry), as Mil. R. Evid. 321 does in courts-martial and a Dessureault hearing does in Arizona. Expert testimony is governed by Rule 702 and Houser, and Arizona recognized its value in Chapple.
- Issues for each side: Prosecutors can show that a careful first identification, recorded with an immediate confidence statement, is probative, without offering courtroom certainty as proof. Defense counsel can show how viewing conditions and procedures bear on reliability, while recognizing that strong conditions and sound procedures support an identification and that an expert cannot say a particular witness is wrong.
Eyewitness memory refers to what a person who saw an event, usually a crime, retains of it and later recounts to police, lawyers, and the court. It includes the account of what happened, the description of the people involved, and any identification of a suspect. Ordinary experience treats such memory as a recording that testimony plays back. Memory research describes something different: what is retrieved is reconstructed from what was attended to and encoded, what was learned afterward, and the cues present when the person is asked to remember (Lane & Houston, 2021, pp. 15–18). That does not make eyewitness memory unreliable. It makes its reliability depend on identifiable conditions, some fixed by the event and some within the control of the people who collect the evidence.
Eyewitness evidence reaches court in nearly every kind of criminal case, and it is persuasive. Errors documented in DNA exonerations have made eyewitness identification among the most studied topics in forensic psychology, and research has changed police practice. The same research has also shown when eyewitness memory is reliable, a point that receives less attention.
This paper is the starting point for the Eyewitness Identification papers in this hub. It covers estimator and system variables, the exoneration data, the strengths and limits of witness accounts, testing memory once, the 2020 scientific recommendations, and the governing federal, military, and Arizona law. Confidence, lineup construction, show-ups, the cross-race effect, and weapon focus have their own papers.
Estimator and System Variables
Wells (1978) proposed dividing the factors that affect eyewitness accuracy into two groups. Estimator variables are features of the event and the witness that the justice system cannot control, such as lighting, distance, viewing time, stress, a weapon, intoxication, age, race of the witness and the person observed, and delay; after the event, their influence can be assessed but not changed. System variables are features of evidence collection that the system does control, such as how an interview is conducted, how a lineup is built and administered, what the witness is told, how confidence is recorded, and whether the witness is tested more than once.
The distinction has practical consequences for both sides. Estimator variables bear on how much weight a particular identification can carry, but nothing done later can improve the original viewing. System variables are where errors can be prevented, and they are also where sound practice produces evidence that deserves confidence. High stress was associated with lower accuracy in both identification and recall in a meta-analysis of 27 identification and 36 recall tests (Deffenbacher et al., 2004, as cited in Lane & Houston, 2021, p. 104); the weapon-focus effect is robust across studies (Lane & Houston, 2021, pp. 109–111); and people recognize faces of their own race more accurately than faces of other races, with other-race faces producing more false identifications, a finding pooled across 39 articles and nearly 5,000 participants (Meissner & Brigham, 2001). Some estimator variables matter less than intuition suggests: in laboratory and field studies, alcohol at the levels tested impaired the completeness and accuracy of recall but has not been shown to impair lineup identification (Lane & Houston, 2021, pp. 33–35).
Worse conditions of either kind tend to raise mistaken identifications, not only lower correct ones. In a study that degraded either the viewing conditions or the fairness of the lineup, mistaken identifications of innocent lineup members increased in both cases (Smith et al., 2019).
What the Exoneration Data Show
A DNA exoneration is a case in which a conviction was vacated, and the person cleared, after post-conviction DNA testing. The Innocence Project counted 375 DNA exonerations in the United States from 1989 through 2020 and reported that 69% involved eyewitness misidentification as a contributing factor, and that 42% of those misidentifications were cross-racial (Innocence Project, n.d.; see also Lane & Houston, 2021, p. 3). “Contributing factor” means the misidentification was one of the circumstances that led to the conviction, not necessarily the only one; most cases had more than one, and the categories overlap.
