Fingerprints are something that have long been used to identify us all as unique individuals. Even identical twins are usually found to have different sets of fingerprints. For the majority of us, they are something mostly taken when you get a driver’s license or identification card in order to have them on record. For others, they are used to tie them back to some sort of event or crime that happened, where fingerprints come up as matches in a database of all fingerprints taken. Just how does this work and how accurate is fingerprinting in solving crimes and placing a person at a crime scene? Read on to find out more.
Fingerprinting has been used as a widespread identification tool in the US since the 1920s. It’s been used as the only hard evidence used to execute people believed to be connected with crimes and until DNA testing came along was one of the most relied on “scientific” data pieces that would help convict people. In the last 10 – 15 years, there’s been a lot of debate about the just how calculated and precise fingerprinting techniques are, hence, scientific being in quotations above. In the last 10 years, there have been at least five people wrongfully identified through fingerprints, four of them ended up behind bars. Most famously was Brandon Mayfield, an Oregon lawyer who spent two weeks in jail in 2004 because three FBI experts matched his prints with evidence found in connection with the Madrid train bombings. Luckily, Spanish authorities continued to try to match the prints after Mayfield was arrested and eventually they linked the prints to an Algerian man which then allowed Mayfield to be released and left him acquitted of charges. Defendants rarely have fingerprints double or triple checked with multiple agencies to ensure the accuracy of fingerprint evidence, which means there could be many more undiscovered mistakes.
Perhaps what’s most interesting about fingerprinting still holding up as hard evidence in court is that the lack of scientific statistical studies of the accuracy of latent fingerprint identification. Latent fingerprint identification is performed by human examiners and typically most human pattern recognition abilities and algorithms have significant false negative and false positive errors. There are automatic recognition programs know as AFIS (Automated Fingerprint Identification System) but the advancement in technology regarding accuracy of these programs has not been enough to surpass human beings – pretty unusual these days. For the most part, humans are able to recognize fingerprint patterns better than automatic methods based on mathematical, statistical, or scientific methods.
The fact that fingerprints are examined and determined by human eyes and not anything more scientific is interesting for a couple reasons. First, scientists, attorneys and juries often don’t put much validity into eyewitness testimony, but identifying fingerprints applies the same logic – it’s examination and testimony from what humans know to be true. Second, human pattern recognition abilities are imperfect. Just as eyewitnesses can misidentify people and objects, human fingerprint examiners almost certainly have flawed abilities and different approaches to reviewing fingerprints. Sometimes to the point where they don’t actually always agree with each other on correctly identifying fingerprints.
In addition to human error in identifying finger prints, some people do not actually have finger prints. The ridges on fingers that typically stand out in finger prints are very shallow, making it hard to get an accurate fingerprint read by just ridges. Other times people who have a lot of wear and tear on their hands or scarring can have altered fingerprints. These types of situations make it especially hard for automatic programs to recognize prints, but it also makes it hard for human fingerprint identifiers to recognize them.
In the Mayfield case, an FBI laboratory review committee evaluated the scientific basis of the fingerprint identification and recommended the investigators do more scientific research including a study of the accuracy of latent fingerprint examiners. The National Research Council has also called for a need to do better evaluations of fingerprint examination decisions in a study in 2009. Since then, the FBI recently published a report on such a study in the Proceeding of the National Academy of Sciences (Accuracy and reliability of forensic latent fingerprint decisions, PNAS, April 25, 2011). This study found a 0.1% false positive rate and a 7.5% false negative rate. What that’s like it saying two plus two is 3.999 plus or minus 0.012. But it isn’t!
What does this all mean? Fingerprinting has a long way to go in terms of accuracy but has now fallen to the wayside because DNA profiling has come around. Fingerprinting is now becoming a thing of the past, or second best as far as hard evidence goes because DNA seems to be the most reliable hard evidence we now have. Perhaps a larger take away is the fact that fingerprint identification has been unchallenged and taken as evidence to convict in many cases over eighty years. This is because of the popular belief that mathematics and science were highly involved in fingerprint identification, when, in fact they were taken as exact answers without performing adequate studies of the accuracy. Despite that two plus two equals four, something so black and white is very rare in the real world. Math teaches us to critically think and factor error into almost everything. With this said, we should almost always demand to know the error rates of numbers and be suspicious of numbers quoted without error rates -especially when freedoms depend on it.
