How do we weed out false (or faulty) evidence?
I recently read a Law Review article titled "Judging Innocence", by Brandon Garrett a professor at the University of Virginia Law School. If you are so inclined, The article was based on a study of all the DNA exonerations, which was 200 at the time. The purpose was to review the claims made in the post-conviction litigation, and see how successful courts were at identifying factually innocent defendants. As you would guess, the answer is not too well. The author also identified a control group of similar cases to determine if there is any difference in how they were handled.
The article contains a lot of data, and enough tables to make a scientist proud. I cannot hope to cover everything in one post, so I thought I would divide it into several.
We already know the types of evidence used to obtain wrongful convictions. 79% involved eyewitness identification, 57% involved faulty forensics, 18% involved informant testimony and 16% involved confessions. Given the prevalence of eyewitness ID's, you would think that challenges in the post-conviction process would be fairly common. After all, you know you didn't do it, so the ID has to be bad. Only 28% percent of the wrongfully convicted defendants challenged the constitutionality of the identification though. That result is no doubt due to the difficulty of making such a claim; of those who made such claims, none were successful. In other words, there is almost no way to challenge a mis-identification on appeal.
Challenges to forensic evidence didn't fare much better. 32% of defendants challenged such evidence on some basis, and only 8% were successful. What I found interesting was how forensics were used. The study found that more than 1/2 of the cases involved improper testimony by the forensic examiners; in other words, it was improper based on the science known at the time. Even with that, challenges were rarely successful.
There was little new about hair evidence. It's not only unreliable; the experts also tended to say more than should have been able to.
Another interesting fact is that many of the cases involved more than one type of evidence. For example, there were cases involving both bad ID's and faulty forensics, as well as other combinations. The fact that the one corroborated the didn't make them reliable.
So what does this tell us? Among other things I believe it is that the post-conviction is not good at providing relief for innocent defendants. For those who faith in the justice to uncover mistakes, its misplaced. There a probably many reasons for that, will be discussed in future posts. Stay tuned.
