Ronald Cotton and Jennifer Thompson: Eyewitness Misidentification, Memory, and Reform

4 days ago

19 min read

Author: Ukrainian Psychological Hub · Published: October 11, 2026 · Editorial Policy

The case of Ronald Cotton and Jennifer Thompson is among the clearest documented examples of how a sincere eyewitness identification can become a confident mistake. Thompson survived a sexual assault in Burlington, North Carolina, in July 1984. She identified Cotton as her attacker, testified against him, and believed she had correctly recognized the person responsible. Cotton was innocent. DNA testing in 1995 excluded him and implicated another man, Bobbie Poole. After more than ten years in prison, Cotton was exonerated. The case is recorded in the 1987 North Carolina Supreme Court opinion, the National Research Council's scientific review, and the Innocence Project's case record.

This case matters because it makes several processes visible in the same legal history: an uncertain first choice, repeated identification procedures, confirming comments from investigators, courtroom certainty, exclusion of alternate-perpetrator evidence, and finally DNA evidence capable of testing the central identity claim independently. Each process requires its own analysis. The criminal case establishes what happened in court; psychological science explains why certain identification procedures can increase error or distort confidence; later reforms show how institutions can protect the evidentiary value of memory.

A central finding follows: the accuracy of an eyewitness identification depends on the witness's opportunity to perceive, the quality of the identification procedure, and the history of information received afterward. Confidence is evidence about accuracy only under particular conditions. The scientific question concerns the reliability of a memory-based identification; the legal question concerns how a jurisdiction decides admissibility, weight, and guilt.

Who were Ronald Cotton and Jennifer Thompson?

Ronald Junior Cotton was a young man living in North Carolina when he became the suspect in a 1984 sexual-assault investigation. Jennifer Thompson, sometimes identified in later accounts as Jennifer Thompson-Cannino, was a college student who survived an assault in her Burlington apartment. Her testimony became an important part of the prosecution. Their names later became closely associated with eyewitness-reform education after forensic DNA testing established Cotton's innocence and the two participated publicly in conversations about memory, wrongful conviction, responsibility, and reconciliation. These biographical and procedural facts are documented in the Innocence Project case history and in National Research Council Box 1-1.

Thompson experienced a genuine violent crime. Her identification of the perpetrator was mistaken. Those two propositions coexist. The failure of an identification does not erase the assault or make the survivor dishonest. Equally, a witness's sincerity does not establish that the accused is the perpetrator. Distinguishing credibility, honesty, memory accuracy, and legal proof is fundamental to the interpretation of this case.

The case also illustrates why wrongful convictions have effects beyond the person imprisoned. The survivor received an incorrect answer about the identity of the attacker, the innocent defendant lost years of liberty, and the actual perpetrator remained outside the reach of this conviction. The record supports these consequences without requiring an invented psychological diagnosis for any participant.

The events of July 29, 1984

The North Carolina Supreme Court's account in State v. Cotton (1987) states that Thompson was attacked in her apartment in Burlington around 3 a.m. on July 29, 1984. The assailant threatened her with a knife, sexually assaulted her, and she escaped. Other similar break-ins and sexual assaults occurred nearby within hours. Those similar incidents became central to the defense's effort to introduce evidence that someone else had committed the crime.

After the assault, Thompson described the attacker and helped produce a composite sketch. Investigators subsequently considered Cotton a suspect. The National Research Council case reconstruction reports that a photograph-based identification procedure presented six photographs and that Thompson had difficulty choosing, initially considering two images before indicating Cotton. That initial uncertainty is historically significant because later testimony expressed much greater certainty.

The next key procedure was a live lineup. Cotton was the only individual who appeared in both the photographic display and the later live lineup, according to the National Research Council. The report also recounts Thompson's hesitation at the live lineup and the subsequent confirming statements by investigators. The case should therefore be evaluated as a sequence of memory tests, rather than as one isolated courtroom identification.

Case chronology: investigation, appeals, retrial, and DNA testing

In 1984, Thompson's assault was investigated, Cotton was identified, and criminal charges followed. The first prosecution produced convictions and a prison sentence. The precise counts and procedural posture should be read in the court record rather than compressed into a generic statement that a mistaken witness alone caused the conviction.

