Two hundred and sixty-six people have been exonerated by DNA in the United States after spending years—sometimes decades—in prison for crimes they did not commit. More than half of these people were convicted on the basis of bad forensic evidence.
In some cases, the forensic methods themselves had few or no professional guidelines or objective criteria. In other cases, analysts used scientifically valid tests but fudged or misstated the science in their testimony. Here’s a list of some common areas for mistakes or malice – and men who have suffered as a result.
Blood Type Matching
Before DNA became widely available, forensics analysts relied heavily on serology, or blood typing (above). Eighty percent of the population – known as ‘secretors’- expresses its blood type in body fluids other than blood, such as semen or saliva, and prosecutors often tried to identify rapists by matching the blood type in their semen – a flawed system, since the samples often mixed together, making it difficult to tell which secretions came from the victim and which came from the attacker.
At Gary Alvin Richard’s 1987 rape trial, a supervisor from the Houston Police Department Crime lab told jurors that Richard (bottom) was a non-secretor. This was damning evidence, the supervisor told the jury, because the attacker in the case was a non-secretor; only 20 percent of the population are non-secretors. Twenty years later, investigators re-tested Richard and discovered that he is, in fact, a secretor. He was freed in 2009 after 22 years in prison.
Blood Spatter Analysis
When Warren Horinek’s (pictured above) wife Bonnie was shot at close range in 1995, all the evidence seemed to point to suicide. But Horinek was convicted of murdering her and sentenced to thirty years in prison. At his trial, blood spatter analyst Tom Bevel told jurors that the fine spray of blood on Horinek’s t-shirt were a sure sign that Warren had shot Bonnie. The jury foreman later confirmed that this testimony convinced the jury to convict.
A 2009 report by the National Academies of Science warned that “the opinions of bloodstain pattern analysts are more subjective than scientific” and said the “uncertainties associated with bloodstain pattern analysis are enormous.” That same year, a reconstruction of the crime scene by a more qualified spatter analyst shows that the spatter on Horinek’s t-shirt was most likely created by performing CPR. Horinek’s lawyers have filed a writ of habeas corpus to try to have him released.
In microscopic hair comparison, an analyst compares hair from a suspect with hair found at a crime scene under a microscope (as seen at left above). If the hairs appear similar, or share unusual characteristics, they are often said to “match.” But the matching process is entirely subjective, and whether characteristics are unusual or noteworthy is based only on the analyst’s experience; there is no data on how common various characteristics are. Even if an analyst is cautious and says on the stand that a defendant “cannot be excluded”—that the hairs could be his—“the jury has no idea how many people are included,” says UVA Law Professor Brandon Garrett. “It’s incredibly vague terminology.”
Larry Peterson (top) was convicted of a 1987 rape and murder when a forensic analyst concluded that hairs found on and near the victim were Peterson’s. When the hairs were later tested for DNA, all matched the victim. Peterson was exonerated in 2006.
William Gregory (center) served 7 years of a 70 year sentence for convicted of rape, attempted rape, and burglary before he was exonerated in 2000. The forensic analyst at his trial testified that hairs found at the crime scene were “more than likely” from Gregory due to their “Negroid origin” and unusual characteristics.
In some cases, questions about the accuracy of the science take a backseat to outright fraud: Curtis McCarthy (bottom) was convicted based on hair and fluid evidence, but was released from prison in 2007 when it was revealed that a member of the forensics team willingly tampered with evidence and gave false testimony.
Steven Barnes was convicted of a 1985 rape and murder on the basis of several unvalidated forensics techniques, chief among them an unusual type of fiber analysis. When police couldn’t find any other evidence linking him to the crime, they turned to an imprint made in the dirt on Barnes’s truck. An analyst testified that the imprint and fabric of the victim’s jeans (at right) made a similar pattern. The prosecution then called a manufacturers’ representative to the stand to testify that these jeans were unusual—as few as 200 pairs had been sold in the county that year. Jurors were convinced the imprint must have been made by the victim’s jeans. Barnes served almost 20 years of a 25-year-to-life sentence before he was exonerated by DNA. Here he hugs his sister after being released from prison in 2008.
