On the "science" of fingerprints

[Dogma]

I recently viewed the video of the Massachusetts Supreme Judicial Court hearing regarding the Patterson Case and the issue of simultaneous prints. One of the justices, early on, referred to the “science of fingerprint analysis.” He was subsequently “corrected” by defense counsel who stated that the analysis of fingerprints “clearly is not scientific.” This lawyer has himself been counseled by the sty of usual anti-fingerprint suspects. So, please consider and opine on the following observations:

Science does not like skill. Science is a socialist, a true egalitarian. All Scientists educated by rote are equal. Everything is repeatable or it is not science. God forbid (sorry, not very scientific mentioning God there) factors like talent enter the equation. Then, we have no science at all.

Think about some of the early pioneers of science. People like Galileo, Newton and Darwin. What diploma mill spit them out? How is it so many others did not have their insight, their curiosity or their ability to observe and understand the world around them? What we now call science or specifically forensic science is more properly called forensic technology. It’s a sterile environment: no room for anyone not preprogrammed by some packaged curriculum at a cost of thousands of dollars. No room for those who develop their abilities through years and years of experience, observation and trial and error.

Science is routinely defined as :
1. The observation, identification, description, experimental investigation, and theoretical explanation of phenomena.
2. Such activities restricted to explaining a limited class of natural phenomena.
3. Such activities applied to an object of inquiry or study.
4. Knowledge, especially that gained through experience
Wow. Look at that last defining statement: science is “knowledge, especially that gained through experience!”
A scientist is defined as: A person having expert knowledge of one or more sciences, especially a natural or physical science. There are many out there who call themselves forensic scientists. When asked how they established such a credential they will no doubt show you some sort of civil service ID card displaying their employment title. Sorry folks, that “sanitation engineer” who wakes me from a sound sleep once a week is still just picking up garbage.
Some would suggest that when examiners accomplish each task in the same way, following a specific protocol and interpreting the results within predefined parameters, that process is scientific. I suggest that it is a mindless effort unworthy of the label “science.” Enter the crime scene investigator. He must observe the circumstances before him. Evaluate his observations by comparing them to the knowledge he acquired, some through training but mostly through experience at other crime scenes. He must decide how to process the scene in order to preserve and yield the most and best evidence. He must evaluate the evidence that is collected and decide what items are suitable for further analysis. He must document not only what was found at the crime scene but also the collection process and disposition of the evidence. And, after the evidence has been studied and examined, he most likely will be asked to offer an analysis of what occurred. Sounds like a scientist to me! But, to many who read this, he is just a dumb cop.
As with my previous posts, my intent here is not to punch anyone in the nose, but to inspire (incite?) debate. Have at it!

[Free Thought]

Dogman,

You da bomb! Keep up the good work.

There is a benifit to all sciences to leave the public relation image that their science is sterile and 100% objective. Dissenters are silenced! The fingerprint science has contributed to this wrong as well.

In reality skill is a requirement in the practice of all sciences. It must be managed with a good understanding of scientific philosophy, but the need for skill exists everywhere.

Thanks for bringing this up again,

FT

[L.J.Steele]

Everything is repeatable or it is not science.

That is one of the fundamental features distinguishing science from pseudo science.

http://www.quackwatch.org/01QuackeryRel ... seudo.html

I'm not saying that fingerprint comparison is or isn't a science. I have concerns how well some of the fundamental concepts have been critically tested with appropriate blind testing and statistically valid samples. I've got concerns, heighted by Dror's initial study and his follow-up work about confirmation bias as ACE-V is practiced. Human factors such as confirmation bias, tunnel vision, and pressures of high profile cases are mentioned by the FBI itself in the Stacey report. I'm not sure one can credibily deny these aren't valid issues that examiners should be concerned about.

No room for those who develop their abilities through years and years of experience, observation and trial and error.

Under Khumo Tire, authority based on experience, observation, etc. would be deemed a technique, not a science -- the difference between an automotive engineer and a car mechanic or a psychatrist and a therapist. As a practical matter, the analysis under Khumo Tire and Daubert is identical. There are some very good car mechancs and therapists out there who are more useful to the trier of ract than a PhD or MD in the field who don't have the right practical knowledge.

To make something a science, one needs to keep track of all the observations, generate statistically valid samples and tests, create hypothesis, critical test thm using various precautions against subconscious bias, and submit the results to peer-reviewed bodies. Otherwise, memory of experiences can lead on into all sorts of problems with heuristics and selective memory.

He was subsequently “corrected” by defense counsel who stated that the analysis of fingerprints “clearly is not scientific.”

The Daubert process is a remarkably odd way to settle this kind of dispute. One has two (or more) experts, who's explanations are filtered through the direct & cross-examination system, with little ability for the trier of fact to ask questions. Neither the folks asking the questions, nor the trier of fact, necessarily have any prior knowledge of the field.

On appeal, the material is filtered through two appellate attorneys (who may not even be the trial attorneys), distilled into cold record, with citations to various journals and books that the appellate judges may, or may not read. And again, the judges themselves do not necessarily have any prior knowledge of the field.

