FROM THE START, A FAULTY SCIENCE
Testimony on bite marks prone to error
By Flynn McRoberts and Steve MillsTribune staff reporters
October 19, 2004
The nation's leading forensic experts held their annual meeting in 1970 at Chicago's Drake Hotel, and all of the old guard was there. Fingerprint experts. Document examiners. Pathologists.
Mingling among them on that late-winter day was a cluster of dentists who shared an interest in a budding discipline. They called it forensic odontology, a decidedly novel application of dentistry--identifying violent criminals based on the bite marks they leave on the bodies of their victims.
But to create their own division within the American Academy of Forensic Sciences and gain the credibility this would bestow, 10 of these forensic odontologists were needed. Only eight were in the room.
The solution: They trolled the meeting rooms of the Drake and recruited a couple of pathologists who also held dental degrees.
With that, a new discipline was born, joining other more commonly known investigative tools such as toxicology and bullet matching.
Since that day in Chicago 34 years ago, bite-mark comparison has become a regular weapon in the forensic arsenal, with odontologists testifying in courtrooms hundreds of times.
They're usually brought in for cases of child abuse, rape and sex murders, where bite marks sometimes are found on victims. But concerns about forensic dentists center less on how often they testify than on how easily judges and jurors accept their opinions as scientific proof.
Courts frequently do so even though there is no accurate way to measure the reliability of bite-mark comparisons, and the method has gained acceptance without benefit of broadly reviewed research and scientific validation, elements that separate true science from guesswork.
The consequences of such subjective testimony are becoming clear. In recent years, new evidence, including DNA, has proved that even a number of the discipline's pioneers have contributed to wrongful arrests and put innocent people behind bars.
In some instances, odontologists can't even agree on the most basic issue--whether a wound is a bite mark at all.
Forensic odontology has come to represent a case study in how easily forensic science's false aura of infallibility can distort the adversarial system of American justice.
In that system, judges and juries are responsible for sifting through often-contradictory evidence. But when experts are allowed to overstate their findings and an unvalidated technique is equated with science, then the system can fail.
"I think bite marks probably ought to be the poster child for bad forensic science," said David Faigman, a professor at the University of California Hastings College of the Law and co-editor of "Modern Scientific Evidence."
It's not simply outsiders or defense attorneys asking fundamental questions. Inside the tight fraternity of odontologists, skeptics are raising concerns about bite-mark comparisons.
"Those comparisons are flawed and based on wishful thinking, as far as being conclusive scientifically," said Dr. Michael Bowers, an odontologist and lawyer who served on the examination and credentialing committee of the American Board of Forensic Odontology, the discipline's leading professional organization.
Bowers angered many of his peers when he co-wrote a controversial study in 2002 that estimated the performance of board-certified odontologists in a workshop exercise. Bowers figured that on average, they falsely identified an innocent person as the biter nearly two-thirds of the time.
Though that figure is hotly debated, even founders of the discipline express deep reservations about bite-mark comparisons.
"They have their place," said Dr. Edward Woolridge, a forensic odontologist from North Carolina. "But I know there's innocent people in jail because of bite-mark testimony."
In one notorious case, such evidence helped send Ray Krone to Death Row for the 1991 murder of a Phoenix cocktail waitress. The former postal worker spent more than three years on Death Row and seven years under a life sentence before DNA testing connected another man to the crime and exonerated Krone.
In many other cases, doubts about such testimony don't necessarily mean a defendant is innocent. But they go to an equally fundamental concern: the courts' ability to interpret and weigh forensic evidence accurately when a person's liberty is at stake.
Defenders of odontology contend that it can be valuable. In a case of child abuse, for instance, there usually are a limited number of people with access to the child. If bite-mark evidence can eliminate all but one of the people under suspicion, it can provide circumstantial evidence against the remaining suspect.
The chief goal of odontologists is "to testify as to the truth represented by the evidence, wherever the evidence leads," said Dr. Robert Barsley, an LSU dental professor and the odontology representative on the American Academy of Forensic Sciences board. "... It can be a useful forensic tool."
Still, too many forensic dentists are willing to testify beyond the limits of the evidence at hand, according to their own colleagues.
"Oftentimes these are horrific cases. So you get a lot of pressure by the [authorities] who employ odontologists to come down on one side of the fence--`Yes, it is this guy' or `No, it isn't,'" said Dr. Iain Pretty, a British odontologist who has challenged Bowers' findings but shares some of his concerns.
"There's a lot of pressure, and conclusions are overstated."
Such concerns have been borne out in a host of examples where DNA later exonerated people imprisoned based on bite-mark testimony, or raised questions about their convictions.