The National Registry of Exonerations counts both DNA and non-DNA cases in which a convicted person was relieved of all legal consequences of the conviction based on new evidence of innocence (Gross & Shaffer, 2012, p. 7). Its first report, covering 873 exonerations from 1989 to early 2012, found mistaken witness identification in 43% of all exonerations and in 80% of adult sexual assault exonerations (Gross & Shaffer, 2012, p. 40). By the end of 2025 the Registry had recorded 3,767 exonerations, and mistaken identification was a factor in 22 of the 97 exonerations recorded for 2025 (23%) (National Registry of Exonerations, 2026). The higher share among DNA exonerations likely reflects the kinds of cases in which DNA evidence is available, chiefly sexual assaults in which the identity of a stranger was disputed.
These data establish that mistaken identifications occur, that sincere and confident witnesses can be wrong, and that misidentification can lead to conviction of an innocent person. They do not show how often identifications in general are mistaken, because exonerations are a small and selected group of cases in which later evidence was available. Researchers also disagree about what the exoneration cases reveal. Some argue that in many of them the witness’s first identification was hesitant or uncertain and became confident only through later procedures, so that the error lay less in the witness’s memory than in the system’s reliance on contaminated later tests (Wixted et al., 2018; Wixted & Mickes, 2022). Others caution that the record of initial confidence in those cases is incomplete and that the reliability of witness memory should not be overstated (Berkowitz et al., 2022; Wade et al., 2018).
The Strengths and Limits of Witness Accounts
Witness accounts are often accurate in their central features. In a field study of witnesses to a fatal shooting, those interviewed four to five months later gave accounts that remained largely accurate and resisted misleading questions (Yuille & Cutshall, 1986). That is a single case, but it is consistent with broader findings that memory for the gist of an event and its central details is generally more durable than memory for peripheral details, while emotion can enhance some central details and impair others, and “central” has been defined inconsistently across studies (Lane & Houston, 2021, pp. 105–108). People recognize familiar faces far more accurately than unfamiliar ones, even from poor-quality video (Lane & Houston, 2021, pp. 52–53). Details first reported in a later interview are often accurate, and treating every new detail as a sign of unreliability is not supported by the research (Lane & Houston, 2021, p. 129).
Accounts are also vulnerable. Information met after the event, whether from a co-witness, media coverage, or how a question is phrased, can be incorporated into memory and later attributed to the event itself. In a classic demonstration, witnesses asked how fast cars were going when they “smashed” into each other, rather than “hit,” gave higher speed estimates and were more likely to report broken glass that had not been shown (Loftus & Palmer, 1974). For that reason an account recorded early, before the witness has encountered such information, generally carries the most weight, and a witness who repeats the same story in later interviews has not thereby shown that the story is accurate.
Person Descriptions
Descriptions of an unfamiliar person tend to be brief and general. They usually include sex, age, height, and build, rarely describe inner facial features, and often fit many people in the area (Lane & Houston, 2021, pp. 28–29). Witnesses to staged crimes in research settings give more descriptors than real witnesses do in some comparisons; in one, the averages were about seven per witness in staged events and about four in descriptions of real offenders published in newspapers (Lindsay et al., 1994, as cited in Lane & Houston, 2021, pp. 28–29), while an archival study of 100 police reports of violent crimes found an average of about seven (Kuehn, 1974, as cited in Lane & Houston, 2021, p. 28). Samples and counting rules differ, so neither figure is a norm. How detailed a description is does not reliably predict whether the same witness will later identify the right person (Lane & Houston, 2021, pp. 39–42), which means a sparse description does not show a later identification is wrong, and a rich one does not show it is right.
Interview method affects how much a description contains. In research by Satin and Fisher (2019), college students who watched a simulated robbery and were questioned with the Cognitive Interview, a structured method built on memory research, gave almost three times as many descriptors of the robber as students given a standard police-style interview, at comparable accuracy. When those descriptions were handed to mock investigators (college students and, in one experiment, police officers) who searched a set of suspect photographs, the Cognitive Interview descriptions raised the rate of finding the perpetrator by almost 30% for both groups. These are laboratory figures, not measures of real investigations. The form of a question also matters. Witnesses sometimes volunteer precise details that are wrong while withholding broader ones they could report accurately, and asking specifically for coarser answers (for example, “dark” rather than a specific color) recovered accurate information in one series of studies (Brewer et al., 2018).
