David Grann responds to an op-ed by John Jackson, the prosecutor in the Willingham case:
Jackson...claims in his article that Willingham was offered a polygraph and refused. I do not know if this is true, though it may be. After Willingham was charged with murder, he stopped coöperating with authorities. (On death row, Willingham wrote to several legal organizations asking them if they could give him a polygraph so that he could prove his innocence.) But even if he refused to take a polygraph after he was arrested, polygraphs are notoriously unreliable, and are not admissible in a court of law. (I highly recommend Margaret Talbot’s piece on this subject, “Duped,” which appeared in The New Yorker, in 2007.) As a result, defense attorneys routinely do not let their clients take polygraphs. Ernest Willis, who I discuss in my piece, was also convicted of committing arson, in a case that was eerily similar to Willingham’s. He had taken a polygraph, and the results were interpreted by police and the prosecutor as a sign that he was guilty. Evidence later emerged, however, that he had not set the fire, and he was exonerated and released, after seventeen years on death row. The idea that a lie-detector test (or the refusal to take one) could be considered evidence cuts to the core of the problems in the Willingham case: a reliance on unreliable and unsound scientific techniques.
Jonah Lehrer discussed Grann's article and just world hypothesis last week.