Admissability Challenge in Kansas City

[Steve Everist]

http://www.kansascity.com/news/local/story/549761.html

Hall’s attorney seeks to exclude fingerprint evidence
By DIANE CARROLL
The Kansas City Star

Defense attorneys are asking a Johnson County judge to throw out the fingerprint evidence prosecutors say they have against Edwin R. Hall, who is charged with capital murder.

In a motion filed late Wednesday, defense attorney Paul Cramm states that investigators found at least 29 partial prints on the car that 18-year-old Kelsey Smith was driving when she disappeared June 2 from the Target parking lot next to Oak Park Mall in Overland Park.

Authorities could link only one of those partial prints to Hall, Cramm wrote. The methodology used to make that determination also is subjective and untested, he stated.

The motion to exclude the evidence is supported by affidavits from three forensic fingerprint experts.

Hall, 26, is scheduled to go on trial Sept. 16 in the kidnapping, rape and murder of the Overland Park woman. Her body was found June 6 in a wooded area near Longview Lake in southern Jackson County. Prosecutors are seeking the death penalty.

Cramm said Thursday he could not comment because of a gag order in the case. District Attorney Phill Kline, who has been prosecuting the case with Deputy District Attorney Stephen Maxwell, also declined through a spokeswoman to comment.

According to the motion, a Johnson County sheriff’s deputy processed the interior and exterior of Smith’s car for fingerprints in the garage of the Overland Park Police Department on or about June 4. A second sheriff’s deputy was there to document the process.

The first deputy collected a set of prints from Hall, who was known as the suspect, and made the identification. His work was reviewed by a member of the same agency who knew beforehand that the match had been made.

Cramm argues that the review should have been conducted by someone who was not aware of the initial finding. The methodology used to determine a match is subjective, he said, and having that prior knowledge could have influenced the review.

The methodology used has never been subjected to scientific testing, the motion states, so the error rate in latent print identification is not known. However, it says, there are enough cases of erroneous identifications to establish that a substantial error rate exists.

Judge Peter Ruddick has not yet set a hearing date on the defense request.

[RL Tavernaro]

Anyone know who the "... three forensic fingerprint experts"are?

[Goldencloud]

The three "Experts" noted in the Kansas City Paper are Simon Cole and Mr and Mrs Haber. The defense just filed this motion and no other actions have been taken by the court or District attorneys office as of today.

[sharon cook]

Oh, for God's SAKE!!!!!!

[Red]

http://www.kansascity.com/news/local/story/549761.html

Hall’s attorney seeks to exclude fingerprint evidence
By DIANE CARROLL
The Kansas City Star


[The first deputy collected a set of prints from Hall, who was known as the suspect, and made the identification. His work was reviewed by a member of the same agency who knew beforehand that the match had been made.]

I wonder why the Kansas City Star thought this information should be noted in the article. I can just imaging who they are interviewing.

[Gerald Clough]

Since she mentions that "Cramm argues," it's pretty much right out of the filing. If you think about it from the reporter's point of view of trying to get some of the argument into the story, this is at least something easy to digest without researching the field in general. It's harder to talk about error rate and subjectivity with any hope of making the issue clear in a brief story, or even a long story.

[Michele]

It looks like most of the issues that the defense is raising are Daubert issues. From what I was able to find, I’m pretty sure this is a Frye State. It’ll be interesting to see if the judge cares about these issues if they aren’t part of their state admissibility standards.

Even if the judge allows these issues to be brought forward, I still don’t think they’re valid concerns.

1) Reviewed by a co-worker so the coworker was influenced… Our profession has a standard of verifying conclusions but it’s not legally required (as far as I know). To complain that an agency went above and beyond but didn’t do it the way the defense thinks it should have been done seems a little ridiculous to me.
2) The method is subjective…..we’ve hashed this out several times on the chat board. I know we all have different ideas but I thought that (generally speaking) that the data used is objective and the subjectivity comes into play when determining how much data is enough. That’s involves every scientific conclusion, not just our profession. If that’s a criteria (which I don’t think it is) then all scientific conclusions will be thrown out of court.
3) The method hasn’t been subjected to scientific testing? That’s a big topic for the chat board so I’ll leave that one alone right now.
4) Error Rate isn’t known but is substantial….When did 25-30 erroneous ID’s out of millions or billions of ID’s become ‘substantial’? Does anyone know what percentage rate is considered ‘substantial’? I don’t remember the exact article but Michael Saks wrote a paper with a chart of the error rates of all different forensic disciplines. If I’m recalling it correctly, fingerprints had the lowest rate. I wonder how the defense will support his theory of ‘substantial’ when most critics don’t seem to think this is true. I know Cole wrote an article where he speculated that the number of errors could be very high but did he have anything to support this?