On January 6, 1987, the Supreme Court of North Carolina ordered a new trial on the rape and burglary charges because evidence pointing toward another perpetrator had been wrongly excluded. The court separately addressed the procedural status of a sexual-offense count. The issue resolved on appeal was an evidentiary error involving third-party culpability, not a judicial finding that psychological research had invalidated all eyewitness identification. The 1987 opinion is the primary source for that distinction.

Cotton was convicted again after retrial. The North Carolina Court of Appeals in State v. Cotton (1990) considered further appellate issues and found no reversible error. This procedural history is important because the 1987 new-trial ruling, the later conviction, and the 1995 DNA exoneration are separate judicial and evidentiary events. Treating the initial appeal as the exoneration would misstate the record.

In 1995, physical evidence underwent DNA testing. According to the Innocence Project's documented exoneration profile, the results excluded Cotton and implicated Bobbie Poole; Cotton was exonerated on June 30, 1995. He had spent more than ten years incarcerated. The DNA results changed the assessment of identity because biological evidence could test the claim independently of eyewitness recollection.

What the 1987 Supreme Court decision actually established

The 1987 court reviewed excluded evidence of other strikingly similar crimes committed in the same geographic area and time frame. One witness to another attack had identified a person other than Cotton when presented with the same lineup. The defense wanted the jury to consider this information as evidence that someone else could have committed the assault on Thompson. The trial court had excluded it.

The North Carolina Supreme Court held that the evidence was relevant to the identity issue and that the defendant was entitled to a new trial on specified charges. Its analysis concerned the admissibility of evidence implicating a third party and the application of North Carolina evidence rules, including relevance principles. It did not decide that every feature of the lineup was constitutionally invalid, and it did not have the later DNA results before it. These details can be checked directly in State v. Cotton, 318 N.C. 663 (1987).

This distinction matters in legal psychology: a psychological explanation of misidentification, an appellate finding of evidentiary error, and a subsequent DNA-based exoneration answer different questions. None should be substituted for another.

What happened during the eyewitness identifications?

The National Research Council describes a six-photo display in which Thompson was initially uncertain, then selected Cotton. Investigators subsequently arranged a live lineup in which only Cotton was repeated from the earlier photographic identification. Thompson again experienced difficulty deciding. Her identification was followed by comments that confirmed her choice, including the information that she had selected the same person previously. The report describes a later courtroom identification delivered with strong certainty. See Identifying the Culprit, Box 1-1.

Repeated exposure is important because seeing a person's face in an investigative display can increase familiarity. At a later procedure, the witness may respond partly to the earlier test experience rather than solely to the original event. This is one reason modern recommendations emphasize preservation of the first identification, the first confidence statement, and the complete sequence of presentations. It does not follow that every repeated identification is false; its evidentiary meaning depends on the exposure history and test conditions.

The National Research Council also recounts affirming remarks by officers after Thompson's choices. Those remarks are pertinent to the scientific literature on post-identification feedback. An investigator may intend to be supportive while unintentionally modifying how a witness later evaluates the quality of the original memory. The concern involves the integrity of the evidence, and does not require a claim that the investigator deliberately sought an incorrect result.

Eyewitness memory is reconstructive

Recognition depends on encoding, retention, and retrieval. At the event, the witness perceives a face under a particular combination of illumination, distance, attention, duration, threat, and competing demands. Later, the witness must compare a remembered representation with people or photographs presented by investigators. At still later stages, the witness may retrieve both event-related information and details encountered during the investigation.

A lineup is therefore a memory test with procedural properties. Its outcome can depend on whether the actual perpetrator is present, the similarity and selection of fillers, the administrator's knowledge, the instructions supplied, and the documentation of confidence. These factors are examined extensively in the National Research Council's 2014 consensus review and the 2020 American Psychology-Law Society scientific review.

Reconstruction does not mean that memories are fictitious. A witness may accurately remember the assault, its location, sensory features, and actions while making an error in recognition of an unfamiliar person. The accuracy of one remembered component cannot be mechanically transferred to all other components. The relevant analysis identifies the specific proposition being tested: was this particular person the assailant?

The first identification versus later courtroom certainty

The most consequential distinction in this case is between the earliest recorded identification conditions and the certainty communicated in court. Initial hesitation can be highly informative. Once a witness has repeatedly seen the suspect, received confirming comments, and prepared for trial, a later confidence judgment becomes difficult to interpret as a measure of the original face memory alone.

