Ronald Cotton — a confident eyewitness, a wrong man, and DNA eleven years late

Summary

In Burlington, North Carolina, Ronald Cotton served roughly eleven years for a 1984 rape he did not commit, convicted on the confident, sincere, and mistaken eyewitness identification of the victim, and was exonerated in 1995 when DNA testing excluded him and matched another man, Bobby Poole. Cotton was first convicted in January 1985 and sentenced to life plus fifty years; after a retrial in 1987 he was convicted again and sentenced to life plus fifty-four years. The case turned almost entirely on the identification by Jennifer Thompson, then a 22-year-old college student, who had studied her attacker's face during the assault specifically so she could later identify him — and who pointed, in good faith, to the wrong man.

The outcome is settled. In the spring of 1995, DNA testing of the rape-kit evidence excluded Cotton and matched Bobby Poole, a man already imprisoned for other crimes who had reportedly boasted to fellow inmates that he had committed the assaults for which Cotton was serving time. On June 30, 1995, Cotton was released; Poole pleaded guilty on July 11, 1995; and Governor James B. Hunt Jr. pardoned Cotton the following day. North Carolina later compensated him $110,000.

What makes the case a landmark is not only the error but what followed it. Jennifer Thompson, devastated to learn she had condemned an innocent man, eventually reconciled with Cotton, and the two became close friends and joint advocates for eyewitness-identification reform. Their 2009 memoir, Picking Cotton, and their public testimony helped drive changes to how police conduct lineups — sequential presentation, blind administration, and recorded confidence statements — across many jurisdictions.

This dossier centers Cotton and the assault victims as the wronged parties. The system failure — a conviction founded on honest but mistaken cross-racial eyewitness identification, hardened by suggestive procedures and a denied alternate-suspect defense — is the mechanism of the wrong. Bobby Poole is named as the perpetrator only because DNA and his guilty plea established it.

Timeline

July 29, 1984.

A man breaks into the Burlington, North Carolina apartment of Jennifer Thompson, a 22-year-old student, and rapes her at knifepoint. She deliberately studies his face. A second woman is assaulted nearby the same night.

The composite.

Thompson works with police to build a composite sketch of her attacker, described as a tall, light-skinned Black man.

The tip and array.

A tip names Ronald Cotton, 22, who resembles the sketch. Shown a photo array, Thompson selects Cotton as "number five." She later identifies him again in a physical lineup.

The arrest.

Cotton is arrested. He offers an alibi but is charged with first-degree rape, first-degree burglary, and first-degree sexual offense.

January 1985 — first conviction.

Cotton is convicted of the rape and burglary of Thompson and sentenced to life plus fifty years.

The other suspect surfaces.

In prison Cotton encounters Bobby Poole, who reportedly bragged to other inmates that he had committed the crimes. Cotton's defense seeks to present Poole as the real attacker.

1987 — retrial.

After the first conviction is overturned on appeal, Cotton is retried, now for both assaults. The court bars the Poole alternate-suspect evidence after both victims say Poole was not their attacker. Cotton is convicted again and sentenced to life plus fifty-four years.

Spring 1995 — DNA testing.

At the defense's request, Burlington police turn over the rape-kit evidence for DNA testing, a technology unavailable at the original trials.

The exclusion and match.

The DNA excludes Cotton and matches Bobby Poole, confirming what Poole had reportedly admitted in prison.

June 30, 1995 — release.

Cotton is freed after roughly eleven years. Poole pleads guilty on July 11, 1995; Governor James B. Hunt Jr. pardons Cotton on July 12, 1995.

2009 — reform partnership.

Cotton and Thompson publish Picking Cotton and campaign together for eyewitness-identification reform; North Carolina has by then adopted reformed procedures.

The Face She Memorized

The cruelty of the case lies in how diligent the witness was. During the assault, Jennifer Thompson made a deliberate decision to study her attacker's face — to memorize details so that, if she survived, she could help convict him. She was not a careless or indifferent witness; she was an unusually determined one, and that determination is precisely what made her error so persuasive to a jury. When she identified Ronald Cotton, she did so with the conviction of someone who had consciously worked to remember.

But memory does not store a face like a photograph, and the effort to remember does not immunize the result against contamination. Cross-racial identification compounds the difficulty: research consistently shows that witnesses are less accurate identifying faces of a different race than their own, and Thompson, who is white, was identifying a Black man seen under terror and a knife. The face she had worked to fix in memory was a real face — just not Cotton's.

The case dismantles the intuition that a confident, sincere, well-motivated eyewitness is a reliable one. Confidence and accuracy are only loosely related; a witness can be entirely certain and entirely wrong. Thompson's sincerity was never in question, and that is the point: the danger of eyewitness identification is not lying witnesses but honest ones whose memories have failed them in ways no jury can detect from the witness stand.

How the Mistake Was Locked In

A single misidentification need not become a conviction; the procedures around it determine whether an initial error is caught or cemented. Here they cemented it. Cotton entered the case as a man who resembled a composite sketch — a likeness, not evidence. Shown a photo array, Thompson selected him; shown a later physical lineup, she identified him again. But once a witness has chosen a face from photographs, seeing that same face in a subsequent lineup tends to confirm the earlier choice rather than test it independently. Each procedure, instead of providing a fresh check, reinforced the one before.

