← back to the files
VC-011 Wrongful conviction · North Carolina 2014

Henry McCollum & Leon Brown — two disabled teenagers signed to death row, cleared by a cigarette

Years lost
~30–31 years served (2 brothers)
Charge
Rape & first-degree murder (1983)
Cleared
Convictions vacated, Sept 2, 2014
Status
Exonerated

Summary

In Red Springs, North Carolina, in the autumn of 1983, two intellectually disabled half-brothers — Henry Lee McCollum, then nineteen, and Leon Brown, then fifteen — were arrested for the rape and murder of eleven-year-old Sabrina Buie, and they were exonerated three decades later, on September 2, 2014, after DNA from a cigarette butt left at the scene was matched to a different man, the serial offender Roscoe Artis. There was never any physical evidence tying either brother to the crime. The convictions rested on handwritten confessions that police produced during unrecorded interrogations and that both teenagers recanted almost at once. McCollum, who had an IQ measured as low as 51, and Brown, measured near 49, were precisely the kind of suggestible defendants for whom such statements are least reliable, and the case is now studied as a definitive example of how a coerced confession from a vulnerable juvenile can survive the absence of corroborating proof.

The outcome is a matter of record. The brothers were convicted together in October 1984 and both sentenced to death. After the North Carolina Supreme Court overturned those sentences in 1988 on instruction errors, McCollum was retried in 1991, again convicted, and again sent to death row; Brown was retried in 1992, convicted of rape, and sentenced to life. McCollum spent roughly thirty years on death row — the longest-serving condemned prisoner in the state — while Brown served a comparable term in the general prison system. On September 2, 2014, Robeson County Superior Court Judge Douglas Sasser vacated both convictions after the North Carolina Innocence Inquiry Commission's DNA work pointed away from them and toward Artis; the district attorney did not contest the finding. The brothers were released the following day.

In June 2015, Governor Pat McCrory granted both men pardons of innocence, and the state paid each the statutory maximum of 750,000 dollars in compensation. A federal civil-rights suit followed. On May 14, 2021, a jury awarded the brothers 75 million dollars — 31 million dollars each in compensatory damages, calculated at roughly one million dollars per year lost, plus 13 million dollars in total punitive damages. A federal appeals court later reduced the figure to account for earlier settlements, but the verdict stood as one of the largest of its kind in the country.

This dossier centers McCollum and Brown as the wronged parties. The system failures — the interrogation of disabled children without protection, a confession credited over the lack of evidence, and a suppressed request to compare crime-scene fingerprints against the real killer — are the mechanism of the wrong. Roscoe Artis is named as the perpetrator only because the record, anchored by the DNA match and his own established history, establishes it.

Timeline

September 1983.
Eleven-year-old Sabrina Buie disappears in Red Springs, Robeson County; her body is found in a soybean field near McCollum's mother's home. She had been raped and suffocated with her own underwear.
Late September 1983.
After a tip from a local teenager, police interrogate McCollum (19) and Brown (15) for hours, without a lawyer or guardian present. Both, intellectually disabled, sign confessions written out by detectives and recant soon after.
October 5, 1983.
Red Springs police request a comparison of fingerprints from beer cans at the scene against prints from Roscoe Artis. The request is cancelled and never disclosed to the defense.
October 22, 1983.
Weeks after the Buie murder, Artis confesses to the rape and murder of eighteen-year-old Joann Brockman in the same town; he is later sentenced to death, then life.
October 1984 — convictions.
In a joint trial under District Attorney Joe Freeman Britt, both brothers are convicted of first-degree murder and rape and sentenced to death. Brown, at sixteen, becomes the youngest person then on North Carolina's death row.
1988 — sentences overturned.
The North Carolina Supreme Court vacates both death sentences over jury-instruction errors and orders new trials.
1991–1992 — retrials.
McCollum is retried, reconvicted, and again sentenced to death. Brown is retried and convicted of rape, drawing a life sentence; the conviction is upheld in 1995.
2005–2010 — the reinvestigation begins.
Defense and innocence lawyers pursue DNA testing of crime-scene evidence; testing of a cigarette butt excludes both brothers. Brown asks the North Carolina Innocence Inquiry Commission for help in 2010.
July 2014 — the DNA match.
The Commission's testing matches the cigarette-butt DNA, through the state database, to Roscoe Artis — who had lived roughly a block from the scene.
September 2, 2014 — vacatur.
Judge Douglas Sasser vacates both convictions and declares the brothers innocent; the district attorney does not oppose it. They are released the next day after about thirty years.
June 2015 — pardons.
Governor Pat McCrory pardons both men; the state later pays each 750,000 dollars in compensation.
May 14, 2021 — civil verdict.
A federal jury awards the brothers 75 million dollars — 31 million each in compensatory damages plus 13 million in punitive damages.