Research has refined the relationship between confidence and accuracy. Wixted and Wells (2017) reviewed laboratory and police field evidence and found that high initial confidence can be strongly associated with accurate identification when the first test uses a fair lineup, neutral instructions, no administrator influence, and an immediate confidence statement. When those conditions are compromised, even high confidence may be less informative. The practical lesson is to record precisely when, how, and under what conditions confidence was expressed.

A court may observe a witness speaking confidently many months after the event. That behavior is directly observable; the witness's original memory accuracy is not. Jurors should be able to consider the initial statement, the procedures, and subsequent feedback instead of equating courtroom presentation with recognition accuracy. The National Research Council's recommendations specifically address this issue.

The post-identification feedback effect

The post-identification feedback effect is a documented phenomenon in which information provided after a choice changes a witness's retrospective reports of certainty or witnessing quality. In a landmark experimental study, Wells and Bradfield (1998) showed that confirming feedback could increase reports of confidence and alter perceptions of how clear the original view had been. The study manipulated feedback experimentally, making a causal claim about the feedback effect under its research conditions possible.

The case-specific record contains accounts of affirming remarks made after Thompson's selections. This pattern is consistent with the experimentally demonstrated mechanism; a historical case, however, does not allow scientists to calculate the exact amount by which those remarks changed Thompson's confidence. The record and the experiment perform different evidentiary functions.

Contemporaneous recording reduces uncertainty about these events. If investigators preserve a witness's own words before giving feedback, later experts and courts can evaluate changes across time. A retrospective question such as how certain the witness was at the beginning cannot fully reconstruct an unrecorded first response after years of additional exposure.

Repetition, familiarity, and possible source confusion

When one suspect appears in more than one lineup, the later display may contain a familiar face. Familiarity can arise from the crime scene or from a previous photograph. Distinguishing those sources is the challenge often discussed under source-monitoring principles. In Cotton's case, the repeated appearance of his face is documented; the precise subjective route by which Thompson made her choices remains inaccessible.

A later match to a previously selected photograph can feel internally consistent. Yet consistency across dependent tests has a different evidentiary value from two genuinely independent observations. If the same suspect is displayed repeatedly and the witness has been told that a previous choice was correct, the second result does not provide a fresh, independent measure of recognition.

The procedural lesson extends beyond lineups. Mug books, composite sketches, media exposures, social media, police communications, and courtroom encounters may become part of a witness's information history. Modern evidence preservation requires documenting exposures rather than assuming the final identification arose from a single untouched memory.

Was this a cross-racial identification case?

Thompson is a White woman and Cotton is a Black man, making the case relevant to discussions of identification across racial groups. Decades of research have found an average own-race advantage in recognition of faces; the meta-analytic review by Meissner and Brigham (2001) is an important scientific reference for this group-level phenomenon.

This evidence helps explain why cross-racial identification deserves care in investigative procedures and expert education. It cannot establish that racial difference caused Thompson's particular error, and it supplies no individual probability of being wrong in this case. The identification sequence, feedback, and independent DNA outcome are observable evidence; the causal weight of each psychological factor in this person's memory remains an interpretive question.

A fair approach recognizes both the scientific literature and its limits. Racial category does not determine the accuracy of a given witness, and racialized assumptions about defendants have no place in interpreting responsibility or credibility. Procedures should maximize accuracy for everyone and allow courts to evaluate contextual risks without reducing an individual to a statistical average.

Why was the actual perpetrator not recognized?

The retrospective record indicates that another man, Bobbie Poole, was implicated by DNA evidence. The National Research Council reports that Thompson had previously failed to recognize him when he was presented during post-trial proceedings. This outcome is striking because it shows the difference between the witness's genuine confidence in Cotton and the biological evidence that later identified the perpetrator. See National Research Council, Box 1-1 and the exoneration profile.

There are several plausible scientific pathways to an incorrect identification followed by failure to recognize the actual attacker: initial encoding limitations, comparison-based selection, repeated exposure to an innocent suspect, and reinforcement of a mistaken choice. The available records do not provide an experimental test capable of ranking these pathways for this individual. It would be inappropriate to treat any single mechanism as a proven exclusive cause.

The lesson is also about investigative breadth. When police and prosecutors focus on a mistaken identity, alternate-perpetrator evidence requires serious examination. The 1987 appellate decision illustrates the legal importance of allowing relevant evidence that implicates someone else, independently of whether any psychological account can be conclusively established.

