Bite Mark Evidence Not Ready for Primetime:

Despite what you might see on C.S.I. (or CSI: Miami or CSI: NY), bite mark evidence is not so reliable, the New York Times reports. I am sure the same can be said of many other types of evidence relied upon in these shows -- even if one were to assume that the average police crime lab has anywhere near the level of sophistication and lack of time and resource constraints as the labs on these shows. (Of course, they can be fun to watch anyway.)

Interested Clerk (mail):
Are jurors who have seen the prosecution successfully use a type of evidence on TV more likely to credit it ("If it's good enough for the TV jury, it's good enough for me") or less likely ("This evidence doesn't look as perfect as the evidence on TV")? In a case involving bite-mark evidence, is it valid to ask during voir dire whether a potential juror watches CSI, or has watched a specific CSI episode involving bite marks?
1.28.2007 12:24pm
LTEC (mail) (www):
Actually, bite-mark evidence is perfectly valid, as demonstrated in a television documentary I saw. The murderer was convicted because he was linked to a bite-mark on a week-old piece of cheese that was found in the victim's apartment. (I believe the detective's name was Columbo.)
1.28.2007 12:36pm
B D McCullough (mail):
Neither is "fingerprint evidence" very reliable. See
from which one paragraph to whet the appetite:

"Judge Pollak, who is a former dean of the law schools at Yale and the University of Pennsylvania, also noted "alarmingly high" error rates when fingerprint examiners took proficiency tests; in 1995 only 44 percent of 156 law enforcement examiners could correctly identify all five prints in the test, and in a 1998 study the number improved to only 58 percent."

And, of course, it's not too difficult to believe that if fingerprinting isn't reliable, neither is "DNA fingerprinting". For a non-technical explanation see, specifically, "Statistics, Litigation, and Conduct Unbecoming" by Seymour Geisser in J. Gastwirth, ed., _Statistical Science in the Courtroom_, Springer 2000, esp. section six. In a nutshell, without random sampling, true probabilities cannot be calculated. The DNA fingerprinting methodology uses a "convenience sample", not a random sample. Hence the "probabilities" calculated are not true probabilities, and the asserted "probability of a match" is nonsense.

Bruce McCullough
1.28.2007 12:36pm
Interested clerk: yes, I think inquiry re whether the jurors watch CSI or other shows dealing with solving crimes with novel evidentiary tests is proper inquiry on voir dire.
1.28.2007 12:43pm
Dave N (mail):
A particular pet peeve of mine is "Law as Taught on Television." I am not sure there is enough material for a law school class, but there certainly is for a continuing legal education seminar.

Bad science is rampant on television--and leads to what some commentators call "The CSI Effect." People expect crime labs to be able to do more than they are capable of because the crime labs on television do it.

Bad law is equally rampant of course. For example, many people, including many cops, think Miranda warnings are required any time someone is placed under arrest. When I teach laws of arrest to police officers I tell them, "Unless it is your case or you have been told by the officer whose case it is to do so, you do not, under any circumstances, give a suspect Miranda warnings." I add that if I were in charge, doing so would warrant a suspension.

Last night, Crossing Jordan managed to mix both bad law (search warrants being issued in open court by a judge who admits on the record the request lacks probable cause) and bad science (the ability to "reassemble" a body into a coherent shape) in one mere hour.

A good technical consultant appears out of reach (or out of budget) for most of this dreck.
1.28.2007 1:44pm
Bill Poser (mail) (www):
I understand that prosecutors now consider it a problem that jurors expect magical, conclusive forensic evidence on the basis of their television viewing, but what bothers me even more is the use of unreliable types of evidence by prosecutors. Given Daubert and all the criticism of questionable scientific evidence in civil proceedings, why is it that forensic evidence of types known to be unreliable, such as the instant bite mark analysis, or lacking both theoretical and empirical validation, such as "voiceprint" identification, are admitted?
1.28.2007 2:50pm
Bill Poser (mail) (www):
Dave N: What makes the bad law and science on TV even more unacceptable is that they could eliminate the bulk of the errors without even using fancy, expensive experts. For most of it, they don't need an authority on criminal law or a forensic scientist. They could probably deal with most of it merely by having someone on their staff who had read a book for lay people on criminal procedure and some of the guides to forensics available for lay people, e.g. one of the series aimed at (wannabe) authors of detective stories. Moreover, even hiring a criminal lawyer and a forensic specialist to vet their scripts would probably not be expensive by their standards, not compared to what they pay the actors in such shows. I really doubt that the constraint is financial. I think that they just don't care.
1.28.2007 2:58pm
Tek Jansen:

Bite Mark Evidence Not Ready for Primetime

Marv Albert was ahead of his time.
1.28.2007 3:32pm
William Oliver (mail) (www):
The problem is not that bitemark evidence is not "reliable," its that people push it beyond what it can do. If you have a bite mark that shows that someone has five missing teeth, a chip in one tooth, and another tooth that is going sideways, and that suspect has five missing teeth, a chip in one tooth, and another tooth going sideways, that's pretty "reliable."

The problem comes from forensic odontologists who push this to the point that they have indistinct impressions and from that they extrapolate to fine detail about the teeth. *That's* unreliable.
1.28.2007 4:25pm
Elliot123 (mail):
When writers have to churn out one show afer another, I doubt we will see much attention to the accurate portrayal of any profession. They obviously don't have the experience to know the details involved, but do have the talent to weave the little they know, or more often the little they suspect, into a passable 42 minutes of TV.