Culture and Language
Most eyewitness research has used participants from a narrow range of cultures interviewed in their first language. A review by an international group of researchers and practitioners described cultural differences in memory and communication that bear on witness interviews, identified common mismatches between what interviewers and interviewees expect, and concluded that research has largely overlooked cultural context and that practitioners lack evidence-based guidance (Hope et al., 2022). A sparse or differently organized account from a witness of another cultural or language background may therefore reflect communication conventions or the interview context rather than poor memory or evasiveness, and the research does not yet support confident predictions in either direction.
Testing Memory Once: The First Uncontaminated Test
An identification procedure is a test of memory, and like other memory tests it changes the memory it tests. A witness who views a suspect in a lineup, a show-up, a photograph, or a courtroom has seen that face again, and the later sense of familiarity may reflect the earlier test rather than the crime. Researchers who have disagreed on other questions agree that memory for a suspect should, where possible, be tested only once, early, and under sound conditions, and that later tests with the same suspect add risk without adding much information (Wixted et al., 2021).
The relationship between confidence and accuracy illustrates why the first test matters. Under what Wixted and Wells (2017) called “pristine” conditions, including only one suspect per lineup, a fair lineup in which the suspect does not stand out, double-blind administration, an instruction that the culprit may not be present, and a confidence statement taken immediately, high-confidence initial identifications are generally highly accurate and low-confidence identifications are much less so. Confidence expressed later is far less informative. In a classic experiment, witnesses told “Good, you identified the suspect” became more confident and also reported having had a better view and paying more attention than witnesses who received no feedback (Wells & Bradfield, 1998). By trial, certainty may reflect feedback, repetition, and preparation rather than the original memory. Wixted and Mickes (2022) argue that eyewitness memory is reliable on the initial, uncontaminated test and that errors arise largely when the system relies on later, contaminated tests such as an in-court identification; the other side of that exchange cautions that high-confidence errors occur and that courts should not over-rely on confidence (Berkowitz et al., 2022). Reviews also note that in some data sets more than one in five of the highest-confidence identifications was wrong, and that the confidence of witnesses who reject a lineup tells little about accuracy (as summarized in Lane & Houston, 2021, pp. 89–90; the underlying sample sizes and conditions are not given here).
A 2025 discussion paper written by six researchers whose work has emphasized either the vulnerabilities or the reliability of eyewitness memory set out points on which they described an emerging consensus (Lindsay et al., 2025). Memory is least contaminated on a first test conducted soon after the event under recommended procedures. A suspect should be selected on independent evidence rather than facial resemblance alone, because a person found through a facial-recognition search or a publicized image may closely resemble the witness’s memory without being the culprit and may be identified with high confidence. On a first test under good conditions, high-confidence identifications are more likely to be accurate than low-confidence ones, and a low-confidence identification from a fair lineup is not strong evidence of guilt. Lineup rejections and filler identifications on a first test are evidence pointing toward innocence rather than an absence of information.
The 2020 Scientific Recommendations
In 2020 a committee of the American Psychology-Law Society, a division of the American Psychological Association, updated its 1998 recommendations on identification procedures (Wells et al., 2020). Four recommendations carried over from 1998: choose fillers so the suspect does not stand out, instruct the witness that the culprit may or may not be present, use double-blind administration, and take a confidence statement in the witness’s own words at the time of the identification. Five were new: interview the witness before the identification procedure, require evidence-based grounds to suspect the person before including him or her in a lineup, video-record the entire procedure, avoid repeated identification procedures with the same witness and suspect, and avoid show-ups where possible and improve them when they are necessary. The National Research Council (2014) had earlier reached similar conclusions about blind administration, instructions, confidence statements, and recording.
The recommendations aim both to reduce mistaken identifications and to preserve the value of correct ones by creating a record from which an identification’s strength can later be judged, which serves both sides. Lane and Houston (2021, pp. 12–13) compare eyewitness memory to trace evidence that can be contaminated and requires care in collection.