By the time Thompson testified, her in-court identification may have reflected the cumulative weight of the array and the lineup as much as any memory of the crime itself — yet to the jury it appeared as direct, first-hand recognition of the attacker. The chain of identifications had converted a resemblance into an apparent certainty, with each link drawing its strength from the last.

The retrial then closed the last exit. When Cotton's defense tried to present Bobby Poole — who had reportedly admitted the crimes to fellow inmates — as the true perpetrator, the court excluded the evidence because both victims maintained Poole was not their attacker. The same eyewitness identification that wrongly convicted Cotton was thus used to bar the very evidence that pointed to the right man. The error was not merely made; it was sealed, with the victims' good-faith certainty serving as the lock.

Eleven Years to a Testable Truth

What finally broke the case was evidence that had been physically preserved but technologically unreadable for over a decade. The rape kits taken in 1984 contained biological material from the actual attacker, but DNA profiling was not available when Cotton was tried in 1985 or retried in 1987. For eleven years the answer sat in an evidence locker, conclusive and inert, while Cotton served a life sentence built on memory.

In the spring of 1995, at the defense's request, the Burlington Police Department turned the evidence over for DNA testing. The result was unequivocal twice over: it excluded Ronald Cotton and it matched Bobby Poole, confirming forensically what Poole had reportedly already confessed informally behind bars. On June 30, 1995, Cotton walked out; Poole pleaded guilty within two weeks; and the governor's pardon followed the next day.

The interval is the lesson. Cotton's exoneration did not require any new perception, witness, or investigative insight — only the application of a test to evidence that had existed all along. He lost eleven years not because the truth was unknowable but because the means of reading it arrived late, and because the system, in the interim, had treated a confident identification as proof enough to foreclose the question. The correction came from outside the original case theory entirely, carried in by a technology the trials never had.

The Five Factors

01

Honest but mistaken eyewitness identification

The case rested on a sincere, determined witness who was simply wrong. Eyewitness memory is reconstructive and fallible, and certainty is no guarantee of accuracy. A conviction built on identification alone has no defense against the most dangerous kind of error — the confident mistake that no jury can see through.

02

The cross-racial identification problem

Witnesses are measurably less accurate identifying faces of another race than their own. Thompson, identifying a man of a different race under extreme stress, faced a recognized and elevated risk of error that the trial treated as ordinary eyewitness testimony. Ignoring that documented difficulty builds a known unreliability into the verdict.

03

Suggestive and compounding ID procedures

A composite sketch, then a photo array, then a lineup did not provide independent checks; each reinforced the last. Once a face is chosen from photos, later identifications tend to confirm rather than test the choice. Procedures that compound instead of cross-checking manufacture a false impression of consistency.

04

The same error used to bar the truth

When Cotton tried to name Bobby Poole as the real attacker, the victims' identification of Cotton was used to exclude the evidence. A wrongful identification that is allowed to foreclose alternate-suspect evidence becomes self-sealing — the original error suppresses the very proof that would expose it.

05

Correction required testing evidence the trial never had

Cotton was freed only when DNA testing was applied to long-preserved evidence, not through any internal reexamination. A process whose mistakes can be corrected only by a later technology reading old evidence has no reliable way to catch its own errors — and depends entirely on that evidence surviving to be tested.

Aftermath

North Carolina pardoned Ronald Cotton in 1995 and later paid him $110,000 — a sum that could not return eleven years. But the case's most significant legacy grew from an extraordinary reconciliation. Jennifer Thompson, on learning that her identification had condemned an innocent man and freed the guilty one to assault others, sought Cotton out; he forgave her, and the two became close friends and partners in advocacy. Their account, the 2009 memoir Picking Cotton: Our Memoir of Injustice and Redemption, became a widely read text on the fallibility of memory and the possibility of repair.

Together they became among the most effective advocates in the United States for reforming eyewitness identification. Their testimony helped advance practices now widely recommended: double-blind administration, in which the officer running a lineup does not know who the suspect is; sequential presentation of photos rather than all at once; clear instructions that the perpetrator may not be present; and recording the witness's confidence at the moment of identification. North Carolina became an early adopter of such reforms.

The durable ripple is a reframing of eyewitness testimony itself — from the gold standard juries once assumed it to be, into a form of evidence understood as powerful but error-prone, to be handled with safeguards. Cotton's wrongful conviction, and the partnership it improbably produced, helped move that understanding from academic literature into police procedure and courtroom practice.

Lessons

Do not convict on eyewitness identification alone; treat even a confident, sincere identification as fallible evidence requiring corroboration.
Account explicitly for the heightened error rate in cross-racial identifications rather than treating them as ordinary testimony.
Administer lineups double-blind and sequentially, warn that the suspect may be absent, and record the witness's confidence at the moment of the identification.
Never let a contested identification be used to bar alternate-suspect evidence; the proof of the real perpetrator must not be sealed out by the original error.
Preserve biological evidence and guarantee post-conviction DNA testing, so a truth that is merely unreadable today is not lost for a decade.

References

Ronald Cotton INNOCENCE PROJECT
Ronald Cotton WIKIPEDIA
Jennifer Thompson and Ronald Cotton — Special Courage Award U.S. OFFICE FOR VICTIMS OF CRIME
Justice Department Honors "Picking Cotton" Authors U.S. DEPARTMENT OF JUSTICE