In the case of Kennedy Brewer, a Mississippi dentist testified for the prosecution that 19 marks found on the body of his girlfriend's 3-year-old daughter were human bites made by Brewer.
A defense expert, though, concluded they weren't human bite marks at all. The expert testified the marks actually were insect bites and the result of the girl's tiny body being left undiscovered in a creek for two days in May 1992.
Brewer was convicted and sent to Death Row. Years later, though, DNA evidence from semen found inside the girl excluded him and identified two unknown genetic profiles, prompting a new trial.
Brewer remains behind bars. And prosecutor Forrest Allgood said he would "absolutely" use the bite-mark testimony again, though it was given by Dr. Michael West, a controversial odontologist who once was suspended by the odontology board.
West is ready to repeat his testimony. "I don't know who killed this girl, but Kennedy Brewer's teeth were on her body...," he said. "It's not rocket science."
Birth of a discipline
In the early years of odontology, meetings were so small that conferences amounted to a few colleagues lecturing to one another. "We'd applaud each other like we were a big audience," Woolridge said. "It sounds kind of corny now."
But these informal sessions had a distinct purpose. By creating their own section within the forensic sciences academy, the odontologists helped establish the credentials they needed to go into court and testify as experts.
To this day, most of their forensic work involves identifying the remains of unknown people by matching teeth to dental records. Comparing bite-marks was a natural extension of this kind of work and allowed dentists who spend their workdays, as some say, "drilling, filling and billing" to pursue their interest in helping solve crimes.
It also can be lucrative. Forensic dentists' fees can range from $100 to $350 an hour, and a typical bite-mark case that goes to trial can earn them anywhere from thousands to tens of thousands of dollars. But those cases are rare, Barsley said.
One of the field's most prominent members is Dr. Richard Souviron, the chief forensic odontologist for the Miami-Dade County medical examiner's office. With his black cowboy boots and lightheartedly ghoulish sense of humor, Souviron is a favorite on the seminar circuit for cops learning to be homicide investigators.
He made his reputation--and, in the public mind-set, the reputation of bite-mark evidence--in a 1979 trial for the murder of two Florida State University students.
Theodore "Ted" Bundy was the accused. Investigators had found a pantyhose mask with hairs that state experts said were "consistent with Bundy's hair." But prosecutor Larry Simpson said the primary piece of physical evidence tying him to the slayings of Chi Omega sorority sisters Margaret Bowman and Lisa Levy was a bite mark on the buttocks of Levy.
Prosecutors wanted to introduce the evidence but were concerned about using the nascent technique in such a high-profile case. They knew it would be challenged.
To boost the discipline's credibility at the trial--and to meet the judicial test that such a novel technique was generally accepted in the scientific community--Simpson told Souviron they needed an expert from the West and the North to corroborate him.
So Souviron called two of his colleagues who had helped form the discipline nine years earlier: Dr. Lowell Levine, from New York, and Dr. Norman Sperber, who worked in San Diego.
The state put Souviron on the stand. "Can you tell me within a reasonable degree of dental certainty if those teeth made the marks?" Simpson asked, referring to molds of Bundy's teeth.
"Yes, sir," Souviron replied. "They made the marks."
The defense brought in a Maryland orthodontist who disputed the prosecution's contention that Bundy's teeth alignment was unique. He testified that "the dental pattern is one I'd expect to find in 20 percent of the population of male Caucasians."
Bundy was convicted, thus adding legitimacy to the upstart discipline.
Twenty-five years later, Souviron acknowledges that the Bundy trial left a problematic legacy: "It catapulted bite-mark evidence to the point where [many] were saying, `A bite mark is as good as a fingerprint.'"
In recent years, Souviron has preached caution, not certainty, when it comes to such evidence. At least in theory, so does the American Board of Forensic Odontology.
Board guidelines adopted in 1995 discourage members from using the term "match" because it "will likely be interpreted by juries as tantamount to specific perpetrator identification."
But the more nuanced phrases the board endorses for the clearest cases, such as "reasonable medical certainty," are defined as essentially the same thing--"no reasonable or practical possibility that someone else did it." Leading odontologists concede that such distinctions are lost on many jurors.
This spring, Souviron stood before a room full of police officers in the Miami-Dade medical examiner's office. They were taking a weeklong death investigation workshop, and Souviron had a couple of hours to introduce them to his specialty.
"Have you ever heard that bite marks are as good as a fingerprint?" he asked the group. A dozen or so hands shot up in the room of three dozen officers.
"You've heard it, right?" Souviron said. "Well, don't use it, because it's wrong."