Limits of the Evidence
Most eyewitness research uses staged events or videos, often student participants, short delays, and no real consequences. Laboratory studies isolate single variables, but real cases combine many at levels that may fall outside those tested; for ethical reasons, stress and intoxication are studied below the levels in many real crimes. Field studies of real witnesses have the opposite problem: researchers usually do not know who the culprit was, so accuracy must be inferred. Lane and Houston (2021, pp. 9–10) point out that a study’s relevance depends on whether it tested the mechanism at issue, not on whether it looked like the case at hand, and that field studies are not automatically more generalizable than laboratory ones.
Many estimator-variable effects are modest and vary across studies, and the exoneration data describe a selected group of cases. Research reports averages for groups of witnesses under specified conditions and cannot establish whether a particular witness’s description or identification is accurate. An expert can explain which conditions research associates with higher or lower accuracy and which were present in a case, but not whether a particular witness is right or wrong.
Questions to Ask About This Research
When exoneration figures are cited, counsel should ask what was counted: the Innocence Project’s 69% is the share of 375 DNA exonerations through 2020 in which misidentification was one contributing factor, not an error rate for identifications, and the Registry’s figures cover both DNA and non-DNA cases and vary by year and crime type. When Satin and Fisher (2019) is cited for the value of the Cognitive Interview, counsel should ask what the figures count and whether the conclusion follows for this case: the almost-30% figure describes how often mock investigators picked the perpetrator from photographs in a laboratory study, not the outcome of real investigations, and counsel can ask for the underlying percentages. When Wixted and Wells (2017) or Lindsay et al. (2025) is cited to support a confident identification, counsel should ask whether the conditions those authors specify were present, including independent grounds for suspicion, a fair and double-blind procedure, and a confidence statement recorded before any feedback. When Meissner and Brigham (2001) or another meta-analysis is cited for an estimator variable, counsel should ask how much evidence it rests on and how large the effect was under conditions like those in the case, recognizing that a group-level effect does not show that a particular identification is mistaken.
Counsel can also ask whether a study measured identification or recall, since some variables affect one and not the other, and which of the witness’s tests the findings describe. These questions are explained in “Reading Research Claims: A Guide for Attorneys” on this site.
Eyewitness Memory and the Law
Federal and Military Standards
The Due Process Clause limits the use of identifications obtained through suggestive procedures. United States v. Wade (1967) recognized a right to counsel at post-indictment lineups, and Stovall v. Denno (1967) held that suggestive confrontations are judged under the totality of the circumstances. Simmons v. United States (1968) asked whether a photographic identification was so impermissibly suggestive as to create a very substantial likelihood of irreparable misidentification. Neil v. Biggers (1972) and Manson v. Brathwaite (1977) made reliability the central question and listed five factors for weighing it against suggestiveness: the witness’s opportunity to view, degree of attention, accuracy of the prior description, level of certainty at the confrontation, and the time between the crime and the confrontation. Research bears on several factors: certainty and self-reported view and attention can be inflated by feedback after the identification (Wells & Bradfield, 1998), which is why recording them at the first identification matters (Wixted & Wells, 2017). In Perry v. New Hampshire (2012), the Court held that this due process screen applies only when law enforcement arranged the suggestive circumstances, and it identified cross-examination, jury instructions, and expert testimony as the safeguards for other identifications.
In courts-martial, Mil. R. Evid. 321 generally admits relevant out-of-court and in-court identifications but excludes those resulting from an unlawful identification process or required to be excluded under due process; the rule treats a process as unlawful if it is so suggestive as to create a substantial likelihood of misidentification, or if a lineup violates the accused’s right to counsel (Mil. R. Evid. 321(b)–(c)). When an objection raises unreliability, the prosecution must show by a preponderance of the evidence that the identification was reliable under the circumstances, and a later identification that follows an unreliable one is admissible only if the prosecution shows by clear and convincing evidence that it was not the result of the inadmissible one (Mil. R. Evid. 321(d)(6)(B)). In United States v. Criswell (2018), the Court of Appeals for the Armed Forces applied the two-step analysis (whether the procedure was unnecessarily suggestive and, if so, whether it created a substantial likelihood of misidentification under the Biggers factors), noting that the inquiry is not limited to those factors and that a witness may hold a sincere belief that is mistaken. On its facts, a divided court (3–2) held that the military judge did not abuse his discretion in admitting an in-court identification that followed a single-photograph display he assumed was unnecessarily suggestive, given the witness’s close and repeated view, a description the judge found accurate, and a quick photo identification less than 24 hours after the offense. Expert testimony on eyewitness memory is governed by Mil. R. Evid. 702 and United States v. Houser (1993); in United States v. Garcia (1996), the Air Force Court of Military Review found that excluding a defense eyewitness expert’s testimony on the confidence–accuracy relationship and on memory transference was an abuse of discretion but not prejudicial, and the Court of Appeals for the Armed Forces affirmed, holding that any error was harmless beyond a reasonable doubt and did not deny the accused his right to present a defense (44 M.J. at 28, 31, 33). The testimony reflected 1990s research, before the work on initial confidence described above. The December 1, 2023 amendment to Fed. R. Evid. 702, which requires the proponent to show that it is more likely than not that the admissibility requirements are met, became part of Mil. R. Evid. 702 through Mil. R. Evid. 1102 in 2025 (Joint Service Committee on Military Justice, 2025); practitioners should confirm the current text. Under either rule, an expert may explain general principles but may not opine that a particular witness is accurate or mistaken.
