A publication of the Archaeological Institute of America
The conviction of a renowned handler raises questions about the use of dogs in archaeology.
Michigan native Sandra Anderson is set to be sentenced on August 24 in Federal Court, Southern Division of the Eastern District of Michigan, for planting evidence at a crime scene and making false statements to authorities. A respected scent-detection dog trainer and handler, Anderson pled guilty earlier this year. Since the charges were filed last fall, the work of Anderson and her dog Eagle has been questioned.
Sandra Anderson rose to prominence in 2000 as a dog trainer and cadaver dog handler specializing in human remains detection. Having helped start a dog training and search company called Canine Solutions, Inc., she later became director of the Great Lakes Search and Rescue of Michigan K-9 Unit. Undertaking searches for nothing more than travel costs, Anderson and Eagle became popular with police departments as an inexpensive tool in their investigations. Sandra and her dog were even asked to Panama and Bosnia to search for graves of victims of political oppression and war crimes. Anderson also visited several archaeological sites and old cemeteries, marking burials or establishing area boundaries. From her work, Sandra Anderson and Eagle gained media attention, appearing in at least one documentary. And in 2000, ARCHAEOLOGY ran an article about Anderson, her dog Eagle, and their supposed ability to detect ancient buried remains through the dog's keen sense of smell.
When Sandra Anderson and Eagle were asked to help on a murder investigation in Oscoda, Michigan, in April 2002, nobody could have foreseen the outcome. After an investigator thought he saw Anderson remove something from her boot where she signaled the find of a small bone, a close eye was kept on her. The next day, when she alleged that Eagle had discovered a piece of carpet in an area previously scoured by investigators, Anderson was arrested on suspicion of planting evidence at the crime scene.
According to a Department of Justice press release, on August 20, 2003, Sandra Anderson was charged with five counts of falsifying and concealing material facts, three counts of obstruction of justice, and two counts of lying to law enforcement officials. The charges relate to seven crime scenes in which Anderson searched in the states of Michigan and Ohio. (Eagle reportedly died of heart disease in November 2003.)
On March 10, 2004, the Department of Justice announced that Sandra Anderson had pleaded guilty to the charges. A press release stated that, "Anderson had repeatedly planted human remains, fibers and items stained with her own blood, which she then represented as evidence." The release also stated that "Anderson made false statements to authorities in an attempt to cover up her wrongdoing."
Under the plea agreement, however, Anderson hopes to be sentenced to 18 to 24 months in jail, along with five years supervised release and a fine to be determined. (Check back with Archaeology.org for a sentencing update in August.)
Since her admission of guilt, those who worked with Sandra Anderson have been left questioning the validity of her results. With convictions hanging in the balance, many criminal court cases that Anderson worked on have had to be re-examined to determine how crucial her testimony was and whether or not her evidence was substantiated. In many of the cases, Anderson's testimony was not critical and could be compared with other evidence to assess its veracity. For her archaeological work, however, testing her claims is not as easy. Unlike a crime scene, proof needed to validate or reject Sandra Anderson's claims lies buried beneath the earth.
One case that is suspect is Anderson and Eagle's September 2002 work on a controversial site in the midst of a land-use dispute between the Prairie Band Potawatomi and Ho-Chunk Nation American-Indians and Wood County officials in central Wisconsin. Called in to find any American Indian graves beyond the known boundaries of the "Indian Bill" and "John Ne-We" cemeteries at Skunk Hill, Anderson marked with flags the supposed locations of numerous burials. After reading about Anderson's visit in local newspapers, archaeologists were skeptical of her findings. Robert Birmingham, then Wisconsin State Archaeologist, calls Anderson's extensive grave markings "ludicrous based on the research we had done." He notes, "that many graves would make it the biggest Native American cemetery in the Upper Midwest." Since Anderson claimed that there were burials outside the original cemeteries, Leslie Eisenberg of the Wisconsin State Archaeologist's office probed the ground and found nothing but a few inches of soil resting on bedrock, too shallow to receive graves.
The only sure way to find out whether Sandra Anderson and Eagle were accurate at Skunk Hill and other sites would be to excavate for remains where she said they were, but, at most of the sites Anderson worked on land-rights issues have prevented this. The problems with Anderson and Eagle's work raises questions about the reliability of all scent detection dogs, but few studies have been conducted to evaluate the abilities of such dogs objectively.
A 1998 study by Debra Komar at the University of Alberta, Edmonton, analyzed and interpreted the effectiveness of eight dog and handler teams at locating simulated animal scattered or scavenged human and animal remains among snow and leaf cover. Tests consisted of "blind searches" or trials in which handlers did not how many items to search for or where they were hidden. The items included dry human and animal bone, and gauze and small articles of clothing soaked in human decomposition fluids and then dried. The study revealed considerable variation in the success rates of the dog-handler teams. The individual dog-handler teams had success rates ranging from 55 to 95% over the trials, and the overall recovery rate for the trials was 81%. Testing dog reliability in cold weather prompted the study, and Komar found that "low ambient temperature and snow depth appear to have no effect on the dog's performance." More important was her conclusion that "results indicate the need for a thorough training program which would expose both dog and handler to a wide range of variables in terms of both body elements and scene terrain."
A 2003 study at the University of Alabama, Tuscaloosa, tested the effectiveness of cadaver dogs (dogs trained to detect old and new remains) and their handlers at finding buried human remains. Keith Jacobi, Alanna Lasseter, and their colleagues, tested the abilities of four dog and handler teams to find ten fresh and skeletal human remains items buried between one and two feet deep. Five separate trials were conducted in warm and humid conditions and videotaped. The trials recorded:
Alerts: dog properly marked the site of remains and the handler recognized it.
