Editor's Note: Shortly after this article was first published, Dr. Clifford Spiegelman passed away on May 14, 2020. The author would like to note that the brief time he spent with Dr. Spiegelman was spent with a man who was unflinchingly loyal to the truth and who practiced statistics as a means of discovering that truth. The Spirit magazine editor board and the Texas A&M Foundation extend their prayers and condolences to Dr. Spiegelman’s family and loved ones.
Imagine that you are part of a jury overseeing a murder trial, and the defendant swears he is innocent. The prosecutor calls to the witness stand a forensics expert with more than 30 years of experience in crime laboratories running tests for cases just like this one. The expert brings the jury’s attention to bullet casings found at the crime scene, specifically the tiny marks engraved on the back of each casing. He explains that these are called “bunter marks,” and they are usually impressed on each bullet by the manufacturer during production.
Dr. Clifford Spiegelman's expertise was used in evaluating the effectiveness of compartive bullet lead analysis, a forensic method first used in the investigation of President John F. Kennedy’s assassination.
The expert then reveals that, after rigorous testing, the bunter marks on the casings at the crime scene match almost exactly to marks from a box of cartridges found in the defendant’s home. Murmurs spread throughout the courtroom. When pressed about his confidence in the findings, the expert claims he can match the two bullets’ marks with an error rate of less than 1%. Jury members exchange looks. Certainly, this can’t bode well for the defendant.
But despite what the prosecutor would have you believe, Dr. Clifford Spiegelman says this kind of evidence is statistically meaningless. “The same bunter marks can appear on hundreds of thousands, if not millions, of bullets,” he said. Because the same bunter tools are commonly used for large quantities of bullets, matching bunter marks cannot be statistically relied upon to track bullets back to the same box. “Trial evidence will focus on whether or not they can tell it’s the same bunter mark on two different bullets, but statistically, it doesn’t matter if they can.”
Pro Bono Statistician
A distinguished professor of statistics in the Texas A&M University College of Science, Spiegelman has used statistics to address issues in unexpected fields for more than 40 years. He co-wrote a leading textbook in transportation statistics, was named the official statistician of the Texas Holocaust and Genocide Commission, and helped found the interdisciplinary field of chemometrics, which uses statistical methods to better understand chemical data.
Spiegelman’s most publicized work, however, is in forensic science. He is a key statistical adviser to the City of Houston’s crime laboratory. Working with the Innocence Project, a free nonprofit legal organization for wrongly convicted people, he has testified pro bono on cases in which shaky forensic methods, like bunter mark analysis, were being used against innocent defendants.
“People like flashy science,” Spiegelman said. “But doing simple science well is very hard.” In the bunter mark example, for instance, what mattered to the case was not that the bullets had the same bunter mark, but whether that information was statistically meaningful in proving or disproving the defendant’s innocence.
Many similarly flawed forensic methods are still used and treated as iron-clad evidence in criminal courts, motivating Spiegelman to research different methods’ validity. In 2007, his expertise was used to help evaluate the effectiveness of comparative bullet lead analysis, a forensic method first used in the investigation of President John F. Kennedy’s assassination. He and five other researchers assessed the method the best way they knew how: by investigating its use following the JFK assassination itself.