CSI boffins: You can't ID crims from bitemarks on victims
Corpse-chomp research discredits gnasherprinting
Topflight CSI boffins have cast doubt on the apparently "commonly held belief" in forensics that criminals can be positively identified from the bite marks they leave on their victims.
"Bitemark identification is not as reliable as DNA identification," explains the study's lead author, Prof Raymond G Miller of the University of Buffalo.
"With DNA, the probability of an individual not matching another can be calculated," he says. "In bitemark analysis, there have been few studies that looked at how many people's teeth could have made the bite."
Miller and his colleagues teamed up with Robert Dorion, author of Bitemark Evidence: A Color Atlas, which is apparently "the only comprehensive textbook on the subject of bitemarks". The boffins embarked on a probing analysis of the subject.
According to UBuff:
The study investigated three main questions: is it possible to determine biter identity among people with similarly aligned teeth; is it possible to determine how many individuals from a larger sample might also be considered as the biter; and, if there is bite pattern distortion, is it enough to rule out a specific biter while still including a non-biter?
These knotty issues were investigated by a complex procedure involving a hundred sets of model teeth made of stone, which were used to make bite marks in skin taken from dead human bodies. (The UBuff report notes regretfully that "current human subject restrictions limit experimentation on living subjects".)
The result? "Bitemark evidence should be approached with caution", apparently. There is more from UBuff here. Our moderating staff look forward keenly to a volley of comments along the lines of "there's a subject you can really get your teeth into". ®
@ Joe M
Firstly, thanks for (some) well-considered, constructive observations.
As a profession, we are well aware of the lack of quality research into bitemarks, most importantly the use of a gold-standard that is acceptable both in terms of diagnostic research and is also forensically relevant, but several studies are currently under way to address this as far as possible.
One current PhD thesis is analysing the prevalence of individual dental features (rotated incisors, etc) within the population as a whole and is producing very useful data, but this is unlikely to yield useable results for another 2 years. Other studies have shown that the biting surfaces of the teeth do appear to produce a pattern unique to an individual, but these are based on analysis of models of teeth rather than bitemarks, and typically have involved smaller sample sizes (<100).
The problem in relating the teeth to the mark lies in the fact that a bite is, by its very nature, dynamic. The teeth and jaws move in the action of biting, the biter and victim move relative to one another (unless the victim is restrained or unconscious), the shape of the bitten tissue alters depending on compression from the bite itself, muscle movement below the skin and the presence of the skeleton relative to the skin surface. These can all produce distortion within the injury, and we are trained to assess this in tandem with the physical evidence and statements provided in order to pick up on any discrepancies. Exclusion is a big part of the process.
As far as research goes, we are unfortunately limited by ethical regulations - it's bloody difficult to gain approval to conduct bite tests just on cadaveric animal tissue, and live-subject research is an absolute no-no!
Even where studies have been conducted on cadaveric tissue, the data cannot necessarily be directly translated to the response of living tissue, as tissue reactions, healing, etc all impact on the appearance over time.
We accept the limitations of the analysis and comparison, and do not examine cases in a frivolous manner. Our duty in such cases is not to convict - it is to provide impartial, unbiased expert opinion to the court (at least, that is how we do it in the UK). I have seen plenty of cases where there is insufficient evidence to even support the allegation of a bite, never mind a comparison with a suspect, and have no qualms about stating such in my reports.
The defence always have the opportunity to call their own expert to assess the case, which helps keep everyone up to date with current research and provides scope for an alternative opinion to be presented.
I take no issue with anyone who wishes to query the methodologies or basis for the work we do, it is welcomed as it maintains an impetus for research and improvement. Suggestions for conducting high-standard research will be welcomed gladly, as few of us are based in academia and thus have little scope to undertake research for ourselves. We are heavily reliant on our research colleagues in that regard.
the example of The Wire
This reminds me of The Wire - season 5. The "homeless serial murderer" which McNulty embellished further by creating bite marks on the victim. The FBI made a good point though, there was no trace of saliva so they concluded that the biting was staged... my point here - unless you're acting like McNulty and using fake dentures to do the biting, isn't the biter always going to leave at least a bit of saliva and therefore traces of DNA?
So perhaps the moral of the story is, if you get bitten, wait until you have a swab test taken before you wash and clean the skin.
Mr Odontologist, how come you missed two vital pieces of information in your description of the use of forensic bite-analysis?
1. There are no double blind studies as far as I can find to show that this is an evidence based, scientific technology rather than just you and your colleagues’ opinions (after a big night out... perhaps).
2. Even if matches were shown to be within acceptable rates of error, it would still be useful only for eliminating suspects, not convicting them. (Say the right incisor imprint is missing and the suspect chews with a good set of 32. Shucks!)
The idea that you can identify a person from a bite mark is probably just as ludicrous as it sounds (with the above given exceptions). This probably explains the lack of double blind, statistically significant studies! No pseudo science ever has them.
I guess at this late stage nothing about “forensic science” surprises me. Whenever my brother and I, (with a collective 70 years of scientific research and technology between us), feel like a good laugh we just look up the latest forensic "science" pratfall. Better than the Three Stooges!