Tikalon Blog by Dev Gualtieri

Fingerprints

December 4, 2017

As a child, I was a member of the Boy Scouts of America for about two years. This was long before the many current controversies associated with that organization, and my membership was understandable given the rural nature of my home environment in Upstate New York. Our idea of a family vacation wasn't Disneyland, which was too far away (Walt Disney World didn't open until 1971) and too expensive for a middle class family. Instead, we would camp at one of several New York State Parks.

Scouts could earn "merit badges" by demonstrating skills in selected subject areas. There was electricity, which was an easy topic for me, and there was also fingerprinting. I was able to visit the local police station to see how it was done, and I had my own fingerprints taken, the card on which they were imprinted now long since lost. Later, I needed to have my fingerprints taken in earnest when I got a government security clearance, long since lapsed.

To paraphrase General Douglas MacArthur, who famously said that "...old soldiers never die; they just fade away," old boy scouts never die, they just end up looking like this.

I spent two weeks during one summer at a Boy Scout camp, where we started campfires by striking flint with steel, and planted some fir seedlings.

(Modified Wikimedia Commons image

The common conception of fingerprints is that they uniquely identify a person, and many television shows perpetuate this idea. During the Elliot Ness era, career criminals supposedly defaced their fingerprints with acid so they couldn't be traced to their crimes. There's also the saying when identifying an action that a person's "fingerprints are all over it." Some computer devices will even image a person's fingerprint as a password to unlock a device, the idea being that only that person's fingerprint is valid.

Computer techniques have advanced fingerprint matching to a high degree, greatly enhancing the probability that there will be a "match" found to some person. The US Federal Bureau of Investigation has developed an Integrated Automated Fingerprint Identification System that's been highlighted on various television crime shows; notably, NCIS. The IAFIS contains fingerprints for more than a hundred million subjects, mine being mixed in there, somewhere. This system is routinely searched by various law enforcement agencies in attempts to match latent (crime scene) fingerprints to individuals.

Fingerprints in electronics meant an entirely different thing a decade ago than it does today.

(Fingerprint image via Wikimedia Commons. Circuit image by the author.)

Is there a scientific basis to qualify fingerprinting, as practiced, as an unerring method of identification? Quite a few other "forensic science" techniques have received criticism as not being as precise as claimed. As one example, polygraph (lie detector) testing has been called a pseudoscience.[1] What's especially interesting is the US government's admission that a polygraph test can easily be defeated in its prosecution of coaching services that provided information on techniques to do this.[2] A polygraph in no way passes the same performance standards that I demand of my analytical instruments.

While the scientific case for polygraph testing is weak, the emerging methods of forensic DNA profiling, first introduced as a law enforcement technique in the early 1980s, are on firmer scientific ground.[3] DNA profiling can be a valid forensic tool, provided, however, that collection techniques are adequate, and the methods for statistical analysis of the DNA data are justified. Several prominent criminal cases have illustrated insufficient technical protocols, faulty interpretation of DNA evidence, or the reliance on closed source software for analysis.[3]

The emerging scientific consensus on fingerprint identification, the results of which are considered absolutely certain in the U.S. judicial system, is that such trust is not warranted.[4] A presentation at this year's 69th Annual Scientific Meeting of the American Academy of Forensic Sciences, February 13-18 in New Orleans, claimed that fingerprint analysis should be considered as circumstantial evidence, only.[4]

Author of the presentation, "Fingerprint Science," Carnegie Mellon University emeritus professor of statistics, Joseph B. Kadane, says

"Fingerprints can be useful for eliminating suspects or linking crimes together, but as an identification tool, they are not as infallible as TV shows and judges and jurors are made to believe... Fingerprint analysts are exaggerating what they can do, and policies need to be changed to prevent this from happening. Courts need to stop accepting fingerprint identifications - because they have no scientific basis."[4]

Joseph Kadane at the 2006 Morris H. DeGroot Memorial Lecture

Joseph B. Kadane at the 2006 Morris H. DeGroot Memorial Lecture.

(Still image from a Carnegie Mellon University video.

Fingerprint analysis was the topic of a huge, 160 page report, "Forensic Science Assessments: A Quality and Gap Analysis - Latent Fingerprint Examination," prepared by the American Association for the Advancement of Science.[5-7] This report states that courtroom testimony that fingerprints collected from a crime scene belong to a single person is indefensible and lacks scientific foundation.[7] Says Kaldane, who was also an author of this report,

"Our review of the scientific literature found that there is no scientific way to estimate the number of people in some community - a city, a state, the country, the world - who share the characteristics found, and hence no scientific basis for identification."[7]

The public holds misconceptions about latent fingerprint examination, and this is sometimes called "The CSI Effect," after the popular television crime series.[7] The scenes for that series were likely shaped by the many years of overstatement by latent print examiners. The report urges latent fingerprint examiners to include specific caveats in their reports, including an acknowledgement that their conclusions are opinions rather than facts.[5]

Fingerprint analysis does have some science behind it. First, although there has not been a scientific assessment of the rarity of any particular feature, or set of features, in a fingerprint, the ridge patterns on human fingers vary greatly among individuals.[5] Furthermore, fingerprints change very little over the life of an individual; and, when the prints from all ten fingers are considered, a specific individual can be identified to high likelihood.[5-6] Extrapolating from this basis, latent fingerprint examiners have made the overblown claim of 100% identification. While latent prints can exclude the preponderance of individuals, there is no scientific basis to identify a single person.[5]

Another concern is automated fingerprint identification software, all of which is proprietary so the source code cannot be examined to determine its assumptions and accuracy.[5] Also, these systems can't match a latent crime scene fingerprint to another from a known source.[6] They can only identify similar features.[6] The report does state that "It is possible that automated fingerprint identification systems could evolve over time."[6]

as every scientist knows, the conclusion of every report is to recommend more research. This report concludes that there are insufficient scientific data on the potential pool of human fingerprints to determine how many people might display similar features.[6] The report further proposes research to determine whether latent fingerprint analysis is strong enough to be used as evidence.[6] As stated in the report,

"Uncertainty in this area means there is an inadequate scientific foundation for determining how many features, or what types, are needed in order for an examiner to draw definitive conclusions about whether a latent print was made by a given individual and hence... examiners should not draw such conclusions."[6]

Fingerprints

References: