CLPEX Chat Board

I came across this article today, written by Michael Cherry and Edward Imwinkelried, titled How we can improve the reliability of fingerprint identification

What I found interesting was the thanks given to David Ashbaugh for reviewing the article. After reading it, I would guess that this was done in order to give it credibility. I would also guess that whatever critique it received by David Ashbaugh was likely ignored.

It can be found at:
http://www.ajs.org/ajs/publications/Jud ... ry_902.pdf

All I can say is Wow! Where did Cherry get all of his statistics and information? Nothing is cited in his paper save for the two cases he mentions. I received a degree in biology only six years ago and I don't ever remember a professor ever saying it was okay to write a paper (in which 'facts' were inlcuded) without using sources and listing those sources. I rarely have anything good to say about defense attorneys and this supports my reasonings. There's so many things fundamentally wrong with this article that I don't want to waste time writing about them (I'm sure most of you know what I'm talkin about!).

And to think that (most) defense attorneys take 'opinion magazines' where this article came from as the gospel!

I agree with Steve...I'm wondering how many of Ashbaugh's comments were addressed when publishing this article. I will see him in Phoenix in two months and I'll have to ask him about it!

Josh

josher89 wrote:I rarely have anything good to say about defense attorneys and this supports my reasonings.

[snip]

And to think that (most) defense attorneys take 'opinion magazines' where this article came from as the gospel!

Judicature is published by the American Judicature Society, which its website say is "a nonpartisan organization with a national membership of judges, lawyers and other citizens interested in the administration of justice." If you want to discuss defense-oriented journals, you should probably look at Champion, NACDL's magazine. You'll find a mix of stuff there on various forensic topics.

As to reliance --

An attorney would likely consider an article in such magazines, be it the ABA's Journal, NACDL's Champion, or one's state bar's journal as potential inspiration. I've gotten good ideas and insights from various journals, but the key is the underlying sources, not the author's opinion (unless the author is someone very prominent in that specific field). These sorts of articles are useful to me in finding out about cases or peer-reviewed scientific materials I hadn't known about.

In a few cases, I've used these sorts of magazines to discuss what a competent attorney might have known about a topic at a particular time. (A potential issue in habeas claims where one may argue whether counsel was ineffective for not raising a particular issue.) One might use an article like this to discuss whether an attorney in 2006 would be aware of potential fingerprint issues, but Mitchell, Plaza, Mayfield, and Cowans would all be much better reference to show that fingerprint challenges were by-then well-known potential issues.

Many of these magazines are looking for potential authors -- if you want to write a better guide for attorneys and judges on fingerprint issues, check out the magazines' submission guidelines and send something in.

Reintroducing Henry 10-finger classification to help manage large fingerprint collections would be something akin to reintroducing men with red flags walking in front of the cars to reduce crashes in NASCAR.

And how does Mr Cherry propose that we detect and correct the 5% of duplicate records caused by alternate identities? Doesn't he know that AFIS systems already do this using the fingerprints themselves and that the residual 5% are caused by badly or incompletely taken sets that resulted in a system miss?

And of course here's one we've heard before:
"If analyzed properly, fingerprints can be as accurate as DNA."

2 words Michael....homozygous twins

And one more thing...

Can anyone tell me the core business of this company?

http://cherrybiometrics.com/

Their webiste doesn't appear to mention it

A defense attorney recently questioned me regarding on of Cherry's articles, in which Cherry stated something to the effect that Henry Classification was more accurate than newer methods. At least, that was the spin the defense attorney put on it. I replied something akin to David's comment, that the Model A was a great car in the 1930s, and Henry Classification was on the cutting edge in 1890. But we've come a long way in both automotive science and in fingerprint science. To go back to Henry would be a century jump into the past.

Just to clear things up - Michael Cherry is not a lawyer. He has actually participated in these forums (and the old version of the forum) in the past. As far as his website, I think his business is related to biometrics and national security, although I'm not sure the scope.

