[dhh2106]

Great and infuriating article. There's also issues with eyewitnesses testimony and lineups that have been well documented, meaning the rate of wrongful conviction may be quite high.

"The National Registry of Exonerations has documented 553 cases since 1989 in which someone was convicted on false or misleading forensic evidence and later cleared. The growing list of exonerations includes a Texas man whose 1987 murder conviction, based on bite-mark evidence, was thrown out in December, and an Illinois man declared innocent in January in the retrial of a murder case that hinged on dubious ballistics evidence.

But the exonerations likely represent only a fraction of the cases in which faulty forensics sent innocent people to prison, researchers say."

Is there any research estimating how many wrongly convicted people may be incarcerated? How could that even be estimated?

[darepublic]

can we filter by cases with large part of the case resting on forensics? Then foreach such case see if we can double check the forensics evidence

[SllX]

Not without enough people willing to enter the related forensic science labor pools, to make a point.

Even after you’ve filtered for all the cases you could want, you’re probably going to end up with a large number. Rechecking all of the evidence in each case to either reaffirm or exonerate the people in prison because of that evidence is a labor intensive task.

[ChrisSD]

Just a note that eyewitnesses testimony can be very helpful in most investigations, it just has to be handled very careful. However it often isn't. As you say, lineups can be particularly tricky.

I mention this because often on this topics people say "eyewitnesses testimony is always unreliable!" when actually it's a bit more complicated than that.

[hannasanarion]

Saying that something is always unreliable isn't the same as saying it is never correct. Eyewitness testimony is an important part of any prosecution, but it should never be relied on, and juries need to be informed of the unreliability.