Both overall cite NRE data. So may be better to go to the source (and the latest version). https://www.law.umich.edu/special/exoneration/Documents/NRE%20Annu

Both overall cite NRE data. So may be better to go to the source (and the latest version). https://www.law.umich.edu/special/exoneration/Documents/NRE%20Annual%20Report%202021.pdf https://www.law.umich.edu/special/exoneration/Documents/NRE%... is the 2021 National registry of exonerations annual report. What you'll see, interestingly, is neither the 2021, nor 2019 report, estimate the prevalence of wrongful conviction directly. In particular, the claim "According to the 2019 annual report by the National Registry of Exonerations, wrongful convictions statistics show that the percentage of wrongful convictions is somewhere between 2% and 10%. " It does not say that. It offers no percent view on the rate of wrongful convictions. It only offers the direct number of exonerations, exonerations broken down by race, by crime type, etc. It points out lots of statistical things (exonerations are not evenly distributed by race, crime, etc) Someone else, with some other (not obviously explained) methodology is using that to calculate wrongful conviction rates. That said, roughly all sources i can find are within the 2-12% rate bound (IE i see a number saying 3-6%, a number saying 1%, a number saying 12-13%). The only truly easily found study (IE pops up immediately on google search) is federal: https://www.ojp.gov/pdffiles1/nij/grants/251115.pdf https://www.ojp.gov/pdffiles1/nij/grants/251115.pdf The DOJ funded the urban justice institute to figure this out back in 2017, and the rate they came up with was also ~11%. In there, they point to/compare with previous studies on DNA evidence that suggest a rate of 12.5%