The girl in the photo had brilliant blue eyes. So brilliant, in fact, that there was little doubt on the parts of investigators that the body found murdered and stuffed into a well couldn’t have been their victim: The recovered remains had eyes as brown as a sparrow, and the missing girl didn’t wear contacts. But, in a twist worthy of the movies, both young women were the same, a fact that was proved conclusively by matching dental x-rays.
It wasn’t a case of photo retouching. Her family vouched for the accuracy of the picture. Instead, the mystery of the blue-eyed missing person who surfaced as a set of brown-eyed human remains translated into a case where blunt head trauma resulted in hemorrhaging in the vitreous humor, which combined with slight decomposition, effectively changed the eye color. Says Deputy Coroner Investigator David Van Norman of the San Bernardino County (California) Sheriff’s Department, who worked that case, “I never looked at eye color the same way after that.”
Eye color, like many descriptors used to portray suspects, witnesses and missing persons, may seem like a no-brainer when it comes to checking the box, but it – along with a disturbing amount of other “factual” information – can be very subjective. Van Norman points to a Jane Doe his office recovered whose eyes, under post mortem examination, were described by different people as “gray, brown and green.” Says Van Norman, “Estimating eye color is far too subjective to be used to rule out a match.”
Subjectivity can affect identification
Subjectivity certainly colors the way we view certain things. Consider, for example, how different witnesses to the same crime identify the identical perpetrator. A classic classroom exercise is having someone pop into the room, commit a series of odd acts and then leave as suddenly as they appeared, and later asking the class to describe the individual. This almost always results in widely disparate eyewitness descriptions. Age, dress, height, weight, hair color and even the acts committed vary based on how the individual witness perceived them. Van Norman says that anything that is subjective can lead to misidentification of human remains. Height is a common area where mistakes are often made.
“Men tend to overstate their height, while women tend to underestimate it,” says Van Norman. Height and weight on most official documents; even the information on driver’s licenses in most states are provided by the person applying for the license. A man who claims to be six feet tall may be anywhere from 5’9” to 6”, while a woman who routinely says she is 5’4”, might be 5’3 or 5’5”. In addition, people tend to lose height as they age. A person who is correctly measured at 5’6” at the age of 40 could be inches shorter in his or her 70s and beyond. Additionally, says Van Norman, obtaining accurate measurements of skeletal remains is both tricky and unreliable.
Because a fully fleshed human skeleton will have skin, muscle, hair and other factors that add to a person’s height, skeletal remains can indicate a height range, but an estimate based on forensic examination should be considered exactly that: an estimate. Van Norman cautions investigators tasked with identifying recovered remains to be careful when assigning height as a static qualifier.
“We have a big stick with a ruler (affixed to it) that we use as a measure and I’ve used that stick to measure bodies hundreds of times. I’ve been wrong lots and lots of times. The reason is that I start at the heel and run that up to the head, and then I find that the guy was two inches taller or two inches shorter. When people are dead and lying on a table they are not the same stature as they are while standing,” he says.
Stature, he says, is an estimate that works for an officer who is looking for a suspect that just robbed a store because it gives the officer a general idea of what to look for. But an eyeball estimate of the person’s height is not so good for identifying a missing person. Van Norman says that relativity also plays a role in identification. A very short person may perceive someone of average height as much taller, with the reverse is true of a tall person. The role of perspective is inarguable when identifying remains.
Pinpointing age
Skeletal remains in general provide difficult analyses, even when the entire skeleton is recovered. In the desert, where Van Norman often works, skeletal remains many times are disarticulated, but even in cases where the intact entire skeleton is recovered, pinpointing age can be difficult.
“(Forensic anthropologists) look at markers on the bones and estimate age and, frankly, the best we can come up with is a range of five to 10 years. Quite often the range can be much greater – as much as 60 years,” he says.
That’s because once individuals reach adulthood, their skeletons change very little, except in response to trauma. According to Doug Lyle, MD, one event that can be detected on skeletal remains includes childbirth, indicating that the remains are those of an adult female, but even girls as young as 12 have been known to give birth.
What about age entries into NCIC or other databases? To put it bluntly, whenever a human is involved, the door for error cracks open. Ages can and have been entered incorrectly, both in the missing persons report and in the paperwork revolving around recovered remains. In one case where information in connection to the remains of a young man was entered into a national database, the age of the recovered remains was estimated to be about seven years older than his actual age. More importantly, the date of disappearance was given as several days after the body was recovered. Clearly the two could not be the same individual, but they were: One of the dates had been entered incorrectly. If someone had not investigated further, the case would still be unresolved.
