> At 100% of the 130 examinations analyzed, seized drug analysis was the examination with the highest percentage of Type 2 errors. It was followed by pediatric physical abuse (83% of cases had forensic errors with 22% having Type 2 errors out of 60 cases analyzed), fire debris not including chemical analysis (78% of cases/38% Type 2 in 45 cases), bitemark comparison (77% of cases/73% Type 2 in 44), pediatric sexual abuse (72% of cases/34% Type 2 in 64), serology (68% of cases/26% Type 2 in 204), shoe/foot impressions (66% of cases/41% Type 2 in 32), DNA (64% of cases/14% Type 2 in 64), hair comparison (59% of cases/20% Type 2 in 143), and blood spatter (58% of cases/27% Type 2 in 33).
Also this sample comes from cases where people were exonerated, which could cluster around poor quality police work in general. And is US-based where policing is all kinds of messed up.
A few years ago i picked up the old (but famous) cases of Brad Cooper and Casey Anthony as they have tons of available digital forensics evidence.
I double-checked the forensics work and found several mistakes in processes, assumptions and technical conclusions. I sent off my findings to people associated with Project Innocence - not because I found anything that proved Cooper or Anthony's innocence, quite the opposite. Instead, I wanted to let them know that forensics experts can make mistakes.
It is interesting that scientific work have fault-finding processes like peer-reviews, but forensics investigations in court cases does not.
except for all of those innocent folks that have had their lives ruined by that 'almost all the time' caveat, it's great!
here's a report[0] that says something like 80% + of criminal forensic work has major mistakes within it.
[0]: https://www.criminallegalnews.org/news/2024/may/15/report-fi...