College of Liberal Arts and Sciences at Iowa State University

Firing blanks

Study conducted by statistics' Alicia Carriquiry comes to the aid of defense attorneys, not FBI.
Ironically the initial funding for the project came from the Federal Bureau of Investigation.

But once the report by Alicia Carriquiry, professor of statistics and associate provost, and former Iowa State statistics professors Hal Stern and Michael Daniels was delivered to the FBI in 2000, the trio never heard from the law enforcement agency again.

"The FBI wanted to know the probability of matching (unspent) bullets to bullets used to commit a crime," Carriquiry said. "From watching television, criminals know they have to get rid of firearms used to commit a crime. But because bullets or bullet fragments can be recovered from a crime scene it seemed to the FBI possible that they could determine if unspent bullets in the possession of a suspect could be matched to the bullets used in a crime with some high level of confidence.

"We came up with a method for establishing the significance of a match, by focusing on estimating the chance of a random match."

The four major bullet manufacturers in the nation each make bullets basically the same way but use different concentrations of various elements in their lead alloy. Bullets are typically manufactured in lots of about 200,000 to 700,000 depending on caliber. Theoretically trace concentrations of silver, copper, tin and other elements in each of those bullets in a lot would remain the same, so that bullets with a similar chemical signature might be thought to come from the same lot.

That's not necessarily the case, Carriquiry says.

"That's the million dollar question," she says. "The FBI claims yes, but we say that is not known."

Bullets are manufactured from big vats of molten lead. Bullets from one end of the manufacturing cycle of each batch have been shown to have a different composition from bullets at the end of the batch and even in the middle.

"The trace composition data that are available has not allowed us to determine whether the variability in chemical composition across and within batches is large or small, and this information is needed to decide whether a chemical match between two bullets has any probative value or not," Carriquiry said. "It really should not mean much to a jury to be presented with evidence showing a close match between two bullets if it is known that close matches are quite common."

If that is the case, then the probability that the bullets match even though the suspect had nothing to do with the crime would be uncomfortably high and the evidence would not have any probative value Carriquiry says.

In the case of DNA evidence, experts know that from analyzing hundreds of thousands of samples that pretty much means that the same person is the "owner" of that DNA.

"The data that would allow this type of statement for bullet lead have not been collected," Carriquiry said.

Other statistical issues can be factored into the equation. Carriquiry points out that in cases of crimes committed in small communities, it would be possible for several individuals to have purchased bullets from the same store that were manufactured at the same time.

"The possibility exists that one retail outlet in a small town could have received a shipment of bullets that were all packaged at the manufacturer at the same time may be non-negligible," she said. "Chances are that there would be a lot of bullets from the same batch in that same town.

"On the other hand, in a big city the chances of a large proportion of bullets from just one or a few batches would be relatively low. Furthermore, some studies have shown matches of bullets that were manufactured ten years apart."

After Carriquiry, Stern and Daniels submitted their initial report to the FBI in 2000, the information was shelved. But a reporter, under the Freedom of Information Act, requested the report and published the results.

"Attorneys have been challenging this type of evidence in court ever since," Carriquiry said.

She has testified twice on the issue in murder trials in Alaska and Kentucky. She gets almost daily e-mails from a woman whose brother was convicted of murder solely on bullet identification evidence - a box of cartridges he owned had the same chemical fingerprint as bullets found at the murder scene.

A National Academy of Sciences report, commissioned by the FBI, is due out soon on bullet matching. Initial media leaks show that the NAS report agrees with much of what Carriquiry, Stern and Daniels said four years ago.

Carriquiry and David Baldwin of the Ames Laboratory have plans to do their own bullet study.

"This is a very interesting problem," Carriquiry said. "We know that we can match the bullets but what does that really mean - and the answer is we don't know what that means."