A recent story by NPR highlighted some issues with the National Integrated Ballistics Information Network (NIBIN), a huge ballistics database operated by the U.S. Bureau of Alcohol, Tobacco, Firearms and Explosives. It contains high-resolution images of the marks left on bullets and shell casings fired by specific firearms. The idea is to allow users to match those bullets and casings to each other and, sometimes, to match them to a specific gun.

The problem isn’t with the database itself. It is with the scientific underpinnings of ballistics evidence in general. In the NPR story, a police officer and an assistant U.S. Attorney comment that “the ejector mark is the fingerprint of the shell casing,” and that “each firearm that shoots a bullet leaves an imprint that’s unique to that firearm.” That’s the theory — but the theory isn’t backed up by science.

The founder of the nonprofit Forensics Justice Project, an attorney, challenges the idea that the marks left on bullets and shell casings are unique enough to be used as evidence.

“The problem is no one’s gone out and actually determined that it could only be matched to that gun to the exclusion of all other guns in the universe,” she points out.

She also points out that false or exaggerated ballistics and firearms evidence has resulted in exonerations.

In one of them, a 2013 death row case in Mississippi, the FBI acknowledged the scientific uncertainty. The agency sent a note to the district attorney explaining that “the science regarding firearms examinations does not permit examiner testimony that a specific gun fired a specific bullet to the exclusion of all other guns in the world” and adding that any claim of such certainty was “not supported by scientific standards.”

Indeed, ballistics is not the only type of forensic evidence that lacks scientific underpinnings. Several forensic techniques have been called into question in recent years, including microscopic hair comparison, bite mark analysis, blood spatter analysis, shoe print and tire track analysis, and even fingerprinting.

Most of these techniques are commonly used to convict criminal defendants, but the Justice Department and FBI have now admitted that microscopic hair analysis testimony, at the least, has been biased toward prosecutors in over 95 percent of cases examined.

Yet despite the uncertainty over the reliability of ballistics, the NIBIN and technology allowing for ballistics tests are coming to local police departments all over the country through Justice Department grants.

It’s crucial for police, prosecutors, defense attorneys and the public to understand the limitations of ballistics and other forensic techniques. Real life isn’t as straightforward as “CSI.”