Josh, I'll be interested to hear David Ashbaugh's feedback on this article and his review of it.

This guy does not know what he is talking about. In a section of his website he is quoted as saying...

"The 10-print strategy, Mr. Cherry said, is a "technical nightmare that will produce many Brandon Mayfields."

refering to the FBI change from 2 fingers to 10 in the US VISIT program.

And once you start citing Wikipedia you know something is off.

It seems he is trying to say how bad the current system is so his company can sell you his latest and greatest fingerprint software and thus increase his market share. Doesnt it seem odd that a company that makes fingerprint recognition software and keyboards would make it seem so unreliable? You know you need the latest Cherry fingerprint reader for your Mac because it uses Henry Clasification 2.0.

"A little learning is a dangerous thing/ Drink deep or taste not the Pierian spring/ There, shallow draughts intoxicate the brain/ While drinking largely sobers us again."

There must be some drunk fools at Cherry Biometrics!

Steve Everist wrote: Josh, I'll be interested to hear David Ashbaugh's feedback on this article and his review of it.

Steve, I'll be sure and post his response here in this topic or I may have to create another as it will be two more months before I have a chance to speak with him.

L.J.Steele wrote:
I've gotten good ideas and insights from various journals, but the key is the underlying sources, not the author's opinion (unless the author is someone very prominent in that specific field).

L.J., I agree with what you stated about "underlying sources, not the author's opinion" as to what you give more credence to. My question again is if Mr. Cherry wishes to denounce fingerprint identification as a science now and revert to the science of fingerprints circa late 19th century, isn't that going backwards? We have the ability to perform PCR on LCN DNA (low copy number DNA-esentially a single cell is all that it takes now for DNA amplification and replication) but because there is a mathematical probability that it's possible that in a 1 in 16 quintillion chance someone else's DNA was amplified, we should just go back to ABO blood-typing and call it good? Of course not. We need to improve our science as technology allows us to be able to do so.

Without any sources to back up Cherry's assertations that the FBI's fingerprint matchers (computers) "have significant calibration issues", I can't find any truth to his statements. I just returned from a training class at the FBI/CJIS complex in Clarksburg, WV where the IAFIS computers are kept and never once did I hear any examiner/supervisor mention that their computers were not calibrated correctly. I'm betting that if there were problems (especially post-Mayfield) we would be notified immediately. I know that technology has allowed me to perform and achieve AFIS hits on latent print searches that were ran on an older version of the AFIS software with no identifications being made at that time.

Cherry fails to provide the images of his "alleged" serial killer's fingerprints before and after each arrest in which the IAFIS computer failed to match the two together. The workflow for these systems relies on several portions of a fingerprint card (fingerprint images, name checks, etc.) and again, I don't see a technical work manual for the IAFIS computers that specify that once the index fingers are compared, the remaining fingers aren't even looked at. I would like to see those sources.

L.J.Steele wrote:
Many of these magazines are looking for potential authors -- if you want to write a better guide for attorneys and judges on fingerprint issues, check out the magazines' submission guidelines and send something in.

Instead of submitting an article to Judicature which has already written and published in respected and Daubert-proven journals (JFI, JFS, etc.), judges and lawyers should be subscribing to those very journals. I guess I look at it like an article (topic) published in the Encyclopedia Britannica versus Wikipedia. I'm gonna have to give my vote on EB every time as being the more reliable source.

Enough for now, eh?

josher89 wrote:My question again is if Mr. Cherry wishes to denounce fingerprint identification as a science now and revert to the science of fingerprints circa late 19th century, isn't that going backwards?

Remember that the target audience for most legal articles is judges and attorneys who rarely encounter the described-problem. The folks who deal with it routinely generally don't need the overview. (Tho one can still find interesting stuff in overview articles -- particularly if they mention cases and sources one from a jurisdiction or journal or perspective one might not normally review.)