Arizona
When a defendant challenges a pretrial identification, Arizona courts hold a hearing outside the jury’s presence under State v. Dessureault (1969). If the pretrial procedure was unduly suggestive, the State must show by clear and convincing evidence that a proposed in-court identification is not tainted by it, and the defendant may request a cautionary instruction. Arizona’s standard identification instruction, RAJI (Criminal) Standard Criminal 45, tells jurors that the State must prove beyond a reasonable doubt that the in-court identification is reliable and lists factors drawn from Biggers and Manson, including the witness’s certainty at the time of the pretrial identification. State v. Chapple (1983) is Arizona’s leading case on expert testimony about eyewitness memory: the Arizona Supreme Court held that excluding such testimony was an abuse of discretion in the circumstances of that case, where identification was the central disputed issue; admission otherwise remains within the trial court’s discretion. In State v. Nordstrom (2001), the court approved a ruling that allowed an eyewitness expert to explain the factors affecting reliability while barring opinions on the accuracy of a particular eyewitness. Arizona Rule of Evidence 702 adopted the Daubert standard effective January 1, 2012 and was amended effective January 1, 2024 to track the 2023 federal amendment (Ariz. Sup. Ct. Order R-23-0004). State v. Salazar-Mercado (2014) permits general, educational expert testimony that satisfies Rules 702 and 403, and State v. Lindsey (1986) bars expert opinions on whether a particular witness is telling the truth.
Special Issues in Legal Practice
The principles described above apply with equal force to every party in a legal proceeding. The science operates the same way regardless of who is asking the questions, and the research summarized here does not by itself favor either side. What differs between the parties is not the science but the vantage point — the particular places where these issues tend to surface, and the particular errors each party is best positioned to guard against. The two summaries below identify those recurring issues. They are offered for orientation, not as litigation strategy; in every case the controlling question is what the evidence in that case will, and will not, support, and the expert’s role is to inform the trier of fact rather than to advocate for either side.
Special Issues for Prosecutors
The research supports several propositions prosecutors can legitimately present. A high-confidence first identification made soon after the event under pristine conditions is generally accurate. Witnesses often remember the gist and central details of an event well, recognize familiar people accurately, and add accurate details in later interviews, and a brief description does not show that a later identification is wrong. Prosecutors are well positioned to secure these strengths by ensuring that suspects are developed on independent evidence and that procedures follow the 2020 recommendations, including a recorded confidence statement and video.
The research also counsels caution. Confidence expressed at trial, after feedback, repeated viewing of the defendant, and preparation, is a poor indicator of accuracy, and testimony that treats courtroom certainty as proof of identity goes beyond what the research supports. An in-court identification by a witness who previously failed to identify the defendant, or who identified a filler, adds little independent information, and on a first test, rejections and filler identifications are evidence pointing toward innocence (Lindsay et al., 2025). Suspects developed through facial resemblance alone, such as a facial-recognition search or a publicized photograph, carry a specific risk of confident misidentification and call for independent corroboration. Unfavorable estimator variables may reduce the weight of an identification even when procedures were sound.