On top of all of that, the only relevant issues are those related to the case the Court is deciding, which may have all sorts of interesting quirks in its record that make broader applications of the rule less useful.

Perhaps a better system would be something like the TWGEYE process used to create the DoJ Guide on Eyewitness ID where the experts, critics, prosecutors, and defense attorneys got together outside the limits of specific cases to try to figure out a better way to resolve the issues.

Sounds like a scientist to me! But, to many who read this, he is just a dumb cop.

It sounds more like an experienced technician -- more akin to the wiz of car mechanic who can fix one's problem quickly and efficiently based on experience, the physical therapist who can diagnose a problem based on the oral medical history, and so on. Or even an experienced lawyer who can listen to the client's story, look at the case law, and come to a quick decision based on experience about what tactics are most likely to succeed. What I do isn't a science, but it is a product of observation and experience.

There's no stigma in being a technician.

Just to digress a sec.

I am frankly, quite fed up with the constant accusations that the Science of Fingerprints is fallible.

Manifestly, all human endevors are, but that there's a chance of a good-faith error doesn't make fingerprint matching into a pseudo-science like astrology or phreonology. It merely means that there is a small chance that the examiner's results, no matter how confidently expressed, are wrong. I don't quite understand the reluctance of examiners to admit that there is a chance that a result could be wrong.

There's a chance in every case that I work on that I could miss something vital to the client. So far, the habeas courts have agreed with me when ineffective assistance of appellate counsel claims have been raised ... but I'm glad there is that layer of critical oversight, even if it is a pain the *$% to have to explain one's actions a decade later. The important thing isn't whether I'm right or not, it is whether the right bad guy was convicted in a fair trial.

in the back of my mind I cannot help but wonder how significant fingerprint evidence would be should any of them ever have to rely upon it solve a personal catastrophe?

Cole addressed this at the NEDIAI debate in Nov. He agreed that if he were the victim of a crime and a print were found, he'd want law enforcement to try to match it to the perpetrator. He said he'd be concerned about an error if he learned that a culprit was being prosecuted on the basis of a single latent -- no prior record, no confession, no other forensics. But he, and I think most of the critics, would admit that fingerprint matching likely gets the right bad guy the vast majority of the time. The critics are saying: acknowledge the risk of a good faith error, take all the feasible steps to prevent errors, and buttress the fundamental hypothesis of the field with some good, solid studies. Again, I'm hard pressed to see why this is deemed unreasonable.

[Dogma]

Happy New Year Ms. Steele and welcome to the debate

That is one of the fundamental features distinguishing science from pseudo science.

However it does not take into consideration the ability of the challenger to perform the same tasks as the original examiner. Is it a matter of faulty science or merely incompetence on the part of the challenger?

To make something a science, one needs to keep track of all the observations, generate statistically valid samples and tests, create hypothesis, critical test thm using various precautions against subconscious bias, and submit the results to peer-reviewed bodies.

This takes a very narrow view of what is science. Not all science is conducted in a laboratory with sophisticated technology. In fact, some "scientific work" becomes generally accepted until experience proves it wrong or a specific investigative challenge is mounted. Just look to the pharmaceutical industry for proof of that. Vioxx anyone?

BTW, as an aside, Ms. Steele, I found it interesting, though predictable, that the top two returns, on a Google search for Vioxx, were for law firms!

I've got concerns, heighted by Dror's initial study and his follow-up work about confirmation bias as ACE-V is practiced.

While Dror's study may indicate a potential problem, the sample size is still very small and the work has not been tested and verified. We can't have it both ways. If the problem with fingerprint science is that it has not been critically tested, we can't use equally untested theories to combat it.

The Daubert process is a remarkably odd way to settle this kind of dispute...

I think that one of the problems, in the minds of many, with Daubert is that it is being used retroactively. Some believe that Daubert was intended to protect against novel and unfamiliar practices being used to establish facts in court. It was not intended, some believe, to address practices that had been used forensically for a hundred years. This argument reminds me of the RICO act which was designed to be used against organized crime but was used by some prosecutors to address civil unrest.

It sounds more like an experienced technician --

A true crime scene investigator is no technician. There is a lot of intellectual work going on there. Technician is more aptly applied to those "forensic scientists" working in labs who never need to think outside the box and claim they don't have to because the circumstances of the evidence find has nothing to do with the identification of the specimen. Not a lot of intellectual curiosity there. As I said before, sounds like the CSI is the scientist and the "forensic scientists" are the technicians who help him in his research (investigation).


Here is another interesting observation...although a number of people are challenging rhetorically whether a latent print can be reliably individualized to a known source, only because certain testing has not been done, no one has taken the initiative to prove that a latent print can't be reliably individualized to a known source. Hmm...maybe its a matter of "why waste your time when you already know the answer."

[Michele]

I've got concerns, heighted by Dror's initial study and his follow-up work about confirmation bias as ACE-V is practiced.

Lisa,

Dror’s studies never looked at confirmation bias “as ACE-V is practiced”. The study showed that examiners were inconsistent in their conclusions. Because biasing information was introduced, it was concluded that the biasing information caused the inconsistent conclusions.