The reasons are fairly simple. No research has been done to confirm that people's bite marks are unique. The alignment of people's teeth changes over time. And human skin shifts when it's bitten, and often just leaves bruising.
"You've got to be real careful with this kind of evidence," Souviron told the officers before leading them down a hall to the morgue to demonstrate his techniques on two corpses. It's "one more little nail in the coffin, and that's what you guys need to think of with bite marks."
On witness stands across the country, however, caution is often absent. The nation's certified forensic dentists, who number about 100, have played a key role in convicting hundreds of defendants.
In February, Souviron testified in the trial of Ronnie Keith Williams, who had just finished a sentence for rape and murder when he was charged in the 1993 murder of a pregnant teen.
Using the careful phrasing approved by the forensic odontologists' board, Souviron said, "In my opinion, within reasonable dental certainty, the teeth of Ronnie Williams left that bite mark."
But after he was challenged by the defense over his methodology, Souviron gave the kind of definitive testimony that he often cautions against.
"That bite mark was left by Mr. Williams," he said, "period."
Souviron, interviewed later, acknowledged that he had stepped over the line at Williams' trial, which ended with a guilty verdict and death sentence. "Had I been challenged on it, I would have said I misspoke. I should have said again, to a reasonable degree of dental certainty, period."
A "science-based art"
The tools of bite-mark analysis are strikingly basic. There's a putty-like material to preserve impressions of the mark; molds to reproduce the teeth of suspects; photographs and plastic transparencies to compare the molds with the wound.
One of the discipline's significant advances, introduced in the early 1990s, is an L-shaped ruler, meant to assist analysts in photographing and comparing the marks.
A few years ago, image-enhancing software--purporting to make the features of a bite mark more visible--was introduced in several crime labs.
But neither computer technology nor low-tech tools address the more fundamental question: whether bite-mark comparisons are scientifically sound.
"This is truly an opinion," said Barsley, the odontology representative on the forensics academy board. "It's unlike DNA, where there's a mathematically predictable science behind that."
Barsley calls odontology a "science-based art" and notes that while a comparison of a bite mark and a wound can tell you they are "very similar, it can't tell you they're identical."
Most courts still allow forensic dentists to practice their work with few boundaries. But, in a rare move this year, a judge in Michigan strictly limited what the odontologists could assert, citing a ruling from that state's Supreme Court.
"You will not be discussing probabilities ... or degrees of certainty," the judge said, restricting them to saying whether they could exclude or not exclude certain suspects.
In many cases, superior evidence such as DNA has shown that bite-mark testimony may be wrong.
Dr. John Kenney, a Park Ridge pediatric dentist and odontologist, conducted a bite-mark analysis that proved crucial in the case against Harold Hill and Dan Young, who were convicted of the 1990 murder and sexual assault of a Chicago woman.
Police initially said Hill, Young and a third man all confessed to committing the crime together. But investigators later learned the third suspect's confession was false; he couldn't have done it because he was in jail at the time of the attack, suggesting Hill and Young's confessions may have been false as well.
Prosecutors then argued that Kenney's testimony corroborated the confessions by linking a purported bite mark on the victim to Young and a hickey on her to Hill.
But a recent reanalysis of the bite-mark evidence commissioned by the defense casts doubt on Kenney's testimony. Two other odontologists, including Bowers, examined his work and came to a different conclusion--that there was no match.
More important, perhaps, new DNA tests on scrapings from beneath the victim's fingernails excluded Hill and Young. Cook County prosecutors have been reinvestigating the case for more than a year and hope to complete the inquiry soon.
Kenney said he remains confident in his analysis of the Hill and Young case but expressed concern that he might have played a role in a wrongful conviction. "No one's going to be more upset than me if guys spent time in jail based in part on what I did...," he said. "I want to see justice served."
The inherent subjectivity in odontology, he said, leaves room for misleading testimony. "You get pushed a little bit by prosecutors, and sometimes you say OK to get them to shut up," Kenney said, acknowledging that he now wishes he had tempered his testimony in the Hill and Young case. "I allowed myself to be pushed."
A spokesman for the Cook County state's attorney's office said "prosecutors did not coerce this man into testifying to anything but the truth."
In another case involving DNA, Dale Morris Jr. of Florida was arrested in 1997 based on a bite mark that two forensic dentists used to connect him to the rape, torture and murder of a 9-year-old girl.
Morris was jailed for four months until DNA tests cleared him and authorities set him free. Two other men, including an uncle of the girl, are scheduled to go on trial in her murder next year.
The odontologist who gave investigators a second opinion that led police to arrest Morris? Souviron.