The Rooms Where the Words Were Written

The case against the brothers was built almost entirely from statements they did not write and could barely read. McCollum's interrogation ran for hours and ended near half past two in the morning; he was told a witness had placed him at the scene, was pressed with the suggestion that confessing was the way home, and signed a statement, drafted by detectives, that implicated several other local boys. None of those boys were ever charged. Brown, nearly illiterate, did not so much sign his statement as print his name in block letters at the bottom of a document whose contents he could not have composed.

What the brothers brought to those rooms was precisely what made the rooms dangerous. McCollum's measured IQ sat around 51 and Brown's near 49 — both far below the threshold psychiatrists treat as the marker of intellectual disability, and both consistent with a mental age closer to childhood than to the adults the law was about to make of them. Decades of research, much of it postdating this case, shows that people with intellectual disabilities are disproportionately likely to comply with authority, to be led by suggestion, and to confess falsely in the belief that agreement will end their distress. The interrogation exploited exactly those vulnerabilities, and no counsel or informed adult was present to interrupt it.

The procedural protections that should have governed the questioning of a fifteen-year-old were simply absent. There was no recording of the long hours that produced the statements, only the signed pages at the end — so the process that manufactured the confessions could never be examined, only its product. A confession from the actual killer should lock onto the scene; these did not. They named the wrong people, fit no physical evidence, and were repudiated almost immediately. The failure was not a single rogue detective but a permitted method applied to two children who could not withstand it.

The Fingerprint That Was Never Compared

The most damning fact about the prosecution is not what it presented but what it buried. On October 5, 1983 — three days before the first trial would later be set in motion, and within weeks of the murder — Red Springs police formally asked the State Bureau of Investigation to compare fingerprints lifted from beer cans at the Buie scene against prints belonging to Roscoe Artis. The request was cancelled. It was never turned over to defense lawyers, and the jury that sent two teenagers to death row never learned it existed.

Artis was not a random name. By 1983 he was a man with a sexual-assault history reaching back decades, and within a month of Sabrina Buie's killing he confessed to the strikingly similar rape and murder of another young woman in the same small town. He lived about a block from the field where Buie's body was found. Every element that should have made him the obvious suspect was available to investigators in 1983, and the one forensic step that might have surfaced him — the fingerprint comparison — was initiated and then quietly shut down.

The consequence of that suppression compounds across thirty years. Had the comparison gone forward, the case against the brothers might have collapsed before it began, and an active, dangerous offender would not have been left to the years that followed. Instead the state convicted the two people the evidence did not implicate, and declined to test the one it did. Tunnel vision around a pair of coerced confessions did not merely accompany this wrongful conviction; an affirmative decision not to look at the real killer made it.

Thirty Years, and the Test That Ended Them

The correction, when it came, came from outside the institutions that had produced the error. The brothers' lawyers and the North Carolina Innocence Inquiry Commission — a state body created precisely to reexamine claims like this one — pursued DNA testing of the physical evidence that had sat unexplained since 1983. A cigarette butt recovered at the scene was tested and excluded both McCollum and Brown. Then, in the summer of 2014, the Commission's work matched that same DNA, through the state database, to Roscoe Artis, already serving life for the murder he had confessed to weeks after Buie's.

On September 2, 2014, Robeson County Superior Court Judge Douglas Sasser vacated both convictions and found the brothers innocent. The district attorney did not contest the result. McCollum walked off death row after roughly thirty years — the longest any condemned prisoner had then spent there — and Brown was freed from a life term, the two of them released into a world that had moved three decades without them. The years inside had been brutal: McCollum had attempted suicide and watched dozens of executions from his work as a death-row janitor; Brown had cycled through psychosis and assault; their grandmother and mother had both died while they were imprisoned.