They recently achieved a milestone on CSI. They had a video of a roomful of people. Then they zeroed in on one woman's eyeball and extracted the reflection of what she was watching. Magnificent.
1.28.2007 4:32pm
Bill Poser (mail) (www):
If treated as a technology capable of providing a unique identification, as it too often is, bitemark analysis is not reliable. Even in the example of the person missing five teeth, a chip in one tooth, and another going sideways, while it is certainly true that that only some fairly small percentage of people will match that description, it is certainly not a unique identifier. Moreover, the probability of a random match in such features is not quantified.

When techniques such as these are used, even where they may be used with some validity in some circumstances, just asking the "expert" whether or not there is a match is foolish. The question has to be: using the material available and the technique used, what is the empirical (as opposed to subjective) probability of a match?
1.28.2007 4:51pm
Bill Poser (mail) (www):
Incidentally, has anyone paid attention to "Numbers"? I don't watch it much, but last year at the American Association for the Advancement of Science meeting I walked in on the tail end of a symposium in which the actor who plays the mathematician participated. It was apparently about how they are trying to use real math and science and use the show to educate as well as to entertain.
1.28.2007 4:54pm
Automatic Caution Door:
they could eliminate the bulk of the errors without even using fancy, expensive experts. For most of it, they don't need an authority on criminal law or a forensic scientist. They could probably deal with most of it merely by having someone on their staff who had read a book ... Moreover, even hiring a criminal lawyer and a forensic specialist to vet their scripts would probably not be expensive by their standards, not compared to what they pay the actors in such shows. I really doubt that the constraint is financial. I think that they just don't care.

I seriously doubt the production of these programs takes place in a complete vacuum. It would quite surprise me to learn that the "CSI" series, for instance, doesn't employ legal and forensic consultants. That doesn't mean the writers and directors aren't taking literary license -- obviously -- but I doubt they're just winging it from the get-go, either.
1.28.2007 4:57pm
Automatic Caution Door:
I suppose I should have done the Googling first:

CSI: Our Consultants
1.28.2007 4:59pm
mccullough - your claims about bitemark and fingerprint eveidence aren't relevant to the reliability of those types of evidence - since you are just pointing out that some examiners of that evidence tend to suck. that says nothing about the reliability of fingerprints themselves. and of course, both sides can hire experts to determine the match.

Dave N. i totally disagree about the miranda comment. it IS true that miranda is not required UNLESS one is conducting custodial interrogation (except in my state a truncated miranda known as the "administrative warning" must be given after arrest, even if no questioning commences, but that's another issue), but to say that it's bad for patrol officers to mirandize somebody upon arrest, and that unless told by a detective - they shouldnt - is insane.

the reason is (Among others) - spontaneous statements. yes, it is true that if you don't mirandize somebody and they are in custody (and note that in WA you have to give the admin warnings "right to an attorney, and if you cannot afford one, one will be appointed free of charge) in EVERY arrest), and they start admitting to stuff, then you can't question any further without THEN mirandizing them, which is far more likely to screw up the investigation since they are then gonna be less likely to continue. better to get it out of the way UPON arrest, imo, since that is more mechanical and also doesn't bring up the spontaneous problems.

aLSO, even if you are arresting somebody for something totally minor, MIRANDIZE them, then if they start spontaneously admitting to murdering somebody you have already got that out of the way.

i have many years experience as a detective and other assignments, and at least in my agency (and others i have worked for) it is expected that officers mirandize upon arrest for PC of whatever. whether or not they are to question a suspect in my case, as a detective , is another issue Entirely - but miranda ? absolutely
1.28.2007 6:22pm
markm (mail):

On Tuesday, Mr. Brown was released from prison, after DNA testing on the saliva left by the biter proved his innocence and implicated someone else in the crime. At the time of his conviction, Mr. Brown, 46, was missing two front teeth. The bite marks, meanwhile, had six tooth imprints.

The prosecution's expert testified that Mr. Brown could have twisted the victim's skin to fill the gaps his missing teeth would have left on the bite mark. Despite rebuttal from the defense, whose expert said the bite marks excluded Mr. Brown, it took jurors five hours to return a guilty verdict.

It sounds like the problem in this case wasn't the unreliability of bite mark evidence, but a prosecution expert witness that committed perjury about either the evidence or his qualifications - and a jury that thinks prosecution witnesses are more reliable.
1.28.2007 6:39pm
Hattio (mail):
Surely you arne't suggesting that prosecutors over-reach and then use questionable experts to back up bogus claims? Why, I never.
1.28.2007 7:54pm
Dave N (mail):
The reason my rule of "No Miranda" is very simple. Officer X, who is not involved in the case other than transporting the suspect gives the Miranda warning and the suspect says, "I want to talk to a lawyer."

Later, the detective in charge re-Mirandizes the suspect, who freely and voluntarily confesses--perhaps he even tells the detective where evidence is that will not be inevitably discovered. Under Edwards v. Arizona and its progeny, the statements to the detective is out because some moron driving a patrol car learned law watching Law and Order.

While United States v. Patane at least suggests that evidence from such an interrogation might be admissble, why create a legal issue when none is necessary?

So I stand by my original comment regarding the giving of Miranda warnings. This advice is not "insane" at all. It deals specifically with officers who have NO real role with the case other than providing transportation.
1.29.2007 2:21am
Mississippi Lawyer:
Whether jurors watch television shows about forensic evidence, either fictional or documentary in nature, seems to me to be an interesting avenue in voir dire.

It might even provide a pretext for striking a juror for cause.
1.29.2007 9:48am
and dave, my objection is the same. there is the spontaneous statement issue, as well as the others mentioned (plus, again my state has different laws - requires that truncated warning evne if not questioning)

i've worked as a cop in 3 states, and along with many prosecutors. your method was universally frowned upon in those three states.

this has nothing about "learning law from law and order" by some "moron" (nice bias) driving a patrol car. considering i have seen many prosecutors who know far less about miranda than i do.
1.29.2007 11:43am