What DNA testing resolved—and what it did not

The 1995 DNA results supplied independent biological evidence that excluded Cotton and linked the crime to Poole. This resolved the central identity question that memory-based testimony had answered incorrectly. The Innocence Project case profile reports the exoneration date and the role of physical evidence.

DNA evidence did not produce a time-stamped recording of Thompson's mental processes in 1984. It could not reveal precisely which moment or comment altered her beliefs, how face recognition unfolded internally, or which preventive intervention would alone have avoided the conviction. It established the error in the central identification and permitted the case to be examined against a known outcome.

The distinction between a verified outcome and an inferred mechanism is the core scientific discipline in historical case study. The certainty of Cotton's innocence, established through subsequent exonerating evidence, is stronger than any retrospective attribution of an individual psychological cause to Thompson, an investigator, or a juror.

Evidence hierarchy: court record, contemporary accounts, and later interpretation

Primary judicial opinions are strongest for what courts considered, which issues they decided, and what disposition followed. The 1987 Supreme Court decision documents the legal error involving alternate-suspect evidence, while the 1990 Court of Appeals opinion records later appellate treatment. A judicial opinion also summarizes testimony and trial evidence; its descriptions should be read as the court's account of the proceedings rather than a modern laboratory reconstruction of memory.

The DNA-based exoneration chronology is documented in the Innocence Project's case file. The National Research Council synthesizes the identification sequence as a teaching case within a larger consensus assessment. These sources differ in purpose and evidentiary function. Neither advocacy case summaries nor institutional research reviews replace the underlying judicial dispositions when a claim concerns the law.

Controlled studies test general mechanisms such as feedback, lineup fairness, or confidence calibration. They may identify effects that are relevant to the case, but cannot recreate every condition of a 1984 assault and subsequent legal process. Later first-person accounts can contribute perspectives on experience and reconciliation; they remain retrospective accounts, not direct measurement of cognitive processing at the original event.

Estimator variables and system variables

Eyewitness research differentiates circumstances investigators generally cannot change after a crime from procedures the justice system can control. The first group, commonly called estimator variables, includes viewing distance, exposure duration, lighting, stress, disguise, and aspects of face familiarity. These are evaluated to understand the circumstances in which the event was perceived.

System variables include the choice of lineup fillers, whether a suspect is presented repeatedly, whether the administrator knows the suspect's identity, what instructions are given, the manner of presentation, the recording of the procedure, and comments made after the choice. The distinction is elaborated in the 2014 National Research Council report and Wells and colleagues' 2020 review.

Cotton's case is a particularly instructive example of system variables because investigator choices after the crime were recorded and can be assessed against later scientific recommendations. Recognizing a system variable identifies an opportunity for procedural safeguards. It does not prove malicious conduct or mean an individual officer could have predicted the future DNA result.

How a lineup should be administered today

The National Research Council recommends blind or blinded administration of photo arrays and live lineups, standardized witness instructions, a verbatim confidence statement at the first identification, and video documentation. The American Psychology-Law Society scientific review of 2020 develops recommendations for collecting and preserving identification evidence.

Blind administration minimizes the possibility that an officer who knows the suspect will inadvertently cue the witness. Appropriate fillers reduce the chance that one photograph stands out for reasons unrelated to memory. Witness instructions should make clear that the perpetrator may be absent and that declining to identify anyone is a valid outcome. A first-person confidence statement should be captured before anyone supplies confirmatory or disconfirmatory information.

Recording the full procedure preserves more than a selected final answer. It provides access to hesitation, initial confidence, repeated viewing, officer prompts, and nonidentification decisions. These records can help a forensic expert, judge, jury, defense attorney, or prosecutor evaluate reliability without rebuilding the episode from memory months or years later.

Simultaneous and sequential lineups: a research distinction

In a simultaneous lineup, witnesses view several people together; in a sequential lineup, they see people separately. North Carolina's statutory framework adopts sequential presentation among its requirements. The legal choice of a format and the scientific assessment of comparative accuracy answer related but distinct questions.

Research on lineup presentation has examined the trade-off between correct identifications and mistaken choices under different conditions. The effect of presentation format cannot be considered independently of filler selection, blind administration, instructions, prior exposures, and the method by which diagnostic accuracy is measured. The National Research Council's review and 2020 scientific recommendations provide a more nuanced assessment than a claim that either format guarantees correct identification.

Cotton's case should therefore be used to teach a package of safeguards, not a single universal lineup trick. Reliable decisions emerge from transparent procedure, preservation of the original response, and independent corroboration.