You mentioned the need for “critical testing” and “good solid studies”, which I couldn’t agree more. I found this study to be extremely interesting but it doesn’t appear there were many controls placed on the sample group (how was expert status determined), no controls of how a conclusion was arrived at (whether ACE-V was used), no controls of other factors that may have lead to the inconsistent conclusions (are there standards for what constitutes an identification), and no control group where no bias was introduced. Without considering these factors, it doesn’t appear to me that this should be considered a solid study.

On a superficial level, having this paper (“Contextual information renders experts vulnerable to making erroneous identifications”) published in a scientific peer review journal (Forensic Science International) gives the conclusions of this research credibility. The measure of whether or not a study is sound should be based on how thoroughly and objectively the hypothesis was researched, and not based on the fact that it was written by someone with a PhD or published in a scientific peer review journal.

The results of this study do warrant concern but should be viewed as initial research where more thorough and comprehensive examinations need to be done.

[Gerald Clough]

The results of this study do warrant concern but should be viewed as initial research where more thorough and comprehensive examinations need to be done.

Well to remember, since we're talking science, that despite media headlines that bounce the public back and forth between alternatives in such things as various health practices, few studies do anything more than lead the way to further study. Often, it's more a matter of learning how to design a study.

[Steve Everist]

Well to remember, since we're talking science, that despite media headlines that bounce the public back and forth between alternatives in such things as various health practices, few studies do anything more than lead the way to further study. Often, it's more a matter of learning how to design a study.

Gerald that's an interesting point, regarding all of the media headlines regarding studies. It seems like we're being bombarded with studies on a daily basis through the media. Our perceived need for the newest findings seem to override the potential inaccuracies of those findings.

CNN had an article in July stating that 1/3 of study results don't hold up over time. They dealt mainly with a range of drugs or treatments, however. You can find it here: http://www.cnn.com/2005/HEALTH/07/13/co ... tudies.ap/

I particularly liked these statements:

"The crazy part about science and yet the exciting part about science is you almost never have something that's black and white," said Dr. Catherine DeAngelis, JAMA's editor-in-chief.

Editors at the New England Journal of Medicine added in a statement: "A single study is not the final word, and that is an important message."

The problem is that too many people are reading only the headlines and aren't getting into the content and the design of the study. The notion that; if it's published, it's accurate, can be a dangerous thing. Of course it took further study, as you mentioned, to eventually disprove the earlier results. Of course given enough time they could bounce back.

[L.J.Steele]

I don't disagree with the criticism of Dror's study. I agree ths sample size is small and that the conclusion to be drawn are limited. (Heck, I think I noted that when I posted links to the New Scientist article on Dror's study a couple of months ago.) Dror is working on a larger follow-up study, I'm told. We'll see where that goes. I think the study is worth noting as a flashing "caution" light, even if one takes the conclusion with a grain of salt.

One needs some mechanism for reviewing practices that have long been accepted as a matter of tradition but not critically challenged. In the past few decades we have, for example, learned a great deal about how eyewitness ID works (and doesn't work) and about ways to make IDs more reliable. Daubert is _a_ way to bring that material to the attention of the courts. The goal is to make sure the right guy is convicted for the offense using a fair process and solid methods, not to convict the wrong guy because we've been using the same method for 80ish years and, in this case, the method turns out to give the wrong answer.

And fingerprints are not the only folks feeling the heat. Another lawyer just passed this link on to me...

http://pacer.mad.uscourts.gov/dc/cgi-bi ... mber20.pdf

U.S. District Court in Mass has just excluded some firearms ID testimony -- and you'll see the familiar issues of certification, standards, blind testing, confirmation bias, and so on being discussed by the judge. (This one seems to have fallen below the media's radar, I didn't see anything about it on Google.)

[Gerald Clough]

And fingerprints are not the only folks feeling the heat. Another lawyer just passed this link on to me...

http://pacer.mad.uscourts.gov/dc/cgi-bi ... mber20.pdf

U.S. District Court in Mass has just excluded some firearms ID testimony -- and you'll see the familiar issues of certification, standards, blind testing, confirmation bias, and so on being discussed by the judge. (This one seems to have fallen below the media's radar, I didn't see anything about it on Google.)

In the opinion, there were some serious issues peculiar to firearms examination. But the most damaging was the documentation issue, which the court took as a distinct problem in judging credibility. In the current, somewhat unpredictable, legal climate, documentation becomes an especially powerful issue when a court decides to allow testimony as to observations, but not conclusions.

A great many latent print comparisons can be demonstrated by a detailed reconstruction of the examination and demonstration of all the observations to the point where the examiner's conclusion is not required for the jury to form their own. Such recitations of observations really make up a great deal of what goes on in trials, officers and witnesses recounting many observations about scenes, demeanors and all sorts of things, without articulating any conclusion, and jurors routinely draw reasonable conclusions from the information.

All in all, I wouldn't give this particular case too much weight. The judge's decision is, I think, one driven by the whole array of difficulties, rather than being purely a response to the "science" issue.