Other variants
Hair color is another variant that, like eye color, can truly be in the “eyes of the beholder.” Unless the individual has roots showing natural hair color or body hair that contradicts hair color on the head (for instance, dark pubic hair with blond head hair), it is extremely difficult to determine even the hair color range based on visuals alone.
Becky Castillo, a coroner volunteer also at the SBCSD, encourages officers to both involve the families who make the reports and double-check their information.
“There are going to be discrepancies on weight, there are going to be discrepancies in hair color – we don’t personally like to rule out based on those,” says Castillo.
Even visual identification can be dicey. In 2006, two college students were among the dead and injured involved in a terrible vehicular accident in Indiana. The pair of young women involved in the crash had similar coloring and size: One died immediately and one lingered in a coma. The dead girl’s family identified and buried their daughter, while the family of the coma victim sat vigil at her bedside for five weeks. Then one day the girl began to speak as she emerged from her coma, and the family became suspicious that she might not be their daughter. A comparison of dental records confirmed their concerns: Their daughter had died in the accident and had already been buried. The wrong family had been at her bedside.
Another area fraught with a disproportionate amount of official discrepancies involves date of last contact. Castillo remembers a case that remained unsolved for months as the result of an incorrect date.
“We had a case with a Jane Doe who was killed running across a freeway. She was estimated to be between 15 and 20, had no identification on her and was healthy. There was no reason that she shouldn’t have been identified in a timely manner,” says Castillo.
The young woman died in September of 2005, but by December of that same year, she still had not been identified. The SBCSD worked the case for 19 months before finally catching a break. The mother saw a composite photo on the Internet and finally contacted the sheriff’s department. However, according to information provided to NCIC, the last known date of contact with the missing daughter occurred in March of 2006. When the mother was asked about the date discrepancy, she said the NCIC data was wrong: Her last contact with her daughter had been in September of 2005 – the same month she died.
“The girl was identified through dental records; what had happened was that the mom had been trying to report her daughter missing and (the originating agency) wouldn’t take a report. They finally (took a report) in March of 2006 and they entered that as the date of last contact,” says Castillo.
Both Van Norman and Lyle say that even the presence or absence of medical devices, such as metal plates, screws or pacemakers, should not be used as proof positive for identification.
“Braces, rods, screws – they remain permanent much of the time, but sometimes they’re used to stabilize the bone, then they open them up and unscrew the plate – that does happen,” says Lyle.
As for obtaining an accurate medical history, he says many people hold onto the things doctors say to them, even in passing.
“The doctor says you’ve got a bad valve and may need it replaced and they walk out of the doctor’s office and it becomes family lore, when all it is really is an innocent heart murmur. I see this all the time,” the cardiologist says.
Proper identifiers
Anything that is subject to human error can be misinterpreted. Van Norman says he knows one out-of-state agency that continues to use the date the person is reported missing also as the date of the incident. That means many of the individuals reported missing by that agency have disappearance dates wrong. And, while tattoos certainly aid identification, like implanted medical devices they can also be removed or altered, and aren’t visible on skeletons or corpses in advanced states of decomposition. As both Van Norman and Castillo said, anecdotal information concerning features and dates may be erroneous, but some things never lie.
The big three identifiers, says both Van Norman and Lyle are solidly based in science: Fingerprints, dental x-rays and DNA.
Some states require that licensed drivers have fingerprints on record with that state’s department of motor vehicles. Van Norman says that in addition to service members, individuals who have been processed through the criminal justice system and police, many employers now require fingerprinting. Simply because the individual’s prints aren’t in a national or state database doesn’t mean they don’t exist somewhere else. Van Norman also points out that nearly everyone has visited a dentist at least once in his or her life. He suggests digging deeper for those records, even when the family is positive there are none.
The third method of positively identifying remains is dependent on the presence of specific conditions: There must be source material at both ends. In cases where remains are very old and degraded, working samples can be difficult to obtain, but if they are obtainable, they can move mountains.
In conclusion, Van Norman says that computers are, “Wicked fast but also remarkably dumb – they compare data, but they can’t do our thinking for us.” He believes comparing unidentified remains versus reported missing via computer-generated data, while helpful to some degree, is fallible and the evidence supports that conclusion. Says the veteran coroner’s investigator, “Concentrate on submitting forensic evidence – that’s where you stand the greatest chance of success.”