Bite Mark Matching
In 1991, an Arizona bartender was found dead in the bar where she worked, with bite marks to her breast and neck. During suspect Ray Krone‘s trial, an analyst testified that the impressions of Krone’s teeth matched the bite marks in the victim. Krone (above, left) was sentenced to death, and spent more than a decade in prison before he was exonerated by DNA and released in 2002.
Bite-mark comparison assumes that the alignment of every person’s teeth is unique, and therefore that each of us would leave a different impression when biting skin (right). This assumption has never been proven. This method of forensics has lately fallen out of favor, because bite marks usually include traces of saliva, making it a much more reliable forensic test available: DNA.
In this image, the print on the left—a whorl pattern—was left in blood at a crime scene. The print was then enhanced with dye to make it easier to see (above center). An analyst compared the enhanced print to another print of an identified suspect and found several matching characteristics, marked (at right above) with colored arrows.
It sounds like a slam-dunk match, but the science isn’t so clear. Even the same person touching the same surface with the same finger will leave a slightly different print each time. Once a print is found, scientists have to perform several subjective measures in determining if the print is a match. Stephan Cowans, pictured here in Boston, was convicted of shooting a police officer in 1997. At Cowans’ trial, an expert testified that Cowans’ thumb print matched a print one left by the shooter. Later, after DNA exonerated Cowans, a re-examination revealed that the print did not in fact match Cowans’—and at least one examiner knew it when he testified otherwise. As a result, the Boston Police Department’s latent fingerprint unit was closed, and two of the fingerprint examiners who testified at Cowans’ trial were brought before a grand jury. Neither examiner was indicted. Cowans was found shot dead in his home in 2007.
Cameron Todd Willingham (pictured above with his daughter Amber above) was convicted of homicide when arson investigators determined that the 1991 fire that killed his three children was intentionally set. But his case came under scrutiny from noted fire experts like John Lentini, who said the police relied on outdated science and “mythological indicators,” a term Lentini uses to describe the mixture of facts, falsehoods, and mistakes used by many police investigators. For instance, the investigators argued that the white spots on the floor of his home (above, left) were signs of an accelerant, but Lentini argues those were just patches of floor covered by toys, blankets, or rugs. Willingham’s story was told in depth in the New Yorker after his death, and in 2010, a four-person panel in Texas found that while faulty science may have played a role in the conviction, the investigators involved did not act negligently.
The Detroit Police crime lab was shuttered in 2008 when an audit discovered that 10 percent of its ballistics cases involved major errors such as misidentifications. (Top, prosecutor Kym Worthy testifies in Detroit City Counsel regarding the lab’s shortcomings). The audit was spurred by a 2008 pleaded case in which ballistics reports which supposedly showed that 42 bullets from the crime scene were all fired from the same weapon; later tests revealed that the bullets were from 2 different guns.
Ballistics experts often fire a bullet from a gun they suspect was used in a crime, then try to match the grooves, striations and markings on that bullet to the bullet – or bullet fragment – found at the crime scene (at right are the images of two bullets considered a match to the same gun; at left are the tools used to measure and mark evidence). So far, however, there’s no scientific evidence to validate the claim that no two guns fire the same way. The National Academy of Science agrees that toolmark analysis makes it possible to isolate the manufacturer or model of a gun that fired a particular bullet, but that there is not enough data to say how common various imperfections in the barrel of a gun.
DNA testing is certainly not foolproof–nothing is–but more common than errors in the lab are errors in how results are described to juries. In the case of Gilbert Alejandro, a forensic analyst told the jury that DNA testing matched Alejandro; in fact, the analyst (who was later implicated in dozens of other cases of fraud), had not even performed DNA testing at the time of Alejandro’s trial. Later DNA tests definitively excluded Alejandro. At Josiah Sutton’s trial, a crime lab employee testified that DNA found on the victim was an exact match with Sutton (pictured above); actually, 1 in 16 black men share that particular profile. Both men have since been exonerated.