Proofiness - Charles Seife [80]
Luckily, a few sharp-eyed statisticians cried foul when they heard about Meadow’s testimony. A blistering article in the British Medical Journal entitled “Conviction by Mathematical Error?” attacked Sir Roy’s numbers, and when Clark lost her appeal, the president of the Royal Statistical Society wrote to the lord chancellor—then the head of Britain’s judiciary—declaring the 1 in 73 million figure “invalid.”
In retrospect, it appears that Clark was truly innocent. One fact that prosecutors never presented at her trial is that her second child had been suffering from a significant staph infection of the central nervous system—it was quite likely the cause of his death. Clark’s second appeal in 2003 was successful, and three other women who had been convicted with the aid of Meadow’s testimony were also set free.
Prosecutors rely on expert witnesses such as Roy Meadow to feed facts—real or fabricated—to juries to help their cases. The right expert can make it easy to clear even high legal barriers. In Texas, a convict can get the death penalty only if a jury finds that he would continue to commit violent acts that would be a threat to society. This should be a relatively hard thing to prove—unless you’ve got expert witnesses like psychiatrist James Grigson in your corner. Grigson was nicknamed “Dr. Death” because he waltzed into courtrooms to declare that a convict was a sociopath and would almost certainly kill again, clearing the way for a death sentence. He did so over and over—even in cases where the wrong person was convicted.78
The case of murderer Thomas Barefoot is an example of how Grigson used proofiness to snow juries. In Barefoot’s case, Grigson didn’t even give him the courtesy of a psychiatric examination before testifying, multiple times, that he was 100 percent certain Barefoot would kill again. It didn’t matter whether Barefoot would be in jail or whether he would be free—Grigson asserted with perfect certitude that Barefoot would kill again. It’s a stunning piece of risk mismanagement; there’s no way that Grigson could be so certain. However, Grigson’s cocksure testimony tended to impress juries— and it did in this case too. Barefoot was given the death penalty. The case went up to the Supreme Court.
The American Psychiatric Association wrote a brief that revealed Grigson’s handiwork to be gibberish:
The large body of research in this area indicates that, even under the best of conditions, psychiatric predictions of long-term future dangerousness are wrong in at least two out of every three cases. . . .
To the extent such predictions have any validity, they can only be made on the basis of essentially actuarial data to which psychiatrists, qua psychiatrists, can bring no special interpretative skills. On the other hand, the use of psychiatric testimony on this issue causes serious prejudice to the defendant. By dressing up the actuarial data with an “expert” opinion, the psychiatrist’s testimony is likely to receive undue weight. In addition, it permits the jury to avoid the difficult actuarial questions by seeking refuge in a medical diagnosis that provides a false aura of certainty.
In other words, Grigson’s predictions were worse than worthless because they were dubious opinions dressed up in the clothing of scientific and mathematical certainty. At best, Grigson was guilty of risk mismanagement, as he misrepresented the probability of a convict’s killing again. More likely, though, his predictions were complete nonsense with no rational basis; they were Potemkin numbers.
Three Supreme Court justices were appalled by Grigson’s proofiness:
A layman with access to relevant statistics