[Steve Everist]

I know Cole wrote an article where he speculated that the number of errors could be very high but did he have anything to support this?

That in and of itself is rather subjective, not supported by any statistics (or lacks a contrast with errors per comparison worldwide), not supported by independent research that is available for peer review, and is only speculation based on having found 25-30 (as you just mentioned) over the history. And of those, how many were fraudulent and not really erroneous ID's?

Even the "tip of the iceberg" can come to a very fine point in comparison to the size of an iceberg. But Cole's reference to the iceberg should consider that the erroneous ID's may actually be no more than the tip while the rest of the iceberg are those that stand up as not erroneous. But there's no value in showing the rest of the iceberg when the tip is the only part that supports the notion.

It would be nice to have someone available to sit in when the motion is heard to see how it is all presented. It's possible that the judge could entirely dismiss the motion and a Frye hearing may not even be held.

I think this may happen more often than we realize, but since it never goes through the process of being a hearing, and coming up on appeal; nobody ever knows.

I had a robbery case where there was a Frye motion to suppress fingerprint evidence. Michele sat in on the hearing for the motion, and the judge denied it and a Frye hearing was never held. Yet very few people outside of our own office are aware - since it essentially became a denied motion. I guess the difference is that this was a kidnap/rape/murder case and mine was 'just' a robbery. There was no media in the courtroom.

[L.J.Steele]

Even if the judge allows these issues to be brought forward, I still don’t think they’re valid concerns.

1) Reviewed by a co-worker so the coworker was influenced… Our profession has a standard of verifying conclusions but it’s not legally required (as far as I know). To complain that an agency went above and beyond but didn’t do it the way the defense thinks it should have been done seems a little ridiculous to me.

A Frye/Daubert challenge is one of the ways blind verification may become legally required -- either through a court ruling that non-blind verifications are sufficiently reliable for admissibility or through something like Mass' DiGiambatista instruction for unrecorded interrogations that tells the jury that when an interrogation is unrecorded, courts have concerns about its reliability.

You are likely to keep seeing this issue as part of a suppression motion and/or cross-exam.

2) The method is subjective…..we’ve hashed this out several times on the chat board. I know we all have different ideas but I thought that (generally speaking) that the data used is objective and the subjectivity comes into play when determining how much data is enough. That’s involves every scientific conclusion, not just our profession. If that’s a criteria (which I don’t think it is) then all scientific conclusions will be thrown out of court.

I don't think the question will ever be "is it at all subjective". The "subjective" question will generally be an amalgam of training, certification and proficiency testing, adherence to proper methods, the difficulty of a particular print, and the significancy of any difference/discrepancies/distortions between latent and exemplar.
Again, this one isn't going away either in suppression motions and/or cross-exam, if for no other reason than to debunk the CSI-myth juror may have that print matches are so mechanical that a computer can match up the points and make the ID with 100% accuracy.

[Steve Everist]

A Frye/Daubert challenge is one of the ways blind verification may become legally required -- either through a court ruling that non-blind verifications are sufficiently reliable for admissibility or through something like Mass' DiGiambatista instruction for unrecorded interrogations that tells the jury that when an interrogation is unrecorded, courts have concerns about its reliability.

You are likely to keep seeing this issue as part of a suppression motion and/or cross-exam.

It seems like that would make things very difficult since verification done blindly removes the scientific process of peer review - unless there is a third person to go over the case and perform the peer review separately. Then, that person would not only become a second reviewer, they would also have to review the original examiner's work along with the blind verifier's work. And thus any case going through blind verification would need to be 'worked' four times to approach being done 'correctly.'

There's nothing wrong with doing blind testing, but that happens long before the verification process, and it isn't necessary in all situations. Those that are starting to incorporate it into their workflow in the verification stage likely weren't doing proper peer review to begin with and likely don't understand the difference between blind testing and verification being done blindly. It's a shame that so many are jumping on the bandwagon because that is what the FBI came up with as a result of Mayfield. Unfortunately they also don't realize that it doesn't make the process any more scientific - in fact done as the only method of verification, it could likely make it less scientific and less likely to meet the peer review portion in Daubert. How do you have proper peer review when you can't review what was done originally?

