Monday, January 7, 2019

Junk science and guns

Are the ejector marks from a handgun unique?

How many handguns would you have to fire in order to make that conclusion?

The Bureau of Alcohol, Tobacco, Firearms and Explosives would like you to believe the answer to the first question is yes and that the answer to the second question is irrelevant.

Back in 1999 the BATFE (known then as the ATF) created the National Integrated Ballistics Information Network (NIBIN), a database of high resolution images of bullet casings. The database was used by forensic examiners to testify at trial regarding the likelihood of a particular gun being used in a crime. Thanks to the Department of Justice, local police departments now have access to the database for use in investigating gun crimes.

But has anyone bothered to verify that the assumption that every gun leaves a unique mark on a bullet casing accurate?

This is the problem we run into with every new "forensic tool" created by law enforcement. We've seen it with tire tread analysis, bullet composition, bite mark evidence, and others. A forensic "scientist" comes up with a theory - and without checking it using the scientific method - declares that this new tool will allow the police to catch the bad guys.

But what tends to happen is the police take whatever "evidence" they uncover and use it subjectively to create a narrative that their leading suspect is the bad guy - even if it leads them in the completely wrong direction.

If you've read The Cadaver King and the Country Dentist by Radley Balko and Tucker Carrington then you know where I'm headed. If you haven't read it, then stop what you're doing right now and get the book. In many investigations the police hone in on the person they think is the most likely culprit - often on little more than a hunch. They look for evidence that points to their suspect and ignore evidence that contradicts their narrative (one of the reasons that Brady material is withheld from defense attorneys). And if the narrative makes little or no sense, they bring in a forensic examiner who will gladly resort to pseudo-scientific methods to create evidence pinning the blame on the guy the police were looking at in the first place.

Just think about it, soon the local authorities will take a bullet casing and will compare the marks on it to the NIBIN database. As soon as the algorithm matches the bullet casing the police recovered with an image in the database, the police will go after the owner of that gun -- never once questioning the assumption that no other gun would leave the same mark.
In 2013 a Mississippi man's life was spared hours before his scheduled execution after the FBI said experts had overstated the science. In a note sent to the district attorney in that case, the bureau clarified 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."
Should the case go to trial the court, which long ago ceded its gatekeeping role with regard to scientific evidence will allow the gun examiner to testify to the unique characteristics of the ejector marks without once questioning whether there is any science behind the assertion. Of course this might work out differently in civil court where judges are only too happy to exclude any new scientific evidence that might result in insurance companies having to pay off claims.

Every new forensic tool (for lack of a better phrase) wants to portray itself as some type of a "fngerprint" since we have all bought into the assumption that our fingerprints are unique. While fingerprints and DNA profiles may very well be unique to an individual, there is little or no evidence that anything else is. Tire treads and shoe soles might have unique wear patterns, but the surface in which the imprint is left and the means by which that imprint is transferred to the surface make the analysis anything but precise.

Just remember what they say about assumptions.

No comments: