Junk Evidence

In 2012, DNA testing showed that the hairs the FBI once said "matched perfectly" in a Washington murder case came from several people, and one of them from a dog. The man convicted on that testimony had already served 25 years.

Junk Evidence is the story of how that happened, and how often. For most of the twentieth century, American juries heard forensic experts state conclusions with a confidence no one had ever measured. Bite marks, hair comparisons, arson "pour patterns," fingerprint certainty, bullet-lead analysis, roadside drug kits and shaken-baby diagnoses were presented as science. When researchers finally began testing them, much of that certainty turned out to have been assumed rather than earned.

Across fifteen chapters, the book follows the cases that exposed the problem, from the Corsicana house fire to the Madrid fingerprint error, the Boston crime-lab scandal and the confession that was not true. It explains in plain language what each method can and cannot show, why jurors believed it, and what the 2009 National Academy of Sciences report and the Texas experiment changed. It also counts the cost in lives and asks what a reformed lab actually looks like.

This is a work of nonfiction built on court records, government reports and published research, with sources listed at the end. If you follow true crime, criminal justice reform or the Innocence Project's work, or simply want to understand why a confident expert in a witness chair is not the same as proof, this book is for you.