Special Issues for Defense
The research supports several propositions defense counsel can legitimately present. Estimator variables such as brief or distant viewing, poor lighting, high stress, weapon focus, and cross-race identification are associated with lower accuracy, and degraded viewing or unfair procedures increase mistaken identifications of innocent people. Suggestive procedures, a non-blind administrator, confirming feedback, and repeated identification procedures can inflate confidence and contaminate later identifications, which is why the first test and the confidence recorded then carry the most weight. Sincere, confident witnesses can be wrong. Expert testimony explaining these principles, matched to the conditions documented in the case, is a legitimate way to help the factfinder evaluate an identification.
The research also counsels caution about its limits. The same science that identifies risk factors shows that memory tested early and properly is often reliable, so a high-confidence first identification under sound procedures is evidence the research treats as probative. Estimator-variable effects are average tendencies, and exoneration data are not an error rate. A sparse description, a delayed report of a detail, or a minor inconsistency does not by itself show that an identification is mistaken. An expert can describe conditions associated with error; the expert cannot tell the factfinder that a particular witness is mistaken, and testimony that implies otherwise invites exclusion or limitation under Rule 702, Houser, or Nordstrom.
This paper is for educational purposes only and is not legal advice. The author is a psychologist, not an attorney; summaries of law are general and must be verified against current authority or with a licensed attorney.
References
Berkowitz, S. R., Garrett, B. L., Fenn, K. M., & Loftus, E. F. (2022). Convicting with confidence? Why we should not over-rely on eyewitness confidence. Memory, 30(1), 10–15. https://doi.org/10.1080/09658211.2020.1849308
Brewer, N., Vagadia, A. N., Hope, L., & Gabbert, F. (2018). Interviewing witnesses: Eliciting coarse-grain information. Law and Human Behavior, 42(5), 458–471. https://doi.org/10.1037/lhb0000294
Gross, S. R., & Shaffer, M. (2012). Exonerations in the United States, 1989–2012: Report by the National Registry of Exonerations. National Registry of Exonerations. https://exonerationregistry.org/sites/exonerationregistry.org/files/documents/exonerations_us_1989_2012_full_report.pdf
Hope, L., Anakwah, N., Antfolk, J., Brubacher, S. P., Flowe, H., Gabbert, F., Giebels, E., Kanja, W., Korkman, J., Kyo, A., Naka, M., Otgaar, H., Powell, M. B., Selim, H., Skrifvars, J., Sorkpah, I. K., Sowatey, E. A., Steele, L. C., Stevens, L., . . . Anonymous. (2022). Urgent issues and prospects at the intersection of culture, memory, and witness interviews: Exploring the challenges for research and practice. Legal and Criminological Psychology, 27(1), 1–31. https://doi.org/10.1111/lcrp.12202
Innocence Project. (n.d.). DNA exonerations in the United States (1989–2020). Retrieved November 2, 2026, from https://innocenceproject.org/dna-exonerations-in-the-united-states/
Joint Service Committee on Military Justice. (2025). Amendments to the Military Rules of Evidence since 1984 (G. E. Maggs, Ed.; 2025 ed.). https://jsc.defense.gov/Portals/99/MRE%20Amendments%202025.pdf
Lane, S. M., & Houston, K. A. (2021). Understanding eyewitness memory: Theory and applications. New York University Press.
