CREDIT: MAKI NARO
When courtroom science goes wrong — and how stats can fix it
COMIC: Bite marks, shoe prints, crime-scene fibers: Matches to suspects are often far shakier than courtroom experts claim. Better statistical methods — among them, a little beast known as the “likelihood ratio” — can cut down on wrong convictions.
TAKE A DEEPER DIVE | Explore Related Articles from Annual Reviews
Statistical Issues in Forensic Science
A 2009 National Research Council report raised concerns over how forensic data are analyzed, interpreted and used in court. Hal Stern explains how more rigorous statistical methods such as the likelihood ratio could improve matters.