Key expert takes a hit
One of the few full-time forensic odontologists in the nation, Lowell Levine has a resume that reads like a sampling of the 20th Century's iconic events.
He helped investigate the assassinations of President John Kennedy and Martin Luther King Jr. He traveled with a team to Russia to examine the remains of Czar Nicholas II and his family. He was among the experts who identified the remains of Josef Mengele, the Nazi doctor who selected incoming Jews for labor or extermination at Auschwitz.
When the Sam Sheppard case, which inspired the TV program "The Fugitive," was reopened, he was called in on that too.
Working from his home office in Albany, N.Y., Levine divides his time between work for the New York State Police and consulting for both defense attorneys and law enforcement. His colleagues say they hold him in high regard as one of the field's top practitioners.
So it was no surprise in December 1998 when a team of Massachusetts State Police officers turned to Levine in hopes of solving the gruesome murder of Irene Kennedy.
The 75-year-old grandmother had been beaten and stabbed two dozen times while on a morning stroll with her husband in a park outside Boston. The killer, who attacked Kennedy when she and her husband briefly took separate paths, left a bite mark on her breast.
The investigators drove from Boston to Levine's office. Explaining the circumstances of the murder, they asked him to compare photos of the bite mark on Kennedy's body with a copy of a mold made from the teeth of a suspect, Edmund Burke.
Then in his mid-40s, Burke was a reclusive handyman who lived with his mother and dozens of stray cats in a small, ramshackle home about a quarter-mile from the park where Kennedy was killed. He became a suspect when police dogs led officers to his home.
Levine declined to discuss the case, citing a lawsuit Burke has filed naming him.
But in a sworn deposition taken in the lawsuit, Levine testified that after studying the materials in his office, he told the waiting officers he could not exclude Burke but would need additional information for a more definite opinion.
Three days later, Levine went to Boston to examine more evidence, asking police to provide him with enhanced photos of the bite wound. They did, and that, Levine said, was enough.
In his deposition, Levine said he concluded "to a reasonable scientific certainty" that Burke had left the bite on Kennedy's breast.
Police searched Burke's home, and arrested and jailed him. The county prosecutors called the bite mark the "most compelling evidence" in the case.
Less than six weeks later, though, officials had to admit they were wrong. DNA taken from saliva recovered on the bite mark was analyzed. A genetic profile was obtained, and prosecutors said it was not Burke's. He was set free.
Levine insisted in the January 2003 deposition that he had been correct when he linked the bite mark to Burke, although he also hedged a bit, saying he had never made a definitive "match."
Under questioning by a lawyer for Burke, who sued the police and Levine after he was cleared, Levine stood by his bite-mark analysis.
"Do you think he bit her breasts?" attorney Robert Sinsheimer, who represents Burke, asked Levine in the deposition.
"I think with a high degree of probability he did," Levine said. He offered possible explanations for why the DNA did not match Burke, including that police who had handled the crime scene contaminated the DNA.
He also noted that another prominent forensic odontologist, Dr. Ira Titunik of New York, had examined the evidence and concurred in his opinion. Titunik confirmed that he had informally examined the evidence and agreed with Levine.
But then Levine's analysis took another hit. In June 2003, some five months after Levine testified under oath and held fast to his bite-mark analysis, police announced they had made another arrest in Irene Kennedy's murder.
The genetic profile derived from the bite mark, the police said, had been entered into a database. It hit on a convicted murderer.
William Keating, the district attorney where the crime occurred, said there was "no question" in his mind that Burke was innocent of Kennedy's murder.
Roadblocks to reform
In the face of so many embarrassing mistakes, a small number of self-styled reformers inside the world of odontology have sought in recent years to determine just how accurate their discipline is.
But they've faced fierce resistance. Bowers, the California odontologist, was roundly criticized when he analyzed something innocuously called Workshop No. 4 at the 1999 meeting of the American Academy of Forensic Sciences.
Bowers said the results showed that on average, the examiners falsely identified an innocent person as the biter 63.5 percent of the time. "It's standing on a foundation of assumption," he said of the field.
His peers questioned his statistical analysis and noted that the workshop was never intended to be a proficiency test. But the furor underscored that odontologists have never agreed on any exam that would gauge the accuracy of bite-mark comparison.
Exacerbating the problem are often-weak standards. In Britain, where there are relatively few odontologists, a dentist must have examined 20 bite-mark cases before being certified as a forensic odontologist. In the U.S., the American Board of Forensic Odontology requirements for certification include two bite-mark cases.