Lindsay, D. S., Wixted, J. T., Fitzgerald, R. J., Hyman, I. E., 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., & Palmer, J. C. (1974). Reconstruction of automobile destruction: An example of the interaction between language and memory. Journal of Verbal Learning and Verbal Behavior, 13(5), 585–589. https://doi.org/10.1016/S0022-5371(74)80011-3
Meissner, C. A., & Brigham, J. C. (2001). Thirty years of investigating the own-race bias in memory for faces: A meta-analytic review. Psychology, Public Policy, and Law, 7(1), 3–35. https://doi.org/10.1037/1076-8971.7.1.3
National Registry of Exonerations. (2026). 2025 annual report. https://exonerationregistry.org/sites/exonerationregistry.org/files/documents/2025Exonerations.pdf
National Research Council. (2014). Identifying the culprit: Assessing eyewitness identification. The National Academies Press. https://doi.org/10.17226/18891
Satin, G. E., & Fisher, R. P. (2019). Investigative utility of the Cognitive Interview: Describing and finding perpetrators. Law and Human Behavior, 43(5), 491–506. https://doi.org/10.1037/lhb0000326
Smith, A. M., Wilford, M. M., Quigley-McBride, A., & Wells, G. L. (2019). Mistaken eyewitness identification rates increase when either witnessing or testing conditions get worse. Law and Human Behavior, 43(4), 358–368. https://doi.org/10.1037/lhb0000334
Wade, K. A., Nash, R. A., & Lindsay, D. S. (2018). Reasons to doubt the reliability of eyewitness memory: Commentary on Wixted, Mickes, and Fisher (2018). Perspectives on Psychological Science, 13(3), 339–342. https://doi.org/10.1177/1745691618758261
Wells, G. L. (1978). Applied eyewitness-testimony research: System variables and estimator variables. Journal of Personality and Social Psychology, 36(12), 1546–1557. https://doi.org/10.1037/0022-3514.36.12.1546
Wells, G. L., & Bradfield, A. L. (1998). “Good, you identified the suspect”: Feedback to eyewitnesses distorts their reports of the witnessing experience. Journal of Applied Psychology, 83(3), 360–376. https://doi.org/10.1037/0021-9010.83.3.360
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
Wixted, J. T., & Mickes, L. (2022). Eyewitness memory is reliable, but the criminal justice system is not. Memory, 30(1), 67–72. https://doi.org/10.1080/09658211.2021.1974485
Wixted, J. T., Mickes, L., & Fisher, R. P. (2018). Rethinking the reliability of eyewitness memory. Perspectives on Psychological Science, 13(3), 324–335. https://doi.org/10.1177/1745691617734878
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
Wixted, J. T., Wells, G. L., Loftus, E. F., & Garrett, B. L. (2021). Test a witness’s memory of a suspect only once. Psychological Science in the Public Interest, 22(1 Suppl.), 1S–18S. https://doi.org/10.1177/15291006211026259
Yuille, J. C., & Cutshall, J. L. (1986). A case study of eyewitness memory of a crime. Journal of Applied Psychology, 71(2), 291–301. https://doi.org/10.1037/0021-9010.71.2.291
Legal Authorities
Manson v. Brathwaite, 432 U.S. 98 (1977).
Neil v. Biggers, 409 U.S. 188 (1972).
Perry v. New Hampshire, 565 U.S. 228 (2012).
Simmons v. United States, 390 U.S. 377 (1968).
State v. Chapple, 135 Ariz. 281, 660 P.2d 1208 (1983).
State v. Dessureault, 104 Ariz. 380, 453 P.2d 951 (1969).
State v. Lindsey, 149 Ariz. 472, 720 P.2d 73 (1986).
State v. Nordstrom, 200 Ariz. 229, 25 P.3d 717 (2001).
State v. Salazar-Mercado, 234 Ariz. 590, 325 P.3d 996 (2014).
Stovall v. Denno, 388 U.S. 293 (1967).
United States v. Criswell, 78 M.J. 136 (C.A.A.F. 2018).
United States v. Garcia, 44 M.J. 27 (C.A.A.F. 1996).
United States v. Houser, 36 M.J. 392 (C.M.A. 1993).
United States v. Wade, 388 U.S. 218 (1967).
U.S. Const. amend. V.
U.S. Const. amend. VI.
U.S. Const. amend. XIV.
Ariz. Sup. Ct. Order R-23-0004 (2023) (amending Ariz. R. Evid. 702, effective Jan. 1, 2024).
Fed. R. Evid. 702.
Ariz. R. Evid. 702.
Rev. Ariz. Jury Instr. (Crim.) Standard Criminal 45 (Identification) (6th ed. 2022).
Mil. R. Evid. 321.
Mil. R. Evid. 702.
Mil. R. Evid. 1102.
Related topic papers: Confidence and Accuracy · Lineup Construction · Show-Ups · The Cross-Race Effect · Memory for Violence and Weapon Focus · Improving Eyewitness Memory
Related topics from Arizona Forensic Services: Courts-Martial