Courts & Crime

Report unveils junk science, tainted testimonies at N.C.'s SBI

On July 16, 1993, SBI Agent Mark Isley hauled Floyd Brown to jail. The charge: beating a retired schoolteacher to death.

The proof: a six-page confession that Isley said Brown uttered word for word in a single telling, one elaborate detail stacked on another.

Then and now, Brown can't get past the letter K when reciting the alphabet. He is a 46-year-old man with the mind of a 7-year-old boy, his IQ hovering at about half that of a person with average intelligence.

"The confession is a work of fiction," said Mike Klinkosum, a Raleigh attorney who represented Brown until his release in 2007. "It's that simple."

Doctors employed by the state were skeptical about Brown's confession as early as 1993, and they sounded alarms in court in 2005. In court documents and arguments, those warnings became more convincing as years passed, reaching Attorney General Roy Cooper's most trusted advisers. The Charlotte Observer and national media pressed Cooper for answers in 2007. But for four years, SBI leadership and the Attorney General's Office failed to investigate Isley's work.

Cooper, a Democrat, didn't act until 2009, in the face of a lawsuit that will likely cost taxpayers and their insurers millions of dollars. He ordered a special review of the Brown case, but the agency has refused to provide any conclusions or results.

The Brown case stands out, but it is hardly the State Bureau of Investigation's only troublesome work. SBI agents have cut corners, bullied the vulnerable and twisted reports and court testimony when the truth threatened to undermine their cases, a News & Observer investigation of the SBI's work, policies and practices reveals.

The SBI sends agents to help local law enforcement agencies solve their most complicated cases. But agents sometimes lock on to the wrong suspect and stick to their story for years, as they did with Alan Gell, a Bertie County man who collected $3.9 million after an SBI agent ignored evidence that pointed to his innocence in a 1995 murder.

At the bureau's crime lab, agents charged with using science to solve crimes have hidden test results or withheld notes that suggested the opposite of findings presented to the courts, as they did in the case of Greg Taylor. Taylor was exonerated of a 1991 murder in February. Firearms and blood analysts have stretched the boundaries of science and aligned themselves so fully with police and prosecutors that the examiners manipulated evidence to fit their theories.

Many agents don't cheat. But even those following the rules work within policies, practices and state laws that bias the agency and its scientists toward the side of prosecutors and away from basic fairness.

To read the complete article, visit

Related stories from McClatchy DC