Courtroom and law enforcement television have created a public perception that fingerprinting is an infallible science. While the patterns for each human finger are formed in the womb, it is important to be careful about language that implies absolute certainty.
Looking at the Numbers
The FBI has famously estimated that fingerprint identification is 99.8 percent accurate. This level of accuracy is reassuring when you have a murder case or burglary with a short list of known suspects. It’s less inspirational when you consider the billions of people currently alive and the computer databases that can search through thousands of arrest records. Our conversations often take fingerprints and snowflakes as scientific examples of uniqueness, and yet even the snowflake cliché has its limits. If a researcher catalogues enough snowflake patterns under a microscope, eventually she will find snowflakes that are indistinguishable from that level of magnification. On a practical level, there are only so many possible variations of geometrical flakes or whorled lines. After a certain point, the subtle variations are easy to miss.
The television shows often proclaim, “It’s a match!” or something comparably certain. When you say “possible match,” the television audience starts to get bored. “Why can’t these analysts just use plain language and say that we got the bad guy?” Unfortunately, matching evidence to suspects is rarely certain beyond a doubt. Even eyewitnesses can be completely wrong in their description or identification of a suspect. When it comes to fingerprinting in a crime scene, the evidence often consists largely of partial or smudged prints. Few criminals have the courtesy to dip their fingers in ink and then firmly roll each digit immediately after the crime. The investigators use what they can find on glasses and other hard surfaces. In a relatively public space like a hotel room, there may be several other sets of prints from people who just happened to use the same space.
Databases and Forensics
The term “forensics” has come to be associated largely with courtrooms and detectives, but it is also a term for debate and argumentation. By this secondary definition, anything that is forensic should also be debatable. Fingerprinting is an excellent tool for narrowing down suspects and adding to other evidence, but the public and jurors should be careful about cases that rest almost entirely on this one type of evidence. Even a 0.2 percent rate of failure is problematic when you consider the number of potential matches stored in national databases like America’s Integrated Automated Fingerprint Identification System (IAFIS). The FBI uses the IAFIS, and it contains over 70 million sets. This database contains more than just prints collected at local police departments during bookings. It also includes sets that have been submitted for background checks. When an employer submits the prints of job applicants to the FBI for a background check, those sets are added to the same database with criminal records.
For employers, fingerprinting is still a great way to supplement other types of background check. It’s just important for everyone to realize the chance of error when too much depends on a single form of forensic evidence.