The reckoning did not arrive seamlessly. The same case had, years earlier, been cited by a sitting Supreme Court justice as a justification for capital punishment — McCollum offered as proof that some crimes demand death — even as another justice had pointed to his disability and childlike mental age to argue the opposite. That an innocent man's name had been used to defend the very system that nearly killed him is the sharpest measure of how completely the error had been mistaken for certainty, and of how heavily a justice system can resist conceding a false confession until DNA forces the concession.

The Five Factors

01
Coercive interrogation of the intellectually disabled
Long, unrecorded questioning of suggestible people — here two disabled adolescents — is a near-ideal machine for producing false confessions. Individuals with intellectual disabilities are disproportionately prone to comply, to be led by suggestion, and to confess in the belief that agreement ends the ordeal. Without counsel or an informed adult, such a defendant cannot resist the pressure that yields words mistaken for guilt.
02
A confession credited over the absence of evidence
No physical evidence linked either brother to the crime, and their statements named uncharged boys and fit nothing at the scene. A sound process treats an uncorroborated, recanted confession as suspect and the lack of forensic linkage as exculpatory. Here the order inverted: the signed pages were believed and the empty evidentiary record was disregarded. When a statement outranks the proof that should support it, error becomes a verdict.
03
Suppression of evidence pointing to the real perpetrator
Police asked to compare crime-scene fingerprints against Roscoe Artis, then cancelled the request and never disclosed it. Withholding evidence that implicates an alternative suspect does not merely weaken the defense; it forecloses the truth. A disclosure obligation honored only when convenient is no obligation at all.
04
Tunnel vision around the first suspects
Once investigators had the brothers' confessions, an obvious alternative — a neighbor with a sexual-assault history who confessed to a near-identical killing weeks later — was not pursued. Contrary indications were absorbed into the existing theory rather than allowed to break it. Confirmation bias let a frame survive facts that should have dismantled it.
05
Correction required an outside DNA match, not self-scrutiny
The convictions were undone not by the courts auditing themselves but by a state innocence commission and a database hit to a different man. A justice process that can only be corrected when independent science identifies the true offender has no reliable internal mechanism for catching its own mistakes — so most such mistakes are never caught.

Aftermath

The civil reckoning was substantial. In June 2015 Governor Pat McCrory pardoned both men, and the state paid each the statutory maximum of 750,000 dollars. Local settlements followed, and on May 14, 2021 a federal jury awarded the brothers 75 million dollars — 31 million each in compensatory damages, scaled to roughly one million dollars per year lost, plus 13 million in punitive damages. A federal appeals court later trimmed the total to account for the earlier payments, but it remained among the largest wrongful-conviction awards in the nation. No sum returned the thirty years, the dead relatives, or the youth that the convictions had consumed.

The case reshaped how North Carolina's failure was understood. It became a fixture in the national argument over capital punishment precisely because the state had come within an execution of killing an innocent, intellectually disabled man, and because the same case had been invoked to defend the death penalty before DNA reversed it. The North Carolina Innocence Inquiry Commission — the body that finally undid the error — stood as a rare institutional answer to a problem most states leave to chance, a standing mechanism for reopening convictions that the ordinary appeals process had repeatedly affirmed.

The durable ripple is a warning about who the system convicts most easily. Two children who could not read their own confessions, could not navigate an interrogation, and could not command a competent defense were the ones the machinery found simplest to condemn — and hardest, for thirty years, to release.

Lessons

  1. Record custodial interrogations in full, and never question a juvenile or an intellectually disabled suspect without counsel and a genuinely informed adult present.
  2. Treat an uncorroborated, recanted confession as a red flag, not a foundation; the absence of physical evidence should weigh against conviction, not be waved aside.
  3. Enforce disclosure of all evidence pointing to alternative suspects, including aborted or cancelled forensic tests, as an absolute duty rather than a discretionary courtesy.
  4. Pursue the obvious alternative suspect — here a neighbor who confessed to a near-identical crime weeks later — instead of folding contrary facts into a fixed theory.
  5. Fund standing innocence commissions with real testing authority, so exoneration does not depend on the chance survival of a single cigarette butt and a database match decades later.

References