CLPEX Chat Board

More recent studies have also shown problems. Though a 2005 study in the Journal of Criminal Law & Criminology suggests a fingerprint false-positive rate a bit below 1%, a widely read 2006 experiment shows an alarming 4% false-positive rate.

Yet the public sees errors as gross anomalies. Like the time the FBI wrongly linked an Oregon attorney named Brandon Mayfield to the 2004 Madrid commuter train bombing that killed 200 people. The FBI had claimed a 100% match between fingerprints found at the scene and Mayfield, who was held for two weeks in federal custody. When the Spanish National Police got the real perpetrator, an Algerian named Ouhnane Daoud, the FBI had to admit its mistake. Mayfield accepted a $2 million settlement from the government.


http://www.forbes.com/forbes/2008/0602/038.html

First: comparing real forensic science to "CSI." Oh dear. You know that's like comparing "Perry Mason" with real court.

Second: once again, the writer is confusing individual incompetence with an indictment of the entire field of forensic science.

Third: the fact that the Brandon Mayfield case is so famous should tell you something about how often that sort of thing happens.

Fourth: what never gets reported is how often re-testing in the "Innocence" Project re-confirms the original conviction.

And finally, exactly which "...widely read 2006 experiment shows an alarming 4% false-positive rate?" (Could that have been from the article from the Jan-Feb 2006 JFI "A Report of Latent Print Examiner Accuracy During Comparison Training Exercises" by K.Wertheim et al? I suggest you actually read that article, if so).

Ooooh Forbes---A Peer Reviewed Journal.

We have two legal issues. One is admissibility, in which we address the validity of the discipline as a science-based source of relevant conclusions. The other is credibility of the individual examiner and the individual conclusion. I think the two issues are often addressed improperly in challenges. The known errors are not relevant to the issue of admissibility, so long as the process itself is not shown to be so ill-founded that errors are to be expected. This is where I would like to turn the "no know error rate" argument back on the challengers. If you are going to claim that a procedure produces some number of errors on account of its fundamental nature, you must show me that there is a fundamental flaw in the customary chain of knowledge and analysis to conclusion. You cannot simply say, "Look. There are some errors." If that is the test, we can throw out all medical evidence, too, since, for example, medical examiners have ample documented instances of serious error. You cannot condemn forensic medicine by showing that some M.E.'s have failed to correctly interpret an observation or that they have failed to properly observe.

The other issue is credibility of an individual examination. In challenging the credibility of a conclusion, it is perfectly legitimate to demonstrate that error is possible. This is the rightful place in litigation for known errors to appear. Their significance is not that the discipline is fatally flawed, but that humans can make mistakes. The evolution of admissible expert testimony in such areas as eyewitness identification suggests it may even be appropriate to bring in experts on the nature of latent print examination at this point. And I can attack an eyewitness by asking if they have ever thought they recognized someone and found they were wrong. Any honest person must answer that they have, and anyone who says they haven't will likely not be seen as credible. We would hope the LPE would be able to answer that they've never found that they were wrong in an official conclusion. (I'm not entirely happy with the one-error-and-out philosophy, because it sets up an argument that admitting error is so devastating that many or most would not admit their error.)

We have a tremendous advantage over something like eyewitness identification, because we can take a fact-finder through the examination in fine detail. You can hardly do that with an after-the-fact examination of a eyewitnessing. We have an advantage over DNA, in that it's pretty hard to observe the molecules dance and observe the actual analysis. We can show our analysis and essentially reenact it. The article calls for multiple independent reanalysis, and I don't think defense experts in latent cases is any kind of bad idea. But we have the capability to demonstrate an examination so thoroughly that it doesn't contribute much to the system. Whether such a thing comes to pass depends, I think, on more knowledgeable attorneys and examiners better prepared to present evidence.

Charles Parker wrote:Ooooh Forbes---A Peer Reviewed Journal.

No disagreement as to its authoritative or even persuasive power, mentioned only to let y'all know what's in the popular zeitgiest.

Gerald Clough wrote:We have two legal issues. One is admissibility, in which we address the validity of the discipline as a science-based source of relevant conclusions. The other is credibility of the individual examiner and the individual conclusion. I think the two issues are often addressed improperly in challenges. The known errors are not relevant to the issue of admissibility, so long as the process itself is not shown to be so ill-founded that errors are to be expected.

To some extent, I talk about Mayfield, Cowans, and the other known mis-id cases to get the appellate judges (I generally don't deal with trials) past their pre-existing beliefs in the infallibility of the field as a whole. There's still some folks out there who don't know that mistakes can happen -- and thus there's a reason to ask questions about the underlying science. This gets us into field-wide issues like the Patterson question about simultaneous impressions or confirmation bias issues and whether non-blind verification does what it purports to do in ACE-V.

Gerald Clough wrote:The other issue is credibility of an individual examination. In challenging the credibility of a conclusion, it is perfectly legitimate to demonstrate that error is possible. This is the rightful place in litigation for known errors to appear. Their significance is not that the discipline is fatally flawed, but that humans can make mistakes.

This is more a trial-level problem than an appellate issue, tho one can get into some interesting questions about the difference between how prosecution and defense experts are often treated .

Gerald Clough wrote:(I'm not entirely happy with the one-error-and-out philosophy, because it sets up an argument that admitting error is so devastating that many or most would not admit their error.)

I have similar feelings -- I'd be more comfortable, in the abstract, with someone who admitted to an error, took the reputation hit, and then used that to learn from.

Gerald Clough wrote:We have a tremendous advantage over something like eyewitness identification, because we can take a fact-finder through the examination in fine detail. You can hardly do that with an after-the-fact examination of a eyewitnessing.