I'll leave the technical discussion of the merits of Cherry's article to you folks -- you are the ones with the technical expertise. As an outsider to the field, I'd like to see a significant number of sources for factual claims and for conclusions that strongly challenge the prevailing wisdom, whatever the topic.

josher89 wrote:Instead of submitting an article to Judicature which has already written and published in respected and Daubert-proven journals (JFI, JFS, etc.), judges and lawyers should be subscribing to those very journals.

There's a vast difference in target audience, and thus the length and tone of the articles. The scientific journals are generally targeted to experts in their respective fields. They assume a certain core of information and tend towards a fair amount of jargon. For a typical judge or attorney who does not practice full-time in criminal law, there's not enough applicable to their practice to justify their time. Legal trade journals are targeted to busy attorneys and judges, generally to members of the organization publishing it. The articles are there as introductions to a topic. To some extent they serve to alert the reader that there's a potential issue. If he or she encounters the topic some time down the road, it may trigger a memory and get the reader to do some research or pull the article out of a folder.

Heck, I do a fair amount of forensics discussions in cases, and JFS, JFI, AFTE J, etc. aren't cost-effective for me to subscribe to. When a question comes up, I go to the library of a local college that has a good collection of journals and treatises. (Usually, if I'm there on any topic, I'll take a glance thru the most recent issues of key journals as well, just to keep up with the area.) I often wonder if, when I cite to the journals whether the judges or opposing counsel bother to pull the underlying articles. I've only had one opposing counsel call to ask where to find a moderately obscure cite -- don't know if the others have better library access, or just don't bother.

Interesting philosopy. Read questionable material to find "reliable" sources of background information while ignoring the author's claims? "Dumb down" the discussion of forensic information so judges and lawyers can understand it quickly and easily? Then what? Have them decide the outcome of cases based on half-truths and mis-understanding of technical issues? "Balance" dedicated professional analyst testimony with defense experts paid to obfuscate the facts and findings?

Somewhere, someone must be trusted to be accurate, fair, honest and ethical. If you give me a list I know who I'd ignore.

*******
"This is not a court of justice, young man, this is a court of law!" - Oliver Wendell Holmes

mdavis wrote:Interesting philosopy. Read questionable material to find "reliable" sources of background information while ignoring the author's claims? "Dumb down" the discussion of forensic information so judges and lawyers can understand it quickly and easily? Then what? Have them decide the outcome of cases based on half-truths and mis-understanding of technical issues? "Balance" dedicated professional analyst testimony with defense experts paid to obfuscate the facts and findings?

Now Davis, that isn't what I said.

I'd tend to judge any such article on the strength of its reasoning and, particularly, the sources it cites. Where any author is making a suggestion that's strongly outside the accepted view, I tend to be skeptical, in part because if I wanted to use the theory, I'd have to pursuade a very skeptical judge who's going to want peer-reviewed publications not citation to a legal trade journal. I wouldn't expect any judge or any attorney to make a decision based on something they read in any of the legal trade journals -- any more than I'd expect someone to make a ruling based on an article in the New York Times, Newsweek, or the Economist.

I don't think of providing a readable, interesting, accurate, well-supported summary of a complex topic for someone who is not familiar with the details, preferably with a guide to where the reader can go to find out more, as "dumbing down". The reader who needs more information should be able to use a well-written introductory article as a way to find key treatises and articles.

The more technical discussions about the details can and should be discussed in as much detail as they require in JFS, JFI, and their ilk.

And in court, the decisions are going to be made on the basis of expert testimony, the evidence itself, and, in some cases, on legal memoranda that cite to, or attach, the relevant treatises and peer-reviewed articles.

I realized late last night where Davis and I may be miscommunicating.

I'm writing about articles about forensic topics in the legal trade magazines in general, not about the merits or problems with the specific article under discussion. I'm merely trying to put it in the context of how this kind of article might be used by its target audience.