Unrecognized alerts: an alert in which the dog located the human remains but the handler did not recognize the signal because it was not the alert the dog was trained to give.
Narrowed areas: dog and handler team identified the area in which remains were situated, but the handlers were unable to specify the location.
False alerts: dog and handler signaled remains where they were not present.
No alert: no signal was given by the dog and handler team that remains were present in the test area.
One possible factor in the variable success rates in the Alabama study could be that in warmer weather, dogs face difficulties such as panting and fatigue. To accurately detect the faint scent of buried human remains cadaver dogs must sniff the air slowly and carefully. "Even with multiple breaks and water availability at all times, dogs panted and thus were limited in their smelling ability" the study noted. The study also concluded that communication between dog and handler was also a major problem. "The dogs are giving signals indicating that human remains are present, but the handler ignores those cues." Jacobi and the research team are compiling a video archive of the trials for handlers to use for training purposes. Handlers can watch the tapes and use them to improve their search methods and communication with their dogs. The study also suggested that to increase reliability and effectiveness there is a "definite need for standardized training for all dog and handler teams."
Despite difficulties revealed by the study, one find particularly intrigued Jacobi. "The fourth team made the most surprising positive alert because they were able to locate one small skeletal cervical vertebra buried 2 ft deep in the large heavily wooded area," he recalls. In the study, the find was "a very dry element of an individual who was skeletonized over 15 to 20 years ago."
The Alberta and Alabama studies, though limited in their scope and number of trials, show both positive and negative results for scent detection dogs. In both studies, dogs were able to locate human remains but the overall reliability of individual dog and handler teams remains in question. In the Alberta study, the recovery percentage for one dog was 55% while another was 95%. The Alabama study showed a much larger number of false alerts and no alerts than correct alerts. Together, the studies show a definite variance in accuracy and reliability among dog and handler teams. Possible factors affecting dog and handler team ability could be weather, soil condition, training, and dog-handler communication. Given such findings and a lack of thorough studies, the Alabama team hopes to study scent detection dogs more extensively in the future.
As many scent detection dogs are home trained, training equality may play a role in the varying reliability of cadaver dog training. Accordingly, the Alabama study suggested that the "availability of appropriate materials and methods through sanctioned dog training facilities might assist handlers in accurately training dogs for discovery of not only buried fresh human remains, but also buried human skeletal remains." Though not applied to human remains detection, the Federal Emergency Management Agency (FEMA) has a certification system in place for dogs and handlers used in search and rescue situations. According to the FEMA website, "each canine/handler team must pass rigorous national certification in urban search and rescue." Handlers are certified by "written and verbal tests regarding search-and-rescue strategies, briefing, and debriefing skills, and canine handling skills." In addition, canine certification "includes proper command control, agility skills, barking alert skills to notify rescuers of a victim and willingness to overcome innate fears of tunnels and wobbly surfaces under the guidance of the handler."
Over the years different fields of scent detection have developed along with associated groups and organizations aimed at advancing the use of scent detection dogs. One of these organizations is the Institute for Canine Forensics (ICF), founded by Adela Morris as a nonprofit organization focused on education and exploration in the field of canine scent detection. According to its mission statement, ICF seeks to "promote and elevate professionalism in the use of specially trained canines for forensic evidence and human remains detection." As part of its mission, ICF has also been exploring the use of dogs in archaeology.
In referring to the Alabama study, Morris states that "dogs are only as good as the training they have received and the only way you know the quality is to test." She cautions against directly applying such studies to dogs used for archaeological work. The dogs that can be used in archaeology, she says, are a new type of trained dog called Historical Human Remains Detection (HHRD) dogs, as Morris refers to them. These dogs are specifically trained to find "old" remains only. "The new generation of specifically trained HHRD dogs in general do not work on fresh blood or any fresh scents," says Morris, adding that puppies are instead "imprinted and given a solid foundation on old bones and teeth" and "later they get experience working in old cemeteries."
The question remains, can dogs be used effectively to find ancient buried remains? While the use of dogs to search for live people, recent human remains, bombs, and drugs, is widely accepted, the idea of dogs specifically searching for ancient remains is relatively new. Accordingly, many agree that such dogs must be tested and held to reliable standards in order to be considered reliable, effective tools in archaeology. "When a dog indicates something buried, it should be field verified with coring, test pitting, or other appropriate means," says University of Colorado, Boulder, archaeologist Payson Sheets.
The Institute for Canine Forensics plans on working more with archaeologists to test and improve the ability of the dogs. "We, as handlers know what a valuable resource dogs can be and we need to keep having finds and prove the value of our dogs," states Morris. The ICF has already worked with Santa Clara University as well as San Jose State University on digs involving students of the schools. In June 2003, with ICF advisory board member and Santa Clara University archaeologist Russ Skowronek, the group used Historical Human Remains Detection dogs to search for unknown burials and boundary lines of the cemetery at Mission Santa Clara. A major test will come for the group when the mission cemetery is excavated in the coming year. "The work we did at the Mission Santa Clara will start up again in the fall and winter when they will start excavating where the dogs have alerted," says Morris.
In reference to the Sandra Anderson case, Morris maintains that, "we cannot let an unfortunate experience put a shadow on the Historical Human Remains Detection dog. With the support of archaeologists we can build a valuable tool to detect historical and ancient burials."
Evan Walker, an anthropology and geography major at the University of Colorado at Boulder, is an intern with ARCHAEOLOGY.