[Gerald Clough]

Daubert isn't as formulized as so many imagine. Those things so frequently mentioned, peer-review, reliability testing, and acceptance, are merely things that may be used in judging the true issues, relevancy, grounding in scientific knowledge, and appropriate verification. We are somewhat in error when we strive to represent the examination process as a scientific inquiry. A reliable conclusion is possible because the knowledge that is being applied in making the examination is based upon reliable scientific knowledge. There is no specific checklist, even for the base knowledge applied. There could not be, without casting out large volumes of valuable expert testimony based upon knowledge that is far more difficult to confirm experimentally or observationally that anything we apply.

Known error rate, peer-review of scientific principles, standards, and acceptance are simply not required by anything SCOTUS has to say about expert opinion testimony. Daubert guides the nature of the inquiry into admissibility. It does not, for all that challengers would like to present, require LPE to meet all of those tests. For sure, each could be used to argue the toss, but it is going bizzarely beyond Daubert to imagine that either verification of the identification or peer-review of the specific examination are always necessary.

Nor does Daubert meddle in the fact issues of individual cases. That to be admissible expert conclusions need not be verified or even verifiable by other experts is amply demonstrated by the testimony of psychologists and psychiatrists. It is quite the norm for those experts to testify in direct opposition to each other on the specific critical issues, such as competency or retardation. The decision on which opinion is more credible is not a Daubert question at all. It is a fact issue, and the decision often rests on collateral evidence of behavior and adaptive function of the subject, as well as very much on the ability of the expert to believably expound in court. Imagine going to court with a fingerprint identification with equally well-credentialed experts testifying to different conclusions. We're taking shots, largely I think because we weren't previously considered a target, but on credibility, for now, we get pretty much a free ride.

This is not to say that individual courts may view a discipline's satisfaction of Daubert differently. SCOTUS was wise enough, as they generally are, to rigorously avoid hindering lower courts in their evaluation of the cases before them. This is why it's so important for examiners and local prosecutors to have a firm grasp of both Daubert and the factors that would render examiner conclusions both admissible AND credible. Satisfying Daubert doesn't make you credible. If blind verification makes the conclusion more credible, good. If peer-review of the work lends credibility, good. We're not going to see a SCOTUS opinion certifying any forensic discipline. The fight will always be in the trenches, case by case, jurisdiction by jurisdiction.

It may also become true that those things could conceivably become requirements. Not as a matter of law, but because a body of acceptable scientific knowledge might come to show that the process cannot be reliable without them. It's such a difficult area to research that it's not coming soon. But there are examples of such evolution of knowledge. At one time not too long ago, courts routinely held that expert testimony on the reliability of eyewitness identification was inadmissible because it did not assist the jurors who could evaluate it from common knowledge. Experts were put up and knocked down when they couldn't cite applicable research. Today, many jurisdictions are reversing cases in which expert testimony on the general nature of such evidence was excluded. It happened because people like Wells and Malpass developed a substantial body of work in the area AND because their work, although not a complete reflection of eyewitness evidence in law enforcement, was close enough to demonstrate principles.

[Michele]

Lisa,

I see your point about BV never being legally required unless it’s established through a Daubert type hearing but I still don’t think the courts will ever require scientific conclusions to have requirements that science itself doesn’t require. If they did this then the conclusion would no longer be scientific but an even higher standard.

I agree with you about your expanded version of subjective (training, certification and proficiency testing, adherence to proper methods, the difficulty of a particular print, etc). All of these elements should be open for everyone to see and question but I’m wondering if this is what the defense in this case meant by subjective? The article stated it as “The methodology used to make that determination also is subjective and untested”.

Fom Gerald: "We are somewhat in error when we strive to represent the examination process as a scientific inquiry."

I have heard this by others as well and I think it’s important not to jump to any conclusions about the nature of a comparison. Our profession calls our conclusions scientific but the only way to establish if that’s true is to look at each application of the method applied. While some offices adhere to very strict scientific principles, other agencies may be arriving at technical conclusions based on scientific research and ideas.

I guess what I’m trying to say is that the examination process can be scientific if scientific principles are used in arriving at the conclusion. Basing a conclusion on scientific knowledge doesn’t mean the conclusion is scientific. And, just because a conclusion isn’t scientific doesn’t mean that it’s not reliable (a non-scientific conclusion can be very reliable), just as scientific conclusions may not mean that conclusions are reliable. (I don’t think that clarified the issue )

[Gerald Clough]

The reason I object to too much "science" and "scientific" verbage is that I think it actually presents the conclusions of latent print examination as something that might eventually prove to be wrong. For one thing, with few exceptions, we are not scientists. We may be developing and extending techniques, but we are not in the business of making new discoveries about the nature of things. "Science" has an inherent character as unfinished work, one body of work leading to another body of work that sometimes turns the field on its head.

