Hacker News Reader

Comment by el_nahual | Hacker News Reader

That number doesn't seem right to me so I followed the wikipedia citation. It cites "The Innocence Project", which actually says:

> the few studies that have been done estimate that between 2.3% and 5% of all prisoners in the U.S. are innocent (for context, if just 1% of all prisoners are innocent, that would mean that more than 20,000 innocent people are in prison).

https://innocenceproject.org/contact/#:~:text=How%20many%20i...).

View on HN

The 1% actually cites the estimation method (extrapolation from DNA exonerations). The higher 2.3-5% doesn’t. What studies are those?

https://innocenceproject.org/how-many-innocent-people-are-in...

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=996629 , is the 2.3%

https://www.researchgate.net/publication/228207991_Innocents... is the 5%.

And for context: both of those numbers are based on studying proven false convictions of people given the death penalty, which of course is a) supposed to have a high bar of evidence to convict in the first place and b) does have an incredibly high bar of both evidence and legal argument to be proven innocent after conviction. Since the majority of people convicted of crimes in the USA didn't even go to trial, it is realistically impossible that this number applies to all people in prison - the percentage of false overall convictions must be higher than the percentage of people who were sentenced to death and then proved themselves innocent, given the different pressures and levels of scrutiny on different crimes.

(cite for not going to trial: https://www.nbcnews.com/think/opinion/prisons-are-packed-bec...)

Whoops – https://papers.ssrn.com/sol3/papers.cfm?abstract_id=931454 is a better source for the second one.

Note that the false positive rate for convictions won't necessarily match the percentage of prisoners who are innocent. The latter depends on both the rate of false conviction and the distribution of sentence lengths.

Reply on news.ycombinator.com