Bite marks, blood splatters, polygraphs, hair follicles … these are some of the forensic tools used as evidence by police and prosecutors to convince juries of defendants’ guilt. The scientific validity of these tools has also been questioned, as ProPublica reports in a recent investigation focused on yet another, little-known forensics technique: 911 call analysis, in which police attempt to “divine truth and deception—and guilt and innocence—from the word choice, cadence, and even grammar of people reporting emergencies," writes Brett Murphy. Such evidence was used to help convict Jessica Logan, a young mother in Decatur, Illinois, who called 911 at 3am in 2019 to report that her 19-month-old son was not breathing.
According to ProPublica, training in 911 call analysis is provided primarily by Tracy Harpster, a retired police chief who sells the eight-hour training to police departments for $3,500 a pop. Participants get a certificate of completion and a checklist designed to track certain types of errors guilty people allegedly make on emergency calls. The detective assigned to Logan’s case had just completed the course a couple months before. ProPublica discovered that numerous researchers have tested whether Harpster’s indicators correlate with guilt, and they’ve “consistently found no such relationship,” yet use of the method persists. Logan was convicted of first-degree murder and sentenced to 33 years. ProPublica found holes in the case, not only with the 911 analysis but also the coroner’s report. This story is the first in a ProPublica series on 911 call analysis. Read it here. (More forensics stories.)