Dossier mode
One Hundred Percent Certain
The same investigation, restaged one beat at a time. Drive it with the arrow keys, space, or autoplay. Nothing is cut from the piece — long runs are split across frames. Read the full investigation or open the Judges & Prosecutors hub.
One Hundred Percent Certain.
For three decades, Texas prosecutors seeking death had a psychiatrist who'd swear — with '100% certainty,' sometimes without meeting the defendant — that the man would kill again. The press called him Dr. Death. His profession called him a charlatan and expelled him.
James Grigson testified in 167 capital trials, nearly all ending in death, supplying the 'future dangerousness' prediction Texas law requires before a jury can impose death — with a certainty no honest clinician could offer, often without examining the defendant. The scandal isn't one strange doctor; it's the machine that used him: a death penalty that demanded a scientific-sounding prophecy created a market for a man who'd fake it, and prosecutors kept calling him because he worked.
All FACT (Wikipedia, DPIC, Northwestern Wrongful Convictions, Barefoot v. Estelle). FENCE: NOT a claim of mass innocence (many were guilty — the wrong is the sentence + the faked prediction); NOT anti-psychiatry (his own field's associations expelled him); NOT that he alone caused each verdict (prosecutors chose him, judges admitted him, juries believed him). The indictment is structural.
167 capital trials, nearly all ending in death — the prosecution's reliable witness.
Grigson was an institution, not an occasional expert. Over ~three decades he testified in 167 Texas capital trials, overwhelmingly for the prosecution, nearly all resulting in death sentences — earning the press nickname 'Dr. Death.' His value was predictability: where another psychiatrist might hedge, Grigson could be counted on to tell the jury, plainly and confidently, that this defendant was a sociopath who would certainly kill again. A dependable component of the Texas death-sentencing machine.
What he sold was junk the profession rejected: '100 percent' certainty psychiatry can't deliver.
Texas capital sentencing turns on predicting whether the defendant will be a 'continuing threat.' Grigson answered he was certain — often '100 percent' certain — of future violence, routinely without examining the defendant, working from a prosecutor's hypothetical. The field doesn't support that. In Barefoot v. Estelle (1983), the American Psychiatric Association told the Supreme Court that psychiatric predictions of long-term future dangerousness are wrong more often than right (~2 of 3). Grigson sold, as medical science, exactly what his profession said it couldn't do.
Randall Adams: the innocent man Grigson helped send to death row.
The cost made concrete. In 1977 Randall Dale Adams was convicted of the 1976 murder of Dallas officer Robert Wood and sentenced to death; Grigson testified he was an incurable sociopath certain to kill again. Adams didn't do it — the shooter was teenager David Harris, who'd picked Adams up after his car ran out of gas. Filmmaker Errol Morris unraveled it in the 1988 documentary The Thin Blue Line; the conviction was overturned in 1989 after 12 years in prison. The detail that captures everything: Morris got interested while researching a film about Grigson himself. The pursuit of Dr. Death led straight to an innocent man on death row.
In 1995 his own profession expelled him — for diagnosing people he'd never met and faking certainty.
The reckoning came from inside medicine. In 1995 the APA and the Texas Society of Psychiatric Physicians expelled Grigson on two grounds: diagnosing without examining the individuals, and testifying he could predict with 100% certainty that they'd commit future violence. Both are ethics violations because they dress unsupported conclusions in medical authority. Grigson blamed 'anti-death-penalty doctors' — but the charge wasn't his politics; it was that he'd testified, under oath and in a lab coat, to things his discipline holds cannot be known.
The machine kept the verdicts: even after he was disgraced, the death sentences he built mostly stood.
What makes it a system story, not a biography. Being expelled as a charlatan in 1995 didn't undo the 167 trials or reliably reopen the sentences his testimony secured. Texas courts kept upholding death sentences resting on his evidence after his discrediting — DPIC flagged the Texas Court of Criminal Appeals doing exactly that. The courtroom had extracted what it wanted (a jury's future-dangerousness finding) and had little appetite to revisit it. Junk science, once admitted and acted on, is extraordinarily hard to pull back out of a verdict — the strongest argument against letting it in.
The line we hold.
- FACT: Grigson ('Dr. Death') testified in 167 Texas capital trials, overwhelmingly for the prosecution, nearly all ending in death sentences.
- FACT: he testified to 'future dangerousness' with '100% certainty,' often without examining the defendant; the APA told the Supreme Court (Barefoot v. Estelle, 1983) such predictions are wrong more often than right.
- FACT: Randall Dale Adams — Grigson said he'd certainly kill again — was innocent (real killer David Harris), exonerated 1989 after The Thin Blue Line (a film Morris began about Grigson).
- FACT: the APA and Texas Society of Psychiatric Physicians expelled Grigson in 1995 for diagnosing-without-examining and claiming 100% certainty.
- FACT: his verdicts mostly stood — Texas courts upheld death sentences built on his testimony even after his discrediting.
- FENCE: NOT mass innocence (guilt of the crime ≠ a reliable prediction of future violence); NOT anti-psychiatry (his own field expelled him); NOT he-alone (prosecutors chose him, judges admitted him, juries believed him). The market for false certainty is created by the law, not one doctor.
Why it matters.
A jury can discount a jailhouse snitch and squint at a shaky eyewitness; it's far harder to doubt a doctor who states his credentials and says he's certain. That's the power junk forensic science borrows — not the strength of its evidence, but the authority of the field it claims to speak for. Grigson is the purest case: not a flawed test or overstated match but a prediction the whole discipline says cannot be made, delivered with a confidence the discipline calls fraudulent, with the maximum penalty at stake. He's unthinkable without the machinery that created him: after 1972's Furman struck down the death penalty, Texas rebuilt its statute around a 'special issue' of future dangerousness (upheld 1976), writing a prediction into the path to the execution chamber — and the adversarial system rewards the expert who gives a side a clean, confident answer over the one who admits uncertainty, so the incentives select for Grigsons. When the Supreme Court could have shut it down in Barefoot v. Estelle, with the APA warning the predictions were worse than a coin flip, it let the testimony continue, trusting juries to sort good science from bad; the Grigson record is the answer to whether they can — 167 times a jury heard '100 percent certain,' and 167 times, near enough, it believed him. Randall Adams was a broke stranger with a public defender; the real killer was a local teenager the system was slow to charge. The lab coat didn't create that imbalance, but it put a scientific seal on it — which is what junk science does best. The fix isn't just disbarring the next Dr. Death; it's questioning a death penalty that makes a fortune-teller a prerequisite for an execution.
Help us fill these lines.
- OpenA case-by-case review of the other 166 trials — how many sentences rested materially on Grigson's testimony, and how many were ever re-examined after his 1995 expulsion.
- OpenThe full appellate record of courts upholding death sentences built on his evidence post-discrediting, to map how the system insulates a verdict from a debunked expert.
- OpenA companion on forensic junk science more broadly (bite marks, hair microscopy, arson 'science') as a leading documented cause of wrongful convictions — the pattern Grigson belongs to.