NUMBERS AT THE CRIME SCENE

A forensic number can be accurate and still support the wrong conclusion.
Numbers at the Crime Scene shows readers how to examine probabilities, error rates, database matches, confidence statements, and statistical claims before treating them as proof. Through clear explanations, original diagrams, data tables, audit tools, and carefully bounded examples, André Kologeski separates what a result measures from what it cannot establish.
Readers learn why base rates change predictive value, why a likelihood ratio is not a probability of guilt, how inconclusive results alter error-rate denominators, when evidence may not be independent, and how charts can guide attention while hiding assumptions.
Written for investigators, attorneys, students, journalists, and critical readers, this book offers a practical framework for evaluating forensic claims without reducing complex evidence to a single impressive number.
The objective is not to distrust statistics. It is to understand exactly what they say-and when they are being asked to say too much.