Gallery

Let’s Talk About FINGERPRINTS

Fingerprints, more technically known as “friction ridge analysis”, have been around for well over 100 years; at least since Francis Galton published his seminal work simply titled “Finger Prints” in London in 1892.  In fact, fingerprints are still analyzed according to their “Galton characteristics”, shown below.

In today’s fingerprint jargon, these are called “minutiae”, but they are still the same features that were cataloged by Galton in 1892.  These “minutiae” are meaningful only in relation to their relative position within the swirls and whorls of the overall friction ridges comprising the fingerprint.  Fingerprints that are forensically obtained from crime scene objects and weapon surfaces are called “latent prints”.  There is a level of fingerprint detail even finer than “minutiae” which consists of very small irregularities in common features, but it is my belief that these are rarely seen in latent prints.

Now …. did you know that it has never been statistically validated that fingerprints are unique?  And what’s more, I don’t believe it ever will be.  The experimental population required to statistically certify uniqueness renders the “experiment” totally impractical.  Given sufficient data, it might (I repeat – might) be possible to develop statistics for “probability of occurrence”, but no such data exists.  [This is where DNA sets itself apart from all other forensic disciplines – it yields a hard statistic for probability of a match or non-match.]

However, over 100 years of anecdotal data, empirical observations, and folklore have bestowed upon “fingerprints” the mantle of infallibility within the justice system.  Fingerprints have been responsible for more criminal identifications, by far, than any other forensic discipline.  But are they infallible?  Not so.  Read on.

About AFIS.  In 1999, the FBI (US Federal Bureau of Investigation) launched the computer-based AFIS (Automated Fingerprint Identification System), which has since been superceded by IAFIS (Integrated Automatic Fingerprint Identification System).  The FBI is now also working on NGI (Next Generation Identification) which will include palm prints, iris, and facial identification data.  IAFIS currently has over 120 million peoples’ prints on file, and it processed 61 million 10-print inquiries in 2010.  This has been a boon to law enforcement, because it would clearly be impossible for human examiners to conduct 61 million fingerprint comparisons in a year.  However, this efficiency comes with a price.  Automated fingerprint matches are much more prone to error than those done by humans, and automated matches need to be confirmed by competent human examiners.  But even the combination of automated match with human verification can produce errors.  Just ask Brandon Mayfield.  Brandon Mayfield is a Seattle, WA attorney whose prints were matched by AFIS to a partial print from the Madrid train bombing, and this was confirmed by human FBI experts.  The FBI called it a “slam dunk”, “bingo” match, and Mr. Mayfield was taken into custody.  But guess what – it turned out that the print was actually from a man named Ouhnane Daoud, an Algerian.  And by the way, Mr. Mayfield was awarded a $2 million settlement.

Fingerprint identification is plagued by the problem of “partials”.  A partial print is an area of friction ridge pattern that is smaller than the “whole” print.  When people are “fingerprinted”, a record is made of the entire friction ridge surfaces of the fingers.  It is not common for a forensic latent print to be anything more than a partial, and this creates a problem with “matching accuracy”, because the examiner has to try to match a small piece of a print to the whole-print exemplar.  And the more “partial” the latent print is, the higher the error rate climbs.  In addition, latent prints are frequently “smudged”, and because skin is elastic it can expand and compress and twist, distorting the latent print.  The number of matching minutiae that is generally accepted for declaring a “match” hovers around 14, but matches on partials have been declared with as low as 8.  The only figure I’ve seen published for “false positives” from partials is 1%-4% (I don’t have attribution for this).  False positives from partials gets worse for automated matches.

Still another issue is that fingerprint matching accuracy is dependent upon the competence of individual examiners.  In 1995, the Collaborative Testing Service administered an examiners’ proficiency test that was pre-approved by the International Association for Identification (IAI).  The results were less than good. Four suspect exemplar cards with prints of all ten fingers were provided to test takers along with seven latents.  Of 156 examiners taking the test, only 68 (44%) correctly classified all seven latents. In total, the test results contained 48 incorrect identifications.

In 2007, a Baltimore County (MD) judge ruled that fingerprints were not reliable enough to be used as evidence against a defendant facing a possible death sentence.

Now, even having said all this, the bottom line is – finger prints are “pretty good” (even though there is NO data from which to compute a probability of occurrence).  But are they infallible?  No.

I also strongly encourage you to read the companion post by Martin Yant titled: “How to reduce fingerprint errors.”

It’s difficult to do this topic justice in a blog post, so here is some reference material for you:

1)  SWIRLS AND WHORLS: LITIGATING POST-CONVICTION CLAIMS OF FINGERPRINT MISIDENTIFICATION AFTER THE NAS REPORT, Prof. Jackie McMurtrie, University of Washington School of Law.

Swirls & Whorls

2)  MORE THAN ZERO: ACCOUNTING FOR ERROR IN LATENT FINGERPRINT IDENTIFICATION, Prof. Simon Cole, University of California at Irvine.

More Than Zero

3)   THE ACHILLES’ HEEL OF FINGERPRINTS, Jennifer Mnookin, The Washington Post.

The Achilles Heel of Fingerprints

4)  MOTION TO EXCLUDE EXPERT TESTIMONY ON FRICTION RIDGE ANALYSIS, OR, IN THE ALTERNATIVE, TO CURTAIL SUCH TESTIMONY UNDER DAUBERT V. MERRELL DOW PHARMACEUTICALS, The Innocence Project of New Orleans.    [Editor’s Note:  If interested, the IPNO also has Daubert motion templates for “firearms identification” and “microscopic hair comparison”.]  This is very well done, but we’ll have to see how it does in court.

IPNO – Motion to exclude friction ridge analysis

5)  And if you’re really, really into this, here’s the FBI Guide on fingerprint processing.

FBI Fingerprint Processing Guide

Phil Locke

3 responses to “Let’s Talk About FINGERPRINTS

  1. Pingback: PBS Nova – “Forensics on Trial” – a Review | Wrongful Convictions Blog

  2. Pingback: Why I Think the US Justice System is Broken – and Why It’s Not Getting Fixed | Wrongful Convictions Blog

  3. sir face construction can be done by fingerprints.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s