We have something more reliable. The discipline is supported by a body of scientifically derived knowledge. It is the sort of knowledge that, even if it is not precisely accurate in all its details, is entirely sufficient for use as a basis for making examinations. It matters little if future work refines the way friction ridge skin forms. What is known is enough to know with certainty that the number of variations approaches infinity. Latent print examinations are not "experiments." Just because we don't know the conclusion prior to the examination doesn't make it experimental, except in the sense that getting out of bed in the morning is an experiment.

I don't want engineers "doing science" when they settle on the design of the aircraft I will ride in. I'd rather not be part of that grand experiment. I prefer that they apply established Newtonian physics, knowing that it doesn't explain everything that can be observed, but that it is absolutely reliable in applying a method of hurtling me through the air from one point on Earth to another, at least not without an ejection seat and test pilot pay. A quantum physicist could more precisely explain the nature of vibrating atoms and the uncertainty about the locations of their subatomic parts in the air that's keeping me from traumatically altering the lithosphere, but I'll stick with the engineer. And I don't want my doctor testing his new theories on me, since the knowledge gained might be that when he tried that, I ceased to be. The warm knowledge that I may have helped further scientific knowledge hardly compensates for no longer moving, breathing, and eating donuts.

Am I concerned about mistakes in latent examinations? You bet I am. I think there are things to learn about catching those mistakes. But I deal with cases every day with issues in which it is impossible for all the qualified experts to be right in their conclusions. But these are strictly issues in which the courts will judge credibility, in which admissibility is almost never an issue. There is little attack on admissibility because they are areas in which both sides are provided with experts.

I personally think the answers to the issue of mistaken identification lie largely in knowledgeable people on both sides and the experts' ability to explain and demonstrate conclusions. The general lack of expert access on the defense side is responsible for us demanding extreme sanctions against those who make errors. We need it, because we who are almost always working for the State are almost always the last and only concluders. We have done quite a good job of defining standards and, for the most part, the limits of our ability to make reliable examinations. But much of the vigor in attacking examination in court is, I think, because there are so few latent print resources for defendants.

[L.J.Steele]

It seems like that would make things very difficult since verification done blindly removes the scientific process of peer review - unless there is a third person to go over the case and perform the peer review separately.

I think we're talking about two different things. Peer review in Daubert seems to be for the method -- has ACE-V, for example, been subject to published, peer-reviewed studies. Or if the examiner was testifying about a new method for processing latents, has that new method been subject to adequate peer review.

The blind-testing/verification issue has to do with the reliability of the individual results and the confirmation bias problem. Does or should the "verification" step of ACE-V provide an adequate safeguard if it isn't done in a blind manner? How well does the ACE part work if the examiner is exposed to potentially biasing information before or while examining the latent and exemplar?

[Charles Parker]

Peer review in Daubert seems to be for the method -- has ACE-V, for example, been subject to published, peer-reviewed studies.

Correct me if I am mistaken but Daubert states has the methodology been peer reviewed. It does not say anything about studies of the methodology been peer reviewed. I think the word study has been added to that part of the gatekeepers role.

Or if the examiner was testifying about a new method for processing latents, has that new method been subject to adequate peer review.

Correct me if I am wrong but what is adequate? Is that not a subjective criteria? If 4 articles are published in a peer reviewed journal I would consider that adequate but a defense attorney probably would not. Who is going to define adquate? The defense, the prosecutor, academia. I think it should stay with the gatekeeper.

The blind-testing/verification issue has to do with the reliability of the individual results and the confirmation bias problem

Would not that statement be more correct to state "the possible problem with confirmation bias". I do not think two studies, one with students and the other with LPE of which not one made a Type 1 error is adequate converging evidence to state that it is a problem. It might be a problem but certainly needs more studies before it moves from might be to absolute.

Does or should the "verification" step of ACE-V provide an adequate safeguard if it isn't done in a blind manner?

There are not sufficient adquate studies to comply with the principle of converging evidence to say it does not.

How well does the ACE part work if the examiner is exposed to potentially biasing information before or while examining the latent and exemplar?

I know of no studies that have been conducted where type 1 errors were made from potentially biasing information. Now if you want to quote Mayfield go ahead, but one incident does not equal a study.

The evidence should be judged by the gatekeeper on its weight and not its admissibility! Or should that principle be discarded as well.