Again, the problem goes back to Mayfield, where a court-appointed expert missed the error, and Cowans, were allegedly two defense experts missed the error. As I recall, there's a few more examples of defense experts missing mis-IDs in Cole's list. This makes me very nervous about relying on going back thru the examination after the fact rather than getting it right the first time.

L.J.Steele wrote:To some extent, I talk about Mayfield, Cowans, and the other known mis-id cases to get the appellate judges (I generally don't deal with trials) past their pre-existing beliefs in the infallibility of the field as a whole. There's still some folks out there who don't know that mistakes can happen -- and thus there's a reason to ask questions about the underlying science. This gets us into field-wide issues like the Patterson question about simultaneous impressions or confirmation bias issues and whether non-blind verification does what it purports to do in ACE-V.

I understand that. It's easy to forget that while most examiners keep up, including participating on a board like this one, there are even folks out there operating as latent print examiners who are unaware of the error cases. The appellate world is so different from the trial. Even more so in capital appeals, with its odd mix of civil and criminal. Those trial distinctions between the pretrial issues of admissibility and fact issues of credibility break down at appeal where both issues can be revisited in a single hearing.

Again, the problem goes back to Mayfield, where a court-appointed expert missed the error, and Cowans, were allegedly two defense experts missed the error. As I recall, there's a few more examples of defense experts missing mis-IDs in Cole's list. This makes me very nervous about relying on going back thru the examination after the fact rather than getting it right the first time.

I think that's really what drives a lot of the efforts to attack the fundamentals of latent comparison. We're working with such a very few documented errors that aren't plain and obvious inattention. The bottom line problem is one of determining how likely it is that there is some significant number of similar errors undetected.

There are, of course, two not mutually exclusive approaches. One is to develop means to insure that such mistakes don't happen, which really means they don't get out of the examination process. The other is to develop means to insure that errors that make it out of the examination are discovered in the course of litigation. My feeling is that the measures in place for the primary examination are as much as can reasonably be done at that level.

We have significant evidence that properly conducted reexamination by an independent expert can repair error. I think that the instances where it has failed to detect actual error may well have to do with the rarity of such independent expert review, the fact that it is far from routine to have identifications contested, which probably speaks more to attorney awareness than anything and, of course, the fact that that the chance of an incorrect identification is so very small. As to why supposedly independent re-examiners and even defense consultants have failed to detect error, I think we're maybe back to the severe philosophy of error and the expectation that conclusions from respected examiners are correct.

Sometimes the issue is the defense expert. We have seen the use of defense experts who had only moderate experience in fingerprints, but were hired by the defense. One had a few years of 10-print and very little latent experience and 20 years of auto parts experience. He was hired by the public defender's office for several of our cases and had done some work in Miami as well. What are the chances he could catch an error?

I see some of that, especially in forensics. It might be a retired analyst who worked in one or two areas now claiming expertise in several other areas. Few attorneys have the experience and knowledge to evaluate resumes. I have seen them cite multiple "course" titles in forensic areas. When I research these, I may find the "courses" to be 30 minute or one hour workshops sessions or presentations at conferences not devoted to that field, the entire training in a claimed area amounting to an aggregate three hours and no experience, outside of whatever few consulting jobs they have managed to hook. The tip-off is often that the conclusion in the instant case is absurd. These folks seem to demonstrate that they either could not recognize an error in the official analysis, given that they can't develop a legitimate conclusion of their own, or that they are willing to say anything of benefit to their client. Catching these things is part of my job.

We, as experts, can't fix everything in the system. It still depends on attorneys knowing enough to critically evaluate evidence and experts. If you're doing criminal defense, you ought to have as much knowledge of the nature of evidence as you do about the legalities of how it was obtained. There are plenty of competent consultants around. I'm all for competent experts working the other side. Better defense advocacy makes for better cases.

In one case I examined for the defense a few years ago, I confirmed that the state fingerprint experts had done everything correctly and their conclusions were correct, i.e., a bloody palm print from a murder scene was that of the murder victim. That conclusion did not fit with the defense theory that the bloody palm print was that of the real murderer and thus, the man convicted of the murder was innocent. Since the defense attorney did not like my conclusion, he hired a different defense expert whose training and experience in latent prints was suspect. That "expert" fell in line with the defense theory and pronounced the state experts were wrong, the bloody palm print was not the victim's but must be that of the real killer, and the state's fingerprint experts must have been part of a conspiracy to convict an innocent man. Litigation is still going on in this case and I may end up testifying for the police in civil suits, even though I was originally retained by the defense.

In another case I reviewed for the prosecution as an independent, outside expert, a bloody fingerprint at a double homicide was presented by the defense as that of the real killer. A defense expert in that case had excluded the defendant as the source of the bloody fingerprint. My examination concluded that the defendant could NOT be excluded -- pattern match and a couple of points in a distorted bloody print. The defendant was convicted and that case is in a situation of extended (perpetual?) appeal.

I do my share of defense cases outside of my home state and have caught some bad ones. When an attorney contacts me, the first thing I do is send him a "caveat letter" warning that I will examine the evidence and render a scientifically verifiable conclusion. That letter goes on to state that if the police have been careless, made a mistake, or committed a fraud, I will be an outspoken advocate for the defense; but if the police have done everything correctly, I will not help defeat a good case. In the majority of defense cases I do, that letter meets with instant approval from the defense attorneys, who only want an honest conclusion so they know how to defend the case. But in a small percentage of cases, the defense attorneys do not want an honest second opinion, they want testimony to validate some preconceived theory they have contrived to beat the case. Those are the attorneys who seek out the "Witness Having Other Reasonable Explanation," for which we normally use only the acronym.