This month, as part of its annual meeting, the executive committee of the ABFO discussed changes in its recertification process to meet new standards imposed by the Forensic Specialties Accreditation Board. The committee agreed to recommend adding a written exam every five years that will cover all aspects of forensic odontology, from dental identification to bite-mark evidence.
"Like any organization, this is evolving," said Dr. John Lewis Jr., president of the ABFO. "And we saw it as a very good step, further strengthening our position as a certification body."
The exam will not, however, involve actual bite-mark comparisons.
A Tribune examination of criminal cases in which bite marks played a key role illustrates why credibility is a concern. The survey involved 154 cases, mostly murders and rapes, that reached appeals courts in state and federal jurisdictions around the country--just a sampling of the hundreds of times odontologists have testified.
Though not comprehensive, the survey found a disturbing pattern. In more than a quarter of those cases, the prosecution and defense offered forensic dentists who gave diametrically opposed opinions.
These were not simply another example of the competing experts common to many courtroom faceoffs. In some cases, the difference was so great that an odontologist said what one colleague considered human bite marks were not bite marks at all.
Such divergent opinions were at the center of the murder trial of Carol Ege.
In the years since someone murdered Cindy Thompson--stabbing and disemboweling a woman who was seven months pregnant--no one had been charged in the 1984 slaying. Then authorities took a fresh look at an old autopsy photograph of her body, which was found inside a wood-frame house in the gritty Detroit suburb of Pontiac.
To the original medical examiner, the mark on Thompson's left cheek was simply a pooling of blood left as she lay dying. But in Ege's 1993 trial, jurors heard the bold proclamations of Dr. Allan Warnick, then the chief forensic odontologist for Wayne County, Mich.
Thompson's attacker had bitten her, Warnick told them. Not only that, he said, the bite mark on her cheek was so unique that he could pinpoint who made it: Ege.
"Let's say you have the Detroit metropolitan area--three, three-and-a-half million people," prosecutor Gregory Townsend asked Warnick. "Would anybody else within that kind of number match like she did?"
"No," Warnick replied. "In my expert opinion, nobody else would match up."
But after Ege's conviction, Warnick's ability as an expert came under serious doubt. In two murder cases, authorities relied on his opinion to arrest suspects, only to dismiss the cases after other forensic dentists disputed his comparisons.
"Therefore the office of the Wayne County prosecuting attorney will not approve warrants where the main evidence as to the identity of a potential defendant is the opinion of Dr. Warnick that he/she is the source of the bite marks," wrote Richard Padzieski, the office's chief of operations at the time.
Warnick directed requests for an interview to his attorney, Bruce Leitman, who said his client declined to comment. "Dr. Warnick's done nothing wrong," Leitman said.
In the Ege case, her defense presented a pathologist and a forensic dentist, both of whom testified that the mark on Thompson's cheek was pooled blood.
The dentist, Dr. Irvin Sopher, said that even if it were a bite mark, the pattern did not match Ege's teeth alignment.
Prosecutors offered a host of witnesses who testified to Ege's threats against Thompson, her rival for the affections of a mutual boyfriend. They say the bite mark was not a pivotal piece of evidence.
"It was just a drop in the bucket," said Kathryn Barnes, an assistant prosecutor for Oakland County who argued the state's case during Ege's second appeal of her conviction. "There was a mountain of evidence against Carol Ege."
But that wasn't how some of the jurors saw it. They noted that authorities knew of the threats against Thompson and other evidence nearly a decade before, yet chose not to charge Ege.
They were more impressed by Warnick's certitude and the evidence they were allowed to bring into the jury room, comparing photographs of Thompson's cheek with transparencies of the suspects' bite marks. They were convinced that only Ege's matched.
"The bite mark was the most convincing evidence because really it was the only thing that placed her at the crime...," one juror recalled in an interview. "That was the clincher."
"I remember thinking," said another juror, Nichelle Williams, "they got her now."
A Michigan appeals court called the case "troubling." It noted that while "the crime is horrific," there were "others who are logical suspects" and that "no physical evidence links defendant to the crime except" the bite-mark testimony.
The appellate court, though, upheld her conviction. Her case is now before a federal judge.
Serving a life sentence, Ege spends most of her days working in the prison library. If she isn't there, she is sitting in her cell watching her 12-inch TV. Her favorite show: "CSI: Crime Scene Investigation."
Her attorney, Carole Stanyar, is not a fan. She has a problem with the genre--namely, as with all crime dramas, the tidy separation of the guilty from the innocent at the conclusion of each episode.
"It's this false sense of certainty," Stanyar said. "That's what Warnick did. He gave [jurors] a false sense of certainty: You can feel comfortable convicting this woman."