North Carolina's Eyewitness Identification Reform Act

North Carolina enacted its Eyewitness Identification Reform Act in 2007, long after the 1984 investigation and the 1995 exoneration. The official Session Law 2007-421 supplies the legislative history, and N.C. Gen. Stat. § 15A-284.52 provides the current codified legal requirements.

The statute addresses independent administration or authorized alternative procedures, sequential presentation, instructions that the perpetrator may or may not be present, appropriate lineup construction, and documentation of the witness's words and confidence. These safeguards respond to known vulnerabilities of the identification process. Their precise scope, exceptions, and consequences of noncompliance are matters of North Carolina law, not universal rules of human memory.

The reform history is often associated with lessons from Cotton's case, but a reform statute has multiple sources of scientific, legislative, and institutional influence. The legislation should be described as a later response to systemic eyewitness problems rather than proof that this one case was the sole cause of every statutory provision. Other U.S. jurisdictions use different statutory or judicial approaches.

How should judges and jurors evaluate eyewitness evidence?

An identification should be examined as a sequence of events: the initial opportunity to observe, the earliest description, the first photographic or live identification, all later exposures, the initial confidence statement, any investigator feedback, and the presentation of evidence in court. The National Research Council recommends greater attention to these conditions and to the documentation available to fact-finders.

A witness's confidence, demeanor, or emotional conviction is a piece of courtroom testimony; it is not a forensic measurement of guilt. When initial confidence was never preserved, later testimony may be unable to answer the most relevant scientific question. Independent corroboration, alternative-suspect evidence, and potential contamination deserve consideration in the context of the applicable evidence rules.

Legal thresholds such as admissibility, proof beyond a reasonable doubt, and state-specific safeguards are legal standards. A psychologist may explain mechanisms and empirically supported limitations but does not replace the court or jury as the authorized legal decision-maker. The limits of admissible expert testimony differ across federal and state courts and across countries.

The limits of identifying a person from one case

A single exoneration proves that an individual identification was mistaken. It does not establish the overall population error rate of eyewitness testimony, the predictive validity of all lineups, or a numerical probability that another specific witness is wrong. Those questions require appropriately designed research and base-rate information.

Even a strong general research result needs attention to the population studied, the lineup's conditions, the prevalence of guilty suspects in tested samples, and whether confidence was recorded at the first uncontaminated decision. Wixted and Wells (2017) distinguish strong initial-confidence evidence under fair conditions from weakly diagnostic confidence obtained after contamination.

The case further underscores the absence of an individual diagnostic test for sincerity. A person who gives incorrect testimony may be truthful about what they believe. Conversely, psychological theories of memory cannot be used to classify every confident witness as accurate or inaccurate without examining the procedure and supporting evidence.

Ethical considerations: survivors, exonerees, and public responsibility

A responsible account acknowledges Thompson's experience as a survivor of sexual violence and Cotton's experience as a person wrongfully imprisoned. Explaining misidentification does not require assigning blame to the witness for having survived an attack or suggesting that all witnesses act alike. The social and institutional consequences of the error should be described without sensationalizing either person.

Cotton's exoneration does not erase the years lost to incarceration. Thompson's later understanding of the mistake does not undo the original assault. Their later public engagement may be described as part of the historical story, but a researcher should avoid asserting unverified diagnoses, private emotional states, or universal narratives of forgiveness. No clinical evaluation is established by the published case materials cited here.

At the institutional level, ethical duties include fair investigation, preservation of exculpatory material, accurate presentation of expert evidence, and procedures that protect both victims and suspects. A forensic psychological opinion should declare the facts on which it rests, describe scientific support and uncertainty, and separate observed actions from inferred mental mechanisms.

Common myths the case corrects

One persistent myth says that sincere eyewitnesses are necessarily accurate. Thompson's good-faith identification was mistaken. The scientific response is to preserve and evaluate the conditions under which an identification was made.

Another myth says that confidence has no evidentiary value whatsoever. Research supports a conditional relationship: immediate high confidence from a properly administered first identification may be informative, while confidence obtained after suggestive procedures or feedback may be much less informative. The relevant distinction is documented in the 2017 review.

A third myth suggests that courtroom testimony provides an independent new test of memory. By the time of trial, the witness often has been exposed to earlier identification results, investigators, counsel, and repeated accounts. The testimony is part of a connected information history.

A final myth treats reform as evidence that human memory is generally useless. The goal of reform is to improve the collection, preservation, and interpretation of useful eyewitness evidence. Accurate identifications can contribute to justice; better procedures help distinguish them from potentially mistaken ones.

Practical lessons for police, lawyers, forensic experts, and journalists

For investigators, the highest-value lesson is to protect the first test of identity. Use a fair lineup, neutral instructions, blind administration where prescribed, and a contemporaneous record of the witness's words. Avoid confirming a choice before the confidence statement is documented. Record every exposure and preserve nonidentifications and alternate-suspect leads.

For lawyers and courts, reconstruct the full identification chronology. Compare the earliest description and confidence with later statements, examine whether the same person was repeatedly presented, and distinguish procedural evidence from independent corroboration. Seek expert interpretation when complex scientific issues warrant it, while retaining the jurisdiction's proper legal standards.

For educators and journalists, explain the court's rulings correctly: the 1987 reversal addressed the exclusion of evidence implicating another possible perpetrator; the 1995 DNA findings established Cotton's innocence. Use the case to illuminate established mechanisms, without claiming a retrospective mental diagnosis or a scientifically measured cause of one witness's error.

For readers, the broader lesson is that memory is human evidence whose value depends on collection conditions. Fair procedures protect witnesses from unnecessary uncertainty, protect innocent suspects from mistaken prosecution, and protect the public's interest in identifying the person who committed the offense.

Frequently asked questions

Why did Jennifer Thompson identify Ronald Cotton?

The historical record documents a tentative photographic selection, a later live lineup in which Cotton was the only person repeated, and confirming feedback. These conditions are relevant to established research on recognition, repetition, and feedback. They do not permit a definitive assignment of one psychological cause to the individual error. National Research Council.

How was Ronald Cotton proven innocent?

DNA testing of crime-related biological evidence in 1995 excluded Cotton and implicated Bobbie Poole. Cotton was exonerated on June 30, 1995, after more than ten years in prison. Innocence Project.

Did the North Carolina Supreme Court overturn the conviction because memory is unreliable?

No. In 1987 the court ordered a new trial on specified charges because the trial court had excluded relevant evidence pointing toward another perpetrator. That is an evidentiary ruling. DNA exoneration came later. State v. Cotton (1987).

Is eyewitness confidence correlated with accuracy?

Under carefully controlled initial identification conditions, confidence can be meaningfully associated with accuracy. Feedback, suggestive procedures, repeat testing, and other contamination may reduce the evidentiary value of later confidence. The timing and conditions of the statement matter. Wixted and Wells (2017).

What is the post-identification feedback effect?

It is the experimentally documented tendency for confirming information supplied after an identification to change a witness's later reports of confidence and witnessing quality. It should be evaluated as a research mechanism, not as a diagnosis. Wells and Bradfield (1998).

Did cross-racial identification cause the error?

The case involves identification across racial groups, and meta-analytic research reports average recognition differences across groups. The case records do not isolate that factor as the cause of Thompson's error. Group-level findings cannot supply an individualized causal conclusion. Meissner and Brigham (2001).

What reforms resulted from growing awareness of eyewitness error?

Research and justice-system reviews recommend fair lineups, independent or blind administration, instructions that the perpetrator may be absent, first-confidence documentation, and recording. North Carolina enacted a specific Eyewitness Identification Reform Act in 2007. National Academies recommendations; North Carolina statute.

Can psychological science determine guilt from a witness's memory?

Psychology can evaluate empirical reliability, procedural risk factors, and the limits of particular interpretations. Guilt is determined through the applicable legal process using the full admissible evidence. No single psychological construct or confidence statement substitutes for legal proof.

References

Innocence Project. (n.d.). Ronald Cotton. Case profile.

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 Institute of Justice. (1999). Eyewitness evidence: A guide for law enforcement. U.S. Department of Justice. Official report.

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

North Carolina General Assembly. (2007). Eyewitness Identification Reform Act, Session Law 2007-421. Official legislation.

North Carolina General Assembly. (current codification). N.C. Gen. Stat. § 15A-284.52, Eyewitness identification reform. Official statute.

State v. Cotton, 318 N.C. 663, 351 S.E.2d 277 (N.C. 1987). Full judicial opinion.

State v. Cotton, 99 N.C. App. 615, 394 S.E.2d 456 (N.C. Ct. App. 1990). Full judicial opinion.

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., & 